PAGENO="0001" COMPETITIVE PROBLEMS IN THE DRUG INDUSTRY ~, 7O ~2 ~ii~ HEARINGS BEFORE THE SUBCOMMITTEE ON MONOPOLY OF THE SELECT COMMITTEE ON SMALL BUSINESS UNITED STATES SENATE NINETIETH CONGRESS FIRST AND SECOND SESSIONS ON PRESENT STATUS OF COMPETITION IN THE PHARMACEUTICAL INDUSTRY PART 5 DECEMBER 14, 19, 1967, JANUARY 18, 19, AND 25, 1068 Printed for the use of the Select Committee on Small Business U.S. GOVERNMENT PRINTING OFFICE 81-280 WASHINGTON 1968 ~ For sale by the Superintendent of Documents, U.S. Government Printing Office Washington, D.C. 20402 - Price $2. PAGENO="0002" SELECT COMMITTEE ON SMALL BUSINESS [Created pursuantto S. Res. 58, 81st Cong.J (90th Cong., first and second sess.) GEORGE A. SMATHERS, Florida, Chairman JOHN SPARKMAN, Alabama JACOB K. JAVITS, New York RUSSELL B. LONG, Louisiana HUGH SCOTT, Pennsylvania WAYNE MORSE, Oregon NORRIS COTTON, New Hampshire ALAN BIBLE, Nevada PETER H. DOMINICK, Colorado JENNINGSRANDOLPH, West Virginia HOWARD H. BAKER, Ja., Tennessee E. L. BARTLETT, Alaska MARK 0. HATFIELD, Oregon HARRISON A. WILLIAMS, Ja., New Jersey GAYLORD NELSON, Wisconsin JOSEPH M. MONTOYA, New Mexico FRED R. HARRIS, Oklahoma WILLIAM T. MCINARNAY, Staff Director and General Counsel JAMES H. GROSSMAN, Minority Counsel MONOPOLY SUBCOMMITTEE GAYLORD NELSON, Wisconsin, Chairman JOHN SPARKMAN, Alabama HUGH SCOTT, Pennsylvania RUSSELL B. LONG, Louisiana MARK 0. HATFIELD, Oregoa WAYNE MORSE, Oregon JACOB K. JAVITS,* New York GEORGE A. SMATHERS,* Florida officio member. BENJAMIN GORDON, Staff Economist SUSAN H. HEWMAN, Research Assistant II PAGENO="0003" CONTENTS Statement of- Binkley, Howard, vice president for administration and planning, Pharmaceutical Manufacturers Association, Washington, D.C.; accompanied John M. Firestone, Ph. D., professor of economics, the Page City College of New York, New York, N.Y 1690 Comanor, William S., Ph. D., professor, Department of Economics, Harvard University, 7146 Cambridge Street, Cambridge, Mass~.. - 2041 Conrad, Gordon R., senior staff associate, Arthur D. Little, Inc., 25 Acorn Parkway, Cambridge, Mass 1667 Cootner, Paul H., Ph. D., professor of finance, Massachusetts Insti- tute of Technology, 77 Massachusetts Avenue, Cambridge, Mass_ 1618 Cutler, Lloyd N., special counsel, Pharmaceutical Manufacturers Association, Washington, D.C 1615 Firestone, John M., Ph. D., professor of econOmics, the City College of New York, New York, N.Y.; accompanied by Howard Binkley, vice president for administration and planning, Pharmaceutical Manufacturers Association, Washington, D.C 1690 Kelly, William H., member, Bureau of Economics, Federal Trade Commission, Washington, D.C.; accompanied Dr. Willard F. Mueller, Chief Economist and Director, Bureau of Economics~_ - 1807 Markham, Jesse W., Ph. D., professor: of economics, Princeton Uni-, versity, Princeton, N.J 1667 Mueller, Dr. Willard F., Chief Economist and Director, Bureau of Economics, Federal Trade Commission, Washington, D.C.; ac- companied by Dr. Russell C. Parker, Assistant Director; and William H. Kelly, member, Bureau of Economics 1807 Parker, Dr. Russell C., Assistant Director, Bureau of Economics, Federal Trade Commission, Washington, D.C.; accompanied Dr. Willard F. Mueller, Chief Economist and Director, Bureau of Economics 1807 Plotkin, Irving H., consultant in econometrics and operations research, Arthur D. Little, Inc~, 25 Acorn Parkway, Cambridge, Mass 1638 Schifrin, Leonard G., - Ph. D., director, Department of Economics, College of William and Mary, Williamsburg, Va 1863 Squibb, George S., former vice president, E. R. Squibb & Sons, 3-Cedar Island, Larchmont, N.Y 1555 Steele, Henry B., Ph. D., associate professor of economics, University of Houston, 105-A Heyne Building, Houston, Tex 1901 Whitney, Simon N., Ph. D., professor of economics, New York University, New York, N.Y 1725 APPENDIXES Correspondence between Mr. Benjamin Gordon and Dr. P. M. Costello re technological progress in the pharmaceutical industry 2113 Paper, "Risk and Corporate Rates of Return," by I. N. Fisher and G. R. Hall, The Rand Corp., Santa Monica, Calif 2120 Additional Submissions by the Pharmaceutical Manufacturers Association 2129 III PAGENO="0004" lv CONTENTS SALIENT EXHIBITS Report, "Risk and Return in American Industry-An Econometric Page Analysis," by Gordon R. Conrad and Irving H. Plotkin 1746 Study, "Trends in Market Shares for Ethical Pharmaceutical Products," by Gordon R. Conrad 1783 Supplemental statement of Dr. Willard F. Mueller, Director, Bureau of Economics, Federal Trade Commission 1843 Article, "The Ethical Drug Industry: The Case for Compulsory Patent Licensing," by Dr. Leonard G. Schifrin 1890 Article, "Monopoly and Competition in the Ethical Drugs Market," by Dr. Henry B. Steele, from Journal of Law and Economics, October 1962 1950 Article, "Patent Restrictions and Price Competition in the Ethical Drugs Industry," by Dr. Henry B. Steele, from Journal of Industrial Eco-~ nomics, July 1964 1971 Article, "The Fortunes of Economic Reform Legislation: The Case of the Drug Amendments Act of 1962," by Dr. Henry B. Steele, Rice University 1998 Article, "Income Opportunities and Physician Location Trends in the United States," by Dr. Henry B. Steele and Dr. Gaston V. Rimlinger, from Western Economic Journal, spring 1965 2011 Article, "An Economic Interpretation of the Spatial Distribution of Physicians in the United States," by Dr. Gaston V. Rimlinger and Dr. Henry B. Steele, from Southern Economic Journal, July 1963 2025 Article, "Research and Competitive Product Differentiation in the Phar- maceutical Industry in the United States," by Dr. William S. Comanor, from Economica, November 1964 2069 Article, "Research and Technical Change in the Pharmaceutical Industry," by Dr. William S. Comanor, from Review of Economics and Statistics, May 1965 2078 Article, "The Drug Industry and Medical Research," by Dr. William S. Comanor, from Journal of Business of University of Chicago, January 1966 2086 Article, "Advertising Market Structure and Performance," by Dr. William S. Comanor and Dr. Thomas A. Wilson, from Review of Economics and Statistics, November 1967 2092 HEARING DATES* December 14, 1967 1555 Afternoon 1579 December 19, 1967 1609 Afternoon 1661 January 18, 1968 1807 January 19, 1968 1863 January 25, 1968 2041 * The testimony for May 15, 16, 17, June 7 and 5, 1967, appears in pt. 1 of these hearings; the testimony for June 27, 28, 29, July 24, and Aug. 8, 10, 1967, appears in pt. 2 of these hearings; the testimony for Sept. 13, 14, 29, and Oct. 13, 1967, appears in pt. 3 of these hearings; the testimony for Oct. 31, Nov. 9, 15, 16, and 28, 1967, appears in pt. 4 01 these hearings. PAGENO="0005" COMPETITIVE PROBLEMS IN THE DRUG INDUSTRY Mr. SQUIBB. Not at all, Sena Senator N1~soN. All right, if time to time, as questions occur Mr. SQUIBB. I have a few mt a little background which I maj Senator NELSON. All right, ~ read into the record your biogra Mr. SQUIBB. Yes, I will do th Squibb. I live at 3 Cedar Islai at the request of the committe Harvard University, I have fi [Hatfield. don, staff economist; James H. Gross- H. Hewman, research assistant; and ive director, staff of Senator Nelson. of the Monopoly Subcommittee of the ee will come to order. mnittee is very pleased to have the op- a Mr. George S. Squibb. He was former & Sons, and a leader in the drug in- ne has been advocating reform within egards pricing policies. Mr. Squibb, I ~timony very carefully. Your analysis ik, opened the door to consideration of lustry. It is reassuring to know that udgment, on many quality drugs, and you are willing to fight for reforms The committee is most pleased to have u may present your testimony in any you do not have any objection to inter- oceed to present your testimony. or. As you see fit any time. rou will proceed, Mr. Squibb, and from e will interrupt you. oductory remarks as to who I am and read first. r. Squibb, you may proceed. Will you ~hical sketch? t now, Senator. My name is George S. d in Larchmont, N.Y., and I am here For 30 years, since graduating from led a wide variety of positions with THURSDAY,~ DECEMBER 14, 1967 U.S. SENATE, MON0~OLY SUBcoMMrrrei OF THE SELECT C~OMMITrEE ON SMALL BusINEss, Washington, D.C. The subcommittee met, purskiant to call, at 10:20 a.m., in room 1202, New Senate Office Building, S~nator Gaylord P. Nelson (chairman of the subcommittee) presiding. Present: Senators Nelson an Also present: Benjamin Go man, minority counsel; Susa: William B. Clierkasky, li~gisla Senator NELSON. The hearin Senate Small Business Commi Today, the Monopoly Subc portunty to hear testimony fro vice president of E. R. Squib dustry. Mr. Squibb for someti the industry, particularly as have read your 28 pages of t of the drug industry has, I thi major reforms within that h prices can be reduced in your that industry leaders such as which will make this possible. you here this morning, and y fashion you see fit. I trust that: ruption for questions as you p STATEMENT OF GEORGE S. QUIBB, FORMER VICE PRESIDENT, E. R. SQUIBB & ONS, LARCEMONT, N.Y. 1555 PAGENO="0006" 1556 COMPETITIVE P1 )BLEMS IN THE DRtTG INDUSTRY TE. R. Squibb & Sons, and that organization. I am a great-grandson Squibb, and represent the pany, which was founde Squibb I have served in the plants and laboratori in production planning then served as headqua points. I was elected secr boards of directors of all Chemical Corp. I was th director of administratio ment as director of trade ant director and directo marketing. After a series of corp vice president for legal a the company. During my career I h management committees, types of management fun I have participated as of professional and trade I have been active in co~ major industry tradi~r4 man of the National Phar I am a lawyer, member study at the Food and Dr business backward and fc concerned with its future all matters affecting pubi us in the years ahead. I have followed very nesses who have a~pea~ studied it, and familiariz least two major conclus all those concerned with health plans insofar as the industry will not adn help clear the air of misu I would like to try to marks based on long mdi ful analyses of things pa in the future. Now, if I may go ahead Senator NELSON. Yes; i Mr. SQUIBB. The publ prices are too high-tha scribed medicines-and t cal companies were not so Because the Governme am currently on a consultantship basis with of Dr. Edward R. Squibb, the founder of fourth generation of the family in that corn- in 1858. During my 30 active years with very part of the organization. I worked in ~. I was a salesman and detail man. I served nd control. I managed a branch office and ters manager for all Squibb distribution tary of the company in 1949, and served the ~quibb Cos. until the merger with Mathieson n made assistant to the president, and later * In 1954 I started back in the sales depart- and customer relations, progressed to assist- of sales and finally to vice president for rate management changes I was appointed d public affairs and am now a consultant to ye served on boards of directors, executive patent and research committees, and in all ;ions at all levels. peaker or panelist on programs of all kinds groups in nearly every State in the country. aitteewoçk for the PMA and NPC, the two ips, and was elected to two terms as chair- riaceutical Council. of the New York Bar, with special graduate ig Law Institute. I know the pharmaceutical rward, most everybody in it, and am vitally both personally and from the conviction that c health must be of vital importance to all of losely all the testimony of the various wit- I before this subcornmitee. I have re,ad it, d myself with it, and have come up with at ons: First, there is a real problem facing ~he establishment of Government-supported [rug procurement is concerned, and second, it it, talks around it, and contributes little to derstanding, suspicion, and distrust. mprove on this performance with some re- stry experience, careful study, and thought- t with some specific suggestions for changes Senator? ease proceed. c believes that pharmaceutical prescription they should be paying less for their pre- at they could be paying less if pharmaceuti- greedy-or used less wasteful sales methods. it, State and Federal, now is preparing to PAGENO="0007" COMPETITIVE PROBLEMS IN THE DRUG INDUSTRY 1557 pay so mudh of the medical bill of the public-a bill the potentiaJ of which incidentally still is grossly understated and misunderstood- legislators have seized upon this public conviction of high prices on drug products to attack the industry for its behavior and to use it as the justification for proposals of new regulation and control of a magnitude far beyond anything that has been felt to date. Senator NELSON. At the end of the first sentence you say: "Or used less wasteful sales methods." Would you elaborate on what you mean by wasteful sales methods? Mr. SQUIBB. I am talking here, of course, about the expenditures which seem to be large, are actually large for the detail man and for the advertising and promotional methods of one type or another that the industry uses to further the sales of its products. These seem on the surface, to the public and to the industry critics, to be unneces- sarily large and unnecessarily indiscriminately used. I develop that a little bit later specifically when I get to some of these areas of expense. Senator NELSON. All right. Mr. SQUIBB. The problem that faces the pharmaceutical industry today relative to public criticism of its pricing and profit structure can be stated just that simply. What to do about it, if indeed anything, is not so easy for the pharmaceutical manager to determine. All sorts of complex social forces and economic facts must be understood, sorted out, and balanced against each other in an effort to foresee the long- range effects of this concept of pharmaceutical prices, a concept which is now generally accepted outside of the industry itself. The only answer that the pharmaceutical industry has given to date is to "tell its story," to "improve its image," to "explain its con- tributions to medical knowledge and the health of the people." These efforts for the most part, have been ineffectual in changing the mind of the public about the nature of the prices it pays for its drugs, and yet the industry keeps at it with dogged persistence issuing new studies in depth-new comparative analyses of profits country by country and industry by industry, new interpretations of the value of its prod- uct contributions, new explanations and justifications for its modus operandi. To all of its critics, temperate or intemperate, expert or lay, legislative or academic, it keeps on saying, in effect: "You just do not understand us-there is the truth which only we can properly explain." The industry has been primarily concerned, and on doubt properly so, with its own internal problems induced for the most part by reason of its phenomenal expansion and by the possibilities for even further growth in widely divergent areas. It has tried to meet, some- times successfully and sometimes not, the day-to-day problems arising from a complex product mix in a complex marketing situation all the while keeping a sharp eye on earnings and corporate health. It has tried honestly and hard to produce good products, and has gen- erally not been satisfied to rest on today's achievement, but to press on to something better. It has been diligent within its own self-estab- lished limits in assuming its responsibilities, and it is truly hurt by those who suggest anything to the contrary. However, diligence, devotion to ideals, and just plain hard work do not necessarily imply intelligence and good judgment, and it is in PAGENO="0008" 1558 COMPETITIVE PROBLEMS IN THE DRUG INDUSTRY this light that the pharmaceutical managers should look today at their policies and their operational Procedures if they are to control their own futures. Mr. GORDON. What do you mean by "within its own self-establisI~ed limits"? Mr. SQUIBB. Well, I think the industry sets forth for itself the areas in which it is properly concerned. It does not accept limits that are proposed from outside, critics or from the public, even generally as to what should be the proper limits of its responsibility. The in- dustry decides what its responsibility is, aild has carried out that ic- sponsbility very carefully and diligently. It is making its own internal judgments as to what are its own duties in this concept. Let us take a look, just as to how the industry prices its products. Every product in the marketplace has a cost, a price and a value. Costs of products in all industries are difficult to determine precisely, not only because manufacturers are naturally reluctant to discuss them, but also because there is a. real difference in opinion as to what they actually are within a given company's financial structure. Meth- ods of accounting, allocation of overhead and variable production expenses, problems in separating out indirect distribution costs and applying them fairly to diverse items, all cloud over any effort to fix clearly just what the "cost" of a. product is. This is often an even more irritating difficulty to tile manufacturer than to the outside in- vestigator, and many a firm has undergone real financial distress because it could not properly establish its product costs. All this applies especially to the pharmaceutical industry. With their multiproduct lines of greatly dissimilar products requiring all sorts of different production leadtimes, storage and thipping vari- ants, and sales promotion programs ranging from the simplest to the most complex techmcally of any that can be imagined, Pharmaceutical companies find product cost determination a real problem even with- out the almost overwhelming difficulty of proper allocation of basic and specific research and development costs which are a substantial part of their annual operating expenses. However, it seems reasonably certain that costs of major volume products are a much smaller per- centage of trade prices than is usual in consumer goods industries. This can be deduced from much published evidence of costs estimated and real, and the facts that price ranges are enormous, and profits better than average. Senator HATFIELD. Mr. Squibb, could I interrupt? Tile sentence be- fore last: "However, it seems reasonably certain that costs of major volume produc.ts are a. much smaller percentage of trade prices than is usual in consumer goods industries." I presume here ot.her consumer goods industries. Could you give us a. little further explanation of comparisons which you are drawing here? Mr. SQUIBB. Well, what I am trying to compare here are the factory costs, the direct labor and tile actual out-of-pocket costs to create a prodluct for the pharmaceutical business, compared with consumer products, such as food products and pac.kaged goods products that are distributed and bought in the general markets by the public. I think that we have a large mass of evidence that has been deduced and put PAGENO="0009" COMPETITIVE PROBLEMS IN TIlE DRUG INDUSTRY 1559 together from various directions, `that the actual costs, of pharmaceu- tical products, are quite low compared to the price that is set. I do not have tables, and `this evidence here.. This is in the background. Much of this is in testimony already giveli `to the committee. Senator NELsON. Just for clarification, are you referring to the kind of testimony that appeared sometime ago in `the Kef'auver hearings, and I think once or twice here, in which efforts were macic by the committee or by some witnesses `to establish the cost of the basic pro- duct and then acid reasonable markup? I think we had some testi- mony `a while back `that `the cost of making 100 tablets of a particular drug was $1.80 or something like that including profit, and the tablet was selling for $16 or $18 a 100. Is that the kind of differential that you are `talking about? Mr. SQUIBB. That is `the cost I am:'talking `about, the basic cost of `the package. I think we have seen `also laid out before the committee at various times in the past weeks the other elements of expense that go into the final pricing of the product, and because profits, research, and `selling expenses are larger in this industry percentagewise than in most other industries, it follows inevitably that the cost has to be smaller. I think this is a matter of general `agreement by anyone who examines the subj ect `objectively. Mr. GORDON. Mr. Squibb, to corroborate your statement you may be interested to know that according to the 1961 Annual Survey of Manufacturers, U.S. Bureau of the Census, the wages as a percentage of value added for all manufacturing is 33.5 percent. For chemical and allied products it is 17 percent. For drugs it is 11.1 percent. That is labor cost. With respect to materials cost, it is also pretty low for the drug industry; and with respect to investment data, fixed assets, as a per- cent of sales for `all manufacturing is 28.5 percent, chemicals and allied products 37.6 percent, and drugs 25.6 percent. So the statistics of the Bureau of the Census certainly corroborate w'hat you have said. Mr. SQUIBB. I think this is a basic fact which actually determines much of what happens later on in setting the pricing structure for any product the fact that the cost is comparatively low. Generally speaking and within a given industry, the price of a product has a predictable relationship to the cost, and in turn the value is close to the price, such being the nature of our competitive economic system. In the drug business, ho'wever, prices are unpredict- able; often they seem to bear no normal relationship to cost. More often than not there are several widely divergent prices for the same product, the highest frequently in~ the largest market and the lowest in the smallest market. Indeed it `is difficult to `determine just what the price is on many important prescription drug products. Senator NELSON. We have had considerable testimony, which I take it you have read, showing the difference between the high prices charged for drugs `to the retail pharmacist, `and the dramatically lower price that the same companies bid for the same drugs to the Defense Supply AgenQy or the Veterans' Administration or the city of New York or major hospitals. These differences amount to a couple thou- sand percent. There are cases where 100 tablets would sell on the retail market for $17 or $18, while the same company bids $1.20 to New PAGENO="0010" 1560 COMPETITIVE PROBLEMS IN THE DRUG INDUSTRY york for the same drug. What explanation do you have for the same drug on the same day being sold to the Defense Supply Agency or to large hospitals for a tenth or one-twentieth of the price that is charged by the same company to a retail pharmacist? Mr. SQUIBB. Well, you put your finger on the whole difficulty. There seems to me to be no good explanation for this. These price differ- entials have developed over the years out of a series of market con- ditions and other pricing situations which I am going to expand on a little bit, but there is no rational explanation that I think can be offered for these very large price variants. This is precisely what has led the industry into the problems that exist today. I believe that this is the point that the industry should face up to, namely, the fact that they cannot give a reasonably accurate explanation of these differ- entials that is acceptable to outside economists or even the public generally. If they would accept this fact and examine their pricing policies more specifically and more realistically, more in view of the way the outside world looks at them, I think a good part of their difficulties would be over, because they would obviously cut out the practice that you have just mentioned. Senator NELSON. It seems to me in viewing the situation that there might be some explanation in the fact that when drug companies offer bids to big government agencies or to the city of New York they are in a tough competitive situation. Bids are solicited and a number of companies come in with an offer, and compete with each other for the business of the particular big hospital or government agency, whereas in the reta.il market the use of a brand name seems to eliminate coin- petition. The physician does the prescribing, and if a trade name is established among a certain percentage of the doctors for a drug that they want to prescribe, does not that tend to eliminate genuine competition. Mr. SQUIBB. Well, I think there is a bigger factor than the fact that the doctor is presold on what he wants by the trade name, that does tend, as you say, to eliminate competition at the retail level. This is the fact that the price that is paid for the prescription by the patient, is determined basically by the price that the manufacturer has charged the retailer for the merchandise. Now, this price is set and is very rigid. There is no competitive activity on this price from retailer to retailer, from large retailers to small retailers, in any way. This price is affected as I point out later on, by Robinson-Patman concepts, and there is a rigidity of pricing at the retail level. This has existed remarkably firmly, and I think this more than anything else sets the top level for the situation which you have just described. Senator NELSON. But we do have in the testimony that I am sure you have read, an instance of a drug on the market that is selling for as mitch as $17 or $18 a hundred by a large firm down to as low as $2 a hundred by another distinguished brand name company and for as little as 78 cents by some lesser known companies. Yet, the company that is charging $17 or $18 a hundred on the retail market and holding a large share of the market does turn around on the same day and bid one-tenth or one-fifteenth as much as to the New York City or the Defense Supply Agency. Mr. SQUIBB. That is right. PAGENO="0011" COMPETITIVE PROBLEMS IN TIlE DRUG INDUSTRY 1561 Senator NELSON. Now, they would not sell anything to New York City or the Defense Supply Agency, if they were charging $17 or $18 a hundred because they would be outbid every single time. So they offer a price of one-tenth or one-fifteenth as much. On the other hand, they go into the retail market and charge the $17 or $18 a hundred or whatever it may be, depending upon the drug, several times as much, and do manage to hold a share of that: market and make a substantial profit. This is the part that is difficult to understand, unless it is simply the fact `that the trade name has become well enough established with the doctors who will prescribe their, drug regardless of what the price is. Mr. SQUIBB. That is exactly, of course, what has happened. There is no price competition or very little price competition at the so-called retail level. If you get your trade name established with your doctors through your promotion activities and your advertising and the gen- uine value of your product, the largest drug chain in the country or the smallest drugstore or any buying organization of the retail pharma- cist still has to pay the $16 because the price is established and pro- tected for competitive purposes. There is not any price competition that the manufacturer has to worry about at that level, and he is very happy, I think, about this. `Senator NELSON. Go ahead. Mr. SQUIBB. Value, continuing at the top of page 4, value on the other hand, is often far above any of the prices set for the product. If health and indeed life itself are measured in terms of dollar value against the cost of the medicines that contribute to them `and in many cases guarantee them, drug products give probably the finest value of anything that can be purchased today. Yet, in ~pite of these tremen- dous values, the public belief persists that prices of drugs are too high and costs inflated by unnecessary and ,wasteful operational procedures by their manufacturers. The reason for this seems to be in the unusual, to say the least, pricing policies of the pharmaceutical industry. It must be understood, however, that the industry operates in a complex-partly controlled and partly open-market in which it deals on a direct basis with the widest possible range of customers both in term's of their capacity for product utilization, and in terms of their competitive and noncompeti- tive natures. At the retail level, that is the retail pharmacy, prices on direct sales to pharmacies large and small-some are very large, others very small-are affected by Robinson-Patman. Identical prices must be charged to the largest chain and its smallest independent competitor for the same drug package. Senator NELSON. I am not an authority on the Robinson-Patman Act. That act does not prohibit a comp'any from giving advertised price reductions for quantity purchases, does it? Mr. SQUIBB. As long as it gives those advertised quantity price re- ductions to everybody on an equal basis, and that t'hey are really bona fide quantity price reductions. Senator NELSON. That they are related to the cost? Mr. SQumi~. Related to the cost. They have to be related to the cost and also related, as I point out here in this paragraph, to what the PAGENO="0012" 1562 COMPETITIVE PROBLEMS IN THE DRUG INDUSTRY customer could normally buy. The price cannot be established for a large quantity that only the biggest chain drugstore could reasonably use. This is a trick wl~ich can be done, but which is not done in the drug business. It just cannot be done by coming out with a 100,000-tablet price which only Wralgreen, for example, could use. This would be a sham, giving Wa.lgreen a price advantage which is improper because the little fellow cannot either physically or financially buy 100,000 of these tablets. I think the pricing practice in the drug business in this area is extremely careful, is extremely closely advised by legal coun- sel, and is strictly within the spirit and the letter of this law which protects the smaller merchant against the larger potential commercia.l competition that lie has. I think this is a very important fact which has affected prescription prices and is affecting today's prescription prices. Sometimes I think it is not always completely appreciated. No matter what you think about Robinson-Patman, you have to realize that this is one of the controlling factors in keeping these Prices rigid and keeping them at the level at which they were originally established. There are iio pressures to bring them down, no normal competitive pressures in the marketplace. Senator NELSON. We have just talked about the situation where a company will sell 100 tablets of the same drug to the retail pharmacist for $17 and $1 to New York City. If you then allow a company to buy 100,000 tablets, and give it the price which you give to New York City, then you would end up driving the people out of business who are paying the $17 per 100. Mr. SQUIBB. Out of business, that is the idea.. Senator NELSON. If you didn't have that price differential in the first place then you would have quite a different problem. Mr. SQUIBB. Yes, and the fact that the industry has permitted it more and more and still seems to be following the practice as you pointed out, of having this tremendously wide variance, continues to keep them in trouble at their end. They do not seem to learn. They keep developing these very wide variants and I think it is going to get worse because the other area which we have not talked about yet, the private hospital area, is developing more into competition with the retail pharmacy area of operation, and in this latter area Robinson- Patman at this moment does not apply. Hospitals are nonprofit and are not judged to be in competition. Senator HATFIELD. Mr. Squibb, let me clarify the record a bit by asking this. As I understand it, you take issue with some of the in- dustry witnesses we have had who have stated very specifically that quantity buying does have a distinct impact upon the pricing dif- ferentials. You say in your testimony that package size and quantity discounts are minimal. Mr. SQUIBB. I Say that they are minimal at the retail level; that is, in sales to retail phar~nacies. They certainly are not minimal in over- all sales to private institutions, Government and privately controlled hospitals. I am talking here about the rigidity of the retail price list, in which five-percent quantity discounts or ten-percent quantity chs- counts are the maximum that are given. Senator HATFIELD. As it would relate to the average consumer pur- chasing through a corner drugstore, the quantity buying by that drug- PAGENO="0013" COMPETITIVE PROBLEMS IN THE DRUG INDUSTRY 1563 store would have an effective price differential which might exist be- tween that drugstore and some other one? Mr. SQUIBB. Very little, because even if you go to buy your prescrip- tion from the largest chain, that chain taking full advantage of what- ever quantity discounts that are, can only buy perhaps, a or 10 percent less at the most. That is all that any retttiler can get in advantage over another retailer. There is very little at that level. Senator HATFIELD. That is in spite of what other witnesse.s have said? Mr. SQUIBB. Well, I am sure tha.t thete are, if you examine the trade list of manufacturers, quantity discounts to retail pharmacies, but they are small. I am not talking about hospital lists. There are special Prices there, but quantity discounts to retail pharmacies are minimal, m my opinion. They go 5 to 10 percent, and they cannot be any more than that or you are in trouble right away with the Robinson-Patman, unless you can give this discoi.mt to everyone. Senator HATFIELD. But, they are significant in terms of the large institutions? Mr. SQUIBB. Oh, very significant. I think we have covered the para- graph at the end of page 4 and the top of page 5, which I was talking to Senator Hatfield about.. Price cutting, discrimination, favored customers, under-the-counter deals, hidden discounts are no factor in sales of prescri.ption products direct to the retail pharmacy trade by pharmaceutical manufacturers. It must be stressed that the retail Pharmacist does most of the prescription business in the country, and it is from this source, generally speaking, that the industry not only gets most of its sales volume, but an overwhelming percentage of its profit. Pharmaceutical manufacturers also sell to wholesalers who, in turn, fill orders from retail pharmacies. There are sharply competing whole- salers in every city and State in the Nation, and the price to each of them on a given pharmaceutical product from a given manufacturer is identical. It has to be for obvious legal reasons, and there are no exceptions to this rule. The wholesaler price is generally discounted from 5 to 20 percent from the so-called trade-retailer__price. Price competition among wholesalers in passing along to the retailer part of tihis discount is a common procedure, but this also is controlled by Robinson-Patman, and wholesalers must offer these discounts equally to all his competing customers. Attempts have been made to tie in services, order quantity and order timing factors, and credit privi- leges with discount sharing, but generally speaking, ttie variance in price produced by this wholesaler practice is not large and it remains but a small factor in the determination of the ultimate price of the product on the consumer's prescription. Occasionally and especially in large metropolitan areas when competition among wholesalers for retail accounts is bitter, products can `be purchased by the pharmacist from the wholesaler more cheaply than he can buy them direct, which fact in itself reflects the rigidity of the direct-selling manufacturers list., and the firm observance of the general principle of Robinson- Patman. While prices of drug products remain generally rigid to the retail and wholesale druggist, quite a different picture is shown on prices to PAGENO="0014" 1564 COMPETITIVE PROBLEMS 1_N THE DRUG INDUSTRY hospitals-non-profit or Government-operated. Here the widest varia- tions exist and from this area of pricing, evidence is deduced that other pricing is non-competitive, unreasonable and unfair. While many pharmaceutical manufacturers maintain so-called hospital pricelists, most hospitals, large and small, find ways to buy their requirements at lesser figures particularly on important large volume products which bear relatively high-price tickets to start with. It should be understood at once that Robinson-Patman does not apply to non- profit institutions who are, per se, not juged to be in competition with each other. Therefore, while Walgreen Drug Co., the largest drug chain in the country, pays $5 for a bottle of a brand of diuretic tablets even if that purchase is repeated a 100 times in a week as compared with the one-a-month $5 price for the item by an independent phar- macy, a large hospital can and often does beat that price down to $3.50 or less, and even the small hospital can expect to get at least 10 percent off the retail price if it has any buying acumen at all. This is done by one or more of several procedures which hospitals are using to ever-increasing effect as their size and importance grow under cur- rent expansive trends of medical treatment supported by Government and insurer funds. Request for competitive bids, formation of buying associations, threat to manufacture or subdivide themselves, control over brand specification within its own walls by therapeutic or formu- lary committees, exposure of products to large numbers of physicians who can be expected outside the hospital to use a product with which they have become fan'iiliar within-such usage incidentally, at consid- erably higher prices-all these things contribute to the hospital's ability to obtain drug products at lower prices than the retailer or wholesaler. Senator HATFIELD. Mr. Squibb, could I interrupt here? WThat would happen to prices if the Robinson-Patman Act were applied to hos- pitals, in your opinion? Mr. SQUIBB. The prices would firm up, take the same rigid nature that they have in the retail trade. The manfuacturers would probably welcome it in terms of removed competitive factors in pricing to hos- pitals. Somewhere in the middle of the presei~t wide range you would have a hospital price that would be arrived at after a little give and take for a while. There would be a rigidity of prices and a lack of competitive bidding, which I think would tend to raise the overall average of drug prices very substantially. Senator HATFIELD. It would raise the prices? Mr. SQUIBB. It would raise the average. Senator HATFIELD. WTould the companies be able to establish uni- formity on prices? Would they be subjected at that point to- Mr. SQUIBB. Oh, no, they do it themselves. They have a list. They would set a hospital pricelist. The price would be $5 or whatever on an item a.nd that would be the published hospital list price available for all to see, including the Government regulatory agency. The price would be $5 and there would be no hospital large or small of any type which gets anything different than $5. It could not, if Robinson- Patman were applied. This would be the price, just as $5 or whatever the price may be to the retailer. Now, they could have quantity discounts to hospitals, bift they would be published quantity discounts. There would be a 25,000 price PAGENO="0015" COMPETITIVE PROBLEMS IN THE DRUG INDUSTRY 1565 or 100,000 price on a tablet. This would be published in your list and all hospitals would be allowed to buy at~this level. Senator HATFIELD. You mean from all the drug companies within say $5 for the specific product? Mr. SQUIBB. No, they would establish a pricelist. Senator HATFIELD. Individual price? Mr. SQUIBB. Individual pricelist. Senator HATFIELD. You would still have the competition, though? Mr. SQUIBB. Oh, yes, you would have cross competition, because a similar product from another company would be at approximately the $5 level, but prices would tend to freeze at that level, that rigid level, and do away with price competition on that product. Senator HATFIELD. Well, are drug companies taking losses then? Could I infer from your testimony that drug companies are taking losses in selling to hospitals on the basis that they feel that they can make up for those losses by the influence, through having their product in the hospital, upon the physician who goes outside of the hospital and engages in private practice? Mr. SQUIBB. This answer to the question of taking losses in hospital sales or Government sales at low prices is a very difficult one to estab- lish. Obviously if they sold their entire line, their entire product mix or even all of any one product at that price level, they might well take losses. The accounting difficulty here in applying your costs to the hospital part of your sales is fantastic. Within a given company you would probably have to say that if, they sold their entire product output at the hospital level, they would take a loss. Senator HATFIELD. That they would ? Mr. SQuIBB. They probably would frequently, because their whole structure, selling structure and other structures, is predicated on their ability to sell at a higher level to the retail pharmacy. Senator HATFIELD. You have given me some generalities here, but what happens in the specific case of Squibb? Mr. SQUIBB. Squibb has a hospital pricelist. Senator HATFIELD. Do you take any losses? Mr. SQUIBB. You would be very hard put to it to say that we lost at a given price to the hospital. I think the answer to your question would be "No." But you would lose, if you put the burden of advertis- ing and selling expenses at that leiel. You do not price out your product return at each different area of sales. We would probably sell the same product to the Government even cheaper than we sell it to a hospital. There may be another lower price yet. Senator HATFIELD. But, do you have any lines that if you sold out at that price you would take a loss? Mr. SQUIBB. Oh, yes; you have a overall cost figure for your product, and you have an average price level which has to certainly cover your basic cost. Plus your selling costs, research costs, other costs in other brackets of your operating statement. Overall actually, when you look generally at a product you sell, and I cover this later on, there is an average selling price, a return to the company which is the sum of all the sales at all the prices, including the retail price, the hospital price level and the Government prices, and these factors vary pro- portionately greatly from product to product in each category. PAGENO="0016" 1566 COMPETITIVE PROBLEMS IN THE DRUG INDUSTRY Senator HATFIELD. Can Squibb calculate a certain impact of having their product in the hospital and affecting the attitude of physicians who are using their product in the hospital and then what they hope the physician will do outside of the hospital where there would be a different price? Mr. SQtTIBB. Yes; I think they can generally. We have seen this. We know that if a certain product is used in a teaching hospital, where a large proportion of the physicians of a community practice, that the use of this product outside the hospital on private practice prescrip- tions will be larger. We learn this through prescription analyses and studies of the various market research prescriptio1i services. I think this is a real factor in the success of a product. Senator HATFIELD. This is pretty generally practiced among all the pharmaceutical houses? Mr. SQUIBB. Yes. Senator HATFIELD. As well as Squibb? Mr. SQUIBB. Yes; this is well known. Senator HATFIELD. And then, actually you feel that part of that price that you are giving to the hospital will be more than compen- sated for by the influence it will have on the physicians outside of the hospital practice? Mr. SQUIBB. That is right, and certainly the younger physicians, physicians in residence, and interns coming along who see in their teaching hospital a certain product being used and are familiar with it, who watch its reactions and its effect on patients for a number of years, this is a product which they understand. They will naturally pick it up later on in practice. This makes sense. We can prove it by market research examination, and, therefore, this is a reason to justify going to special lengths, pricewise, to get your product into this hospital. Senator HATFIELD. Do you feel that in the teaching hospital or in academic pursuits of medical education there is an adequate academic program in which the physicans, or the medical students and interns, are being given appropriate education in order to make judgments on pharmaceutical products for use? Mr. SQUIBB. No; surely there is not. Now, I judge this just on per- sonal opinion. I have not gone into the subject in specific detail but in conversation with many hundreds of physicians and examination of the testimony before the committee of the leading men that you have had here, it is obvious that there is not adequate background in this area. When I say there is not maybe there can never be, because of the size of the problem and the demands of other areas. Senator HATFIELD. So then, the pharmaceutical houses really sort of depend upon this in some instances, or let us say, they perhaps exploit it? Mr. SQUIBB. They exploit it. Senator HATFIELD. In such terms as getting their products into the hospitals and teaching hospitals at a lesser cost, in order to influence this intern or this young physician who has not been fully and ade- quately prepared in this area through the academic programs, and so it comes in primarily as a sales approach and a promotional approach, rather than so much a matter of analyses or evaluations or comparisons? PAGENO="0017" COMPETITIVE PROBLEMS IN THE DRUG INDUSTRY 1567 Mr. SQUIBB. That is right. Actually the sales in this area are studied and evaluated to establish whether this really does the thing that we think it does. Prescription studies and market research studies indicate that it does in fact. Senator HATFIELD. But, you feel that once he is convinced that Squibb is really a superior product he will then be using and prescrib- ing Squibb products outside? Mr. SQUIBB. That is right. Certainly the specific product that he has used, prescribed and seen used in the hospital, and hopefully this will spread even to others. Senator HATFIELD. That is where you make your profit. Mr. SQUIBB. That is where we make the bigger profit outside. Mr. GoRDoN. Mr. Squibb, even then the products are not being sold to the hospital at a loss, are they? Mr. SQUIBB. No; I do not think so. Mr. GORDON. They are making some money? Mr. SQUIBB. Yes; they are not selling at a loss. The difficulty of ac- counting here from company to company is a major problem. I do not think products are being sold to hospitals at a loss. Senator NELSON. I have the impression from what I have read and from testimony before the committee, that the use of formularies by hospitals and by cities in the purchase of drugs is a rapidly expanding practice. Is that correct? Mr. SQUIBB. Yes. Senator NELSON. Is this a trend that has been relatively accelerated in the past 10 years? Mr. SQUIBB. Yes. I think the development of formularies by hospi- tals has sort of been forced upon them by the problem of inventory, by the problem of drug selection which has become so vast and so com- plicated in the last 10 or 15 years with the emergence of so many new potent drugs. There is a problem of selection. There is a problem of inventory investment by the hospital which is under financial pres- sure anyway. The problem of just carrying every type of product there is just for an occasional use. There is a great deal of effort being spent to bring some order out of this very difficult `situation by the develop- ment of formularies. This has been since the war or even more recently than that. I would say in the last 10 years formularies have come into use in just about every hospital for this reason. Senator NELSON. We have had testimony on formularies from a number of pharmacologists, as well as representatives of some very fine hospitals, such as Montefiore in New `York, Los Angeles County Gen- eral Hospital, and a group health organization in the State of Wash- ington. They have testified that they make their formulary available for sale and use by doctors outside the hospital. In hospitals in smaller communities, 25-bed, 50-bed, 75-bed hospitals, which may not have the expertise to development of a good formulary of their own, is there anything comparable available for use and if so, are these hospitals using it.? Mr. SQUIBB. Well, I think even the smallest hospitals are using for- mularies. I think the American Association of Hospital Pharmacists has developed a st.a.i~dard or a basic formulary which is suitable for hos- pitals of any size, I think every hospital w-ould, as a matter of fact, have S1-2S0-----GS----pt. 5-2 PAGENO="0018" 1568 COMPETITIVE PROBLEMS IN THE DRUG INDUSTRY a formulary. I think it is required now because of the rules of medicare and reimbursement by the Government in that area, that the products be on a formulary in order for it to be reimbursed by the Government. So every hospital has a formulary. I think that there is no problem now on the part of any hospital, in developing a formulary, or borrowing somebody else's if they are so small they have not the time or the know how to get one of their own individually developed. I think most every hospital, whatever its na- ture or size, a. 50 or 100 bed hospital, develops its own, on its own local requirements, from. its own medical committee, pharmacists, doctors, technicians working together. This is a very important development in medicine right now, no question about it. Senator NELSON. I do not even have a good guess as to what the ef- fect will be, but it does seem to me that there is going to be a dramatic effect in several ways on the prescribing of drugs and the price of drugs when the time arrives that practically every hospital in America is using a formu~ary, and when practically every hospital is taking com- petitive bids, and when all the bids are requested on a generic basis. Mr. SQUIBB. That is right. Senator NELSON. I do not know what effect this is going to have, but it seems to me, that the use of formularies will have some dramatic effects in the field of prescribing of drugs by generic name, even though the physician may identify the company. And it seems to me, when formularies are widely available, and when the Government is more deeply involved and concerned about costs because of medicare and medicaid and when all the information is available to develop good formularies, that some kind of dramatic changes are going to occur. I do not know how to evaluate it, but I would like your comments on it. Mr. SQUIBB. You have right in the rec.ord of this hearing the testi- mnony of Dr. Williams, I believe his name is, from Grady Hospital of Emory University in Atlanta, in which he describes for you his own experience in developing a formulary there, and the extraordinary results it has had on his drug bill; and not only on the drug bill, the price paid for drugs themselves, but also on the way drugs were prescribed in that particular hospital. I think this is an extraordinary clear exposition of what happens when a large hospital, well run, with high quality standards without any question, sets itself into the for- rnulary problem. It is an extraordinarily clear example of how this develops. Senator NELSON. I would like your comments on this. I also meant to add that I assume if the medical profession is to improve the way it strives to, it will not be many years before almost every practicing physician will necessarily be affiliated with a hospital. So when formu- lanes are available, competitive bidding is the rule, and differentials in prices are visible to the doctor who sees what the hospital pays and then sees what his patients are paying in the retail market, there are going to be some tremendous pressures developing not only within the medical profession, but also within the hospitals, at the govermimental level, and with the public itself for changes in pricing practices. Do you think that is the case? Mr. SQTJIBB. There is no question about it. This is exactly what the pharmaceutical industry does not appear at this moment to fully ap- PAGENO="0019" COMPETITIVE PROBLEMS IN THE DRUG INDUSTRY 1569 preciate what the effects will be of this development in this noncon- trolled area, where there is price competition and where there will con- tinue to be more and harder price competition as the years go on. This will have a major effect on the phariñaceutical industry, its pricing policies and the operational practices that it can afford to pay for and that even that are advisable to pay for. This is the major changing trend in the medicine business, to use the word generally, in the United States. You have put your finger on it right there. Senator NELSON. And, I take it from your testimony, that you feel that the industry is not creatively undertaking to resolve this prob- lem at this time? Mr. SQUIBB. No, no they are not, because they are riding along with their profits and their procedures geared to their major area of opera- tions, which is at present the return from the retail pharmacist at a rigid price level. This has given them in the past all they needed and more than they needed, and although there are rumblings and, of course, discussions, they do not seem to see and certainly they do not seem to see in their publicly issued statements, the problem that I see and perhaps others see in this area, the problem for them which is developing. Senator HATFIELD. Mr. Chairman. Speaking of these formularies, Mr. Squibb, recognizing that these people who are going to be gov- erning the establishment of formularies in hospitals around the coun- try, are some of the same people if not exactly the same people, who have been siThj ected to your industry's influence through previous con- tacts, can you tell me what kind of objective criteria will be used to e~t~bli~h these formularies? Are they not going to be actually estab- lished by the same people whom we are talking about now? Are they going to somehow; in some transformation, come into a new under- standing of these competitive drugs? How are we going to know that we have some sort of objective criteria upon which these formularies are based, if we are dealing with primarily the same people we are falking about today? Mr. SQUIBB. Of course, you are and they will be making th~ir selec- tions. For example, the Grady Hospital, as spelled out in your hearings here by Dr. Williams, has to use a diuretic. They do not want to buy large quantities of all 15 or 20 diuretic tablets that are on the market. They have to have one. Their committee, which is made up of physi- cians in private practice in Atlanta, and in hospital practice at Grady, makes a selection, an arbitrary selection if you will, after due consid- eration. How do they do it? Senator HATFIELD. Arbitrary? Mr. SQUIBB. Well, they have to decide on one out of 15. It is arbi- trary. I say it is arbitrary because when it gets right down to it they have to cross out 14 and leave one. Now, what goes on in their minds to pick the one that they leave there is, of course, interesting. Certainly, the decision is subject to the background they have had in experience with the drug in its advertising and presentation to them. We find in different formularies around the country in different states, different diuretics, to just take one product, for example. Different products are selected in different areas. In other areas this Grady Hospital choice is not selected. PAGENO="0020" 1570 COMPETITIVE PROBLEMS IN THE DRUG INDUSTRY Senator HATFIELD. So, there is no real objective criteria that is un- related to the individual physician's own background, influence, ex- perience, exposures and pressures? Mr. SQUIBB. No. I do not think you could ever have a physician ar- rive at a position of prominence without having some kind of a back- ground in which he has been subjected to varying degrees of what you are talking about. Senator HATFIELD. On this matter of prices which you said might become one of the most important of all criteria, what is the differential between the percentages of profit that the industry makes between your three major groups to whom you distribute, the hospitals, the retail trade, and the wholesale trade? You say here that the over- whelming percentage of profit of the industry comes from the retail trade. What is that overwhelming percentage in relation to these two other maj or distribution groups? Mr. SQUIBB. I am deducing that from background rather than from any specific figure, because still the private prescription is the major source of outgo of the industry's products. Most of these prescriptions still go through the retail pharmacy, so they are at the high price levels. Therefore, just from deduction most of the profit comes from there. Senator HATFIELD. 70 percent, 60 percent? Mr. SQUIBB. Oh, I would say 70 percent. Senator HATFIELD. 70 percent, or 80 percent? Mr. SQUIBB. It varies for certain products. There are some products that are hospital products, of course, but I would say basically it is 75 percent. Senator HATFIELD. How does it divide as between wholesale and hospitals? Mr. SQUIBB. Well, I am putting the wholesaler into the retailer activity, because the wholesaler is merely a middleman with a corn- mission in the same rigid market. Senator HATFIELD. But, can you divide that for me between the retailer and wholesaler? Mr. SQUIBB. Well, it depends on the company. Some companies do not have any wholesaler business at all. Others have nothing but wholesale. Senator HATFIELD. How about Squibb? Mr. SQUIBB. Squibb has some wholesale business. Squibb splits, say, around 30 percent wholesale business. Senator HATFIELD. So, you could then take a ratio of that 30 percent within the 70? And in the 30 percent of the total overall volume of profit will be from hospitals. Mr. SQUIBB. From hospitals, although some of the wholesale business, if you have wholesale business, goes to hospitals, although most hos- pitals prefer to buy direct. Senator HATFIELD. Most do? Mr. SQUIBB. Most do. They prefer it, if they can. Senator HATFIELD. So, the overwhelming percentage of profit then would be wholesaler and retailer combined and would be about 70 percent total profit? Mr. SQuIBB. I would think so, or more. Senator HATFIELD. Or more? PAGENO="0021" COMPETITIVE PROBLEMS IN THE DRUG INDUSTRY 1571 Mr. SQUIBB. Or more. Senator HATFIELD. Do you know specifically what the Squibb per- centages are? Mr. SQUIBB. No; I do not know. It is not really a matter of interest in that category, although it should be. I think if they studied it more specifically, particularly the growing effect of their hospital business, they would see that the business is growing here, and that they had better watch out and prepare ourselves more firmly pricewise and procedural wise for this changing trend. Senator HATFIELD. What is the trend of profits from drug sales to hospitals today, up or down. Mr. SQUIBB. Up, mostly going up because the volume is going up. They have to be going up. The volume is going up there very sharply. Senator HATFIELD. I am speaking now of profits from the sales to hospitals from the drug companies. Mr. SQUIBB. As the volume goes up the profits would go up. Senator NELSON. As concerns the formulary, I would just like to read into the record the comments of Dr. Williams from Grady Hos- pital, when he testified before the committee about some of the very kind of problems that Senator Hatfield has raised. He says: It was not easy in the beginning. As you may imagine, the medical profession is conservative. This- lie is referring to the adoption of the formulary- this was a radical departure of performance at Grady Hospital. There were com- plaints that the committee was trying to dictate the type of medicine practiced at the hospital, that Grady Hospital patients would be poisoned by cheap inferior drugs, that the change from one color pill to another would upset the Grady Hospital patients in an irremedial manner. That we should buy trade name expensive items to support research done by the large drug companies, and that the committee was attacking the American free enterprise system. We persisted with the support from a large part of the faculty and the total support of the administration of the hospital who were interested in cutting this enormous drug bill. We have been fortunate. Estimated savings during the first year ran as high as $150,000 on a `budget of $480,000 for drug purchases. Now, on the question of how a formulary is developed, it is correct, is it not, that in a good hospital with all the specialties represented the therapeutics committee is composed of the urologist and the heart specialist and the various internists and pharmacologists and pharma- cists, and that based upon experience from medical literature, upon their own experience, they come to a conclusion about what ought to be included in the formulary. When they concluded from clinical evidence that several drugs of the same compound appear to be of equivalent clinical value, then the pressure, of course, is to select the one that is the most clinical. Is not that about the way a formulary is developed? Mr. SQUIBB. I think that would certainly enter into the picture. `F he pharmacist is generally charged with the responsibility of bring- ing in the price background of the different products, and at that juncture, price would certainly be a determining factor, all other things being equal. Senator HATFIELD. But would you not agree, Mr. Squibb, that it is not purely a clinical basis upon which they establish tha formulary? It is, as the chairman has very clearly pointed out, based upon the experience and the involvements of that physician- PAGENO="0022" 1572 COMPETITIVE PROBLEMS IN THE DRUG INDUSTRY Mr. SQUIBB. Oh, yes. Senator HATFIELD (continuing). Serving as a member of that committee? Mr. SQUIBB. That is right. Senator HATFIELD. And his background having been influenced by your industry? Mr. SQUIBB. That is right, certainly. We even see the committee of physicians in a given hospital take turns, rotate in this responsibility over the years, and as a result you see the products on a formulary changed just as the physicians change in this committee. This is evidence of what you are talking about. I think we have probably covered the pharmaceutical pricing back- ground. I would like to raise or point out something here which I think has not gotten into the testimony yet too clearly, and that is on page 8, the top paragraph there. It is significant to note that the industry's pricing policies and practices have not only brought on sharp criticism from those in academic, press, and legislative circles outside the pharmacy and medical professions, but also from those within the professions. The drug press, pharmacy convention pro- ceedings, records of legislative hearings, all are full of complaints of discriminatory pricing on the part of manufacturers from retail pharmacists, wholesalers, their trade associations; from puzzled physicians, purchasing agents, and institutional personnel both pri- vate and govermnent-employed who are faced with responsibilities to maintain and dispense inventories of drugs. Friend and foe alike of the industry point with all sorts of degrees of alarm, shame, clis- gust, distrust, and perplexity at what is certainly a situation at best hard to justify, and at worst completely unreasonable, chaotic as well as shortsighted and stupid. The criticism of drug pricing is even more bitter within the in- clustry, within its customer framework, than it is in the full light of testimony in hearings such as this. Industry is struggling with this, but it does not seem to come to grips with it. Now, let us see what can be done about pharmaceutical pricing, the public conception of which threatens a whole industry, and which couldi lead if not properly handled, to irretrievable steps of retribution andi regulation which very wEll might have serious implications for the whole pattern of medical care in the United States-and not neces- sarily for its improvement. It is important that public confidence in the fairness of drug prices be restored, and that in so doing the future of an important industry's contribution to medical progress not be limited or proscribed to absolutely no one's benefit. First, how can the industry by itself revise its price structure and restore some sanity into its marketing programs? It must be im- mediately realized that the concern here is with prices, and this being so, a concerted, cooperative approach by pharmaceutical companies is impossible, antitrust and conspiracy concepts being what they are. Actions must be taken audi solutions found, company by individual company working within its own framework of costs, prices, and profits-not to mention its own competitive standing in a bitterly contested market. In the first place and by far the most important of all, pharmaceu- tical management must understandi the nature of the social climate PAGENO="0023" COMPETITIVE PROBLEMS IN THE DRUG INDUSTRY 1573 relative to medical care in which it now lives and works. This is not the climate of 10 years ago, or even of last year. It is a moving, expanding, developing concept of prepaid or guaranteed protection from the fi- nancial impact of illness as a basic personal requirement with the added growing idea of a real, affirmative responsibility resting in the public health authorities-State and Federal-for establishing and maintaining conditions leading to optimum health for all. There is no longer any question that it is the duty of government to protect the public from itself as well as to provide for the means to insure sound health in every physical and 1nental aspect of individual well-being. The public wants the government, State and National, to pay for most of this, and for what it does not actually pay for, the public expects its governments to assume roles of stern regulators, overseers, and active supervisors. It is precisely because the public looks for and even demands these services and responsibilities from its elected officials that all of those privately engaged in medical care in any form must exercise extreme caution to conform their activities-particularly financial-to what are considered normal, fair, and reasonable-such judgments to be made and construed within a very narrow range of acceptable conduct and its results. Definitely the pharmaceutical industry must accept special responsibility for its operations, and special burdens and limi- tations not regularly assumed by other industries because of its place in the scheme of medical care now an object of particular public con- cern. It must adapt itself at the very least to the standards of normal economic and industrial practices and profits, and be able to live with them and at the same time continue to make the contributions and perform `the services that it has pointed to so proudly in the past. Probably it must go even further than this, and accept the idea that its profits should be less than those of average comparable industries in recognition of this special social responsibility which is being firmly and irrevocably forced upon it by the general public. Senator HATFIELD. Mr. Squibb, would you identify for me the aver- age comparable industries? You are speaking here of the social respon- sibility of the drug industry? Mr. SQUIBB. Yes. Senator HATFIELD. And, the relation of that to its profits, that they should be less than those of average comparable industries? Mr. SQUIBB. In `the testimony that has been given before the com- mittee, the industry has been compared in its innovative fUnction with industries such as office equipment and other electronic industries on the one hand and with the textile industries and the food industries on the other hand, which are rigid or noninnovative industries. What 1 am really saying here is that you have to take an average return on investment by the representative industries that cut across a broad seg- ment of public life, not by special industries that do not affect a mass of people. Senator HATFIELD. You mean like food? Mr. SQUIBB. Like food, definitely food. Senator HATFIELD. Clothing. Mr. SQUIBB. And clothing, that is right, that type of an industry that everybody or most everybody in the country contributes to by some sort of purchase. PAGENO="0024" 1574 COMPETITIVE PROBLEMS IN THE DRUG INDUSTRY Senator HATFIELD. Tobacco? Mr. SQUIBB. Tobacco, yes. I should think tobacco is now an essential, that is right. Senator HATFIELD. Then you would apply, if I understand you correctly, a double standard in a sense, to the drug industry and to these so-called comparable industries as it relates to their profits? Mr. SQUIBB. Yes, a double standard for the pharmaceutical industry, because of its nature, a special standard. Senator HATFIELD. I see. And in relation to this whole matter of picking out these so-called comparable industries, and, of course, they would no longer be comparable, would- Mr. SQUIBB. No. What I am saying when I talk about comparable, I mean an industry whose products are widely used by the consuming public as are pharmaceuticals. Senator HATFIELD. This is because of the special and unique social responsibilities I understand that you feel the drug industry has. Mr. SQUIBB. Tha.t is exactly it. Senator HATFIELD. And as it relates to this protecting of the public against itself, you have led into this particular point, what about the tobacco industry? As a former member of the National Cancer Society board and as one who is familiar somewhat with these reports made by our Surgeon General, do you not agree that it does threaten the health and well-being of great numbers of American people? Mr. SQUIBB. Well, never having smoked a. cigarette in my life and not a participant in, or supporter of, the tobacco industry I just know what I read in the papers on this. I think that tobacco, how- ever, in spite of those statements, is a necessary fact of life to millions of people, and as such is a very important consumer industry at the moment. It may be hurting people and it may not be, I do not know, but- Senator HATFIELD. But there is a social responsibility, as you point out here, on the part of the drug industry based upon its relation to the health of the people. Mr. SQUIBB. Excuse me, I see what you mean. Yes, if a product made by the tobacco industry, or a product made by any industry, even an automobile has its damaging functions. Any one who makes an instrument that can damage people has some degree of social responsibility, as has been recently expressed. But I think that the drug business far transcends any of these, either tobacco, or any other, because it is given at a time for purposes which are unique and desperately involve life itself in some cases. Senator HATFIELD. Even though we have statistical evidence a~bout lun~ cancer in relation to cigarette smoking? Mr. SQUIBB. Yes. I am not too well informed about that, but even e'oing along with you, although I really do not know about the harm- ful nature of tobacco, but if it is in fact harmful, then the industry has a very heavy social responsibility. Senator HATFIELD. How do we protect the public against itself when we determine there is a sOcial responsibility by a particular industry that is not being upheld, merely regulating profits? ~~:rr. SQUIBB. In a lot more than profits. Senator HATFIELD. What would you suggest further? PAGENO="0025" COMPETITIVE PROBLEMS IN THE DRUG INDUSTRY 1575 Mr. SQUIBB. Well, you have to regulate the whole performance of an industry that is engaged in this type of activity; quality, its prac- tices of distribution, its practices of promotion, presentation of its products. Senator NELSON. Theoretically, `as to quality, with a sufficient num- ber of personnel to inspect, the present required standards are de- signed to insure quality, although these standards are not always complied with as we know from drug recalls and various other things. But there is legally at least a quality standard established for this industry. This is true maybe of some other industries, too. Mr. SQUIBB. Oh, yes. Every industry has its quality standards. Senator NELSON. I take it from what you were saying about profits that you are simply setting some limits. You think the pricing prac- tices as now exist ought to be reformed, and as you have stated, at the very least, the standards for pricing should :be set to accord with the standards of normal economic and industrial pricing practices and profits. Your other sentence is: "Probably it must go further." Mr. SQUIBB. Yes. Senator NELSON. I am assuming what you are saying is that the industry at least must conform to noEmal practices and might even have to go further than that in some cases. Mr. SQUIBB. Perhaps go further than that, because it is so urgently under the eye of people who are concerned with health care problems today. Senator NELSON. As to your comment about social responsibility, I think it is interesting to read a statement from page 139 of the re- port of the Committee on the Judiciary of the U.S. Senate, June 27, 1961: In industry generally, the views expressed by Jefferson and Franklin with respect to patent monopolies appeared to fade rapidly, but for a time they con- tinued to prevail in the drug industry, reflecting a recognition of the peculiar and unique relationship of this industry to the public health. In 1854 when Dr. E. R. Squibb, founder of E. R. Squibb & Son, managed to distill for the first time pure ether of uniform strength, he declined to take out patents. Instead lie published his discovery in September of 1856 issue of the American Journal of Pharmacy. The essential difference between most other countries and the United States on this matter is that the views held by Dr. Squibb have continued to prevail abroad but have long since been abandoned here. But at least it was a matter of discussion and concern over 100 years ago when your great grandfather founded your distinguished company. Mr. SQUIBB. That is right, and I think it always will be. It always will be from the very nature of the problem we are dealing with. I would like to continue on page 11 and point out this area here which I think adds a little bit to what we are talking about right now. A pharmaceutical company management-now enjoying, let us say, a 10-percent return on sales after taxes-sets its objective for the next year at 12 percent-and then 15 and even 18 percent over a 5- or 10-year plan. Often in the past this type of profit objective has been obtained through new products and intensive marketing programs. Recently, such growth has been more exceptional lit actuality, although it still is quite possible. However, pharmaceutical firms still project and plan for steady growth in profits, as does most industry. These pressures PAGENO="0026" 1576 COMPETITIVE PROBLEMS IN THE DRUG INDUSTRY for better profits, quite normal and accepted elsewhere as good busi- ness practice, take on peculiar overtones when sales of essential health care products are involved. The consent of more and more profits from the miseries of the sick, the aged, and the malnourished segment of the population seems to run counter to the swelling trend toward State- supported medicine. If the Government takes over the responsibility of ameliorating or improving the health of a large part of the Nation, this should not result in enriching beyond normal levels at the very most those who participate in the job. Expenditures for medicine by tax-supported Government agencies bring with them strong pressures, both sociological and economic, for the avoidance of profiteering, waste, and misappropriation. The latter two pitfalls are the responsibility of those who administer the program-the first is the prime responsibility of the supplier, in this instance the pharmaceutical manufacturer. Until and unless he sees this clearly, the threat of Government-im- posed profit regulations will always be imminent. Profit objectives must be made subordinate to productivity objectives. Senator NELSON. Some of the testimony we have had from the indus- try, and other statements made by members of the industry and the medical profession have asserted that in order `to attract capital for the purposes of expansion and for research, these higher profit rates are necessary because of the great risk and so forth. Do you have any comment. about that? Mr. SQuIBB. I commented on risk in my earlier paper. I do not feel that the pharmaceutical industry itself, under the definition of the word "risk," as I understand it, is a risky business. It is a profitable business, one in which most of the large companies have been long en- gaged a.nd have many years of profit history behind them. While it is true that a given product in their structure may come and go and may often be relatively short-lived, their financial risk, it seems to me, in my experience, has been very little in the pharmaceutical business. A well-managed pharmaceutical company is as good an investment1 as good a place to put your hard-earned dollars, as safe a spot as any I know of. I do not `think `that there is a risk involved here in terms of cvha't is commonly `thought of `as a risky speculative sort of a business. The drug business is not. Senator NELSON. I have one other question I intended to ask some time back. Although it is not directly related to the context of your present statement, I would like to ask it here. The committee staff has gathered statistics on prices of `a number `of drug products manufactured in this country and sold in foreign coun- tries. We have found by checking through the State Department and the embassies in a number `of European countries as well `as Latin America. and others that it is quite common for `the price charged in the marketplace here to be anywhere from 50 percent higher to twice as high or even 10 times `as high as a price charged by `the same corn- pany for the same product in Paris, London, Berne, `Switzerland, Rome and &sewhere. I have not been able `to get an explanation that `is satisfactory to me as to why that clifferential-why the manufacturer of `a drug here, ships it overseas, and charges a lower price `abroad. What justification is `there for charging a lower price to people in foreign countries than char~ed to our own nationals? PAGENO="0027" COMPETITIVE PROBLEMS IN THE DRUG INDUSTRY 1577 Mr. SQUIBB. I think it is very simple. The justification for this lower price you are referring to is simply the fact that that is the price, lower or higher or whatever it may be with which the manufarturer can get into that particular market. He could not sell the product at the price that he is selling it here. He could not sell it in adequate quantities. He could not make an entry in the market at this level. So, I think that the pric.e that is established in any given country is the price that the market there will bear, and return him some ade- quate sales. That is the judgment decision that is made when a product is marketed in any country. It is as simple as that. I do not think there is any other explanation. There cannot be any other explanation. Senator NELSON. Is there a.ny difference between the marketing and the advertising of drugs, in foreign countries and here? For example, if doctors are convinced in Switzerland that this very fine miracle drug called by a certain brand name is going to help their patients, why cannot the firm establish the same high price that brand names com- mand in our own market even where the same drug is available for far less? Mr. SQUIBB. I think they probably could if they wanted to put the promotion money, the selling and advertising effort in establishing a trade name in Switzerland, to use your example. But the money is not there, the time is not there, the people are not there to do that, and the overall market, the Swiss market or the French market or what- ever market you are talking about, the total market is probably not large enough to justify a great deal of~ money and effort in this direc- tion. Senator NELSON. Is there any difference in the laws applicable to the advertising of drugs in foreign countries? Or, is there any differ- ence in the traditional practice of the medical profession in those countries that might account for this differential? Mr. SQUIBB. I think this is part of it. Promotion and advertising in some countries is limited by law and, restricted as to distribution of samples, distribution of free materials. The media, that is the medical journal media that are available differ widely. The actual procedures of medicine in terms of who buys the product and who `ays for it differs from country to country, and, of course, basically the size of the market itself differs. Some of the smaller countries have less sophisticated developments in this area. But, basically the price level is predetermined. If you decide to go into the Algerian market, you look at the price you are going to have to charge for your product in that market to get any sales at all, and that is the price you set, irrespective of what your priëe may be here. Senator NELSON. Is there any question in your mind as to whether the firms make a profit in these foreign countries, where they charge one-third, one-fourth, one-fifth a.s much as they do in the marketplace here? Mr. SQUIBB. No, because of w-hat we have already talked about, the basic cost of the product is still low in comparison with the price. So, it gives you a wide range of opportunity here. Mr. GORDON. Mr. Squibb, aren't you really saying that prices are low overseas because of competition, and that they are high here because of the lack of competition? ii\'Ir. SQUIBB. No. PAGENO="0028" 1578 COMPETITIVE PROBLEMS IX THE DRUG INDUSTRY Mr. GORDoN. You are not saying that? Mr. SQUIBB. No, I am not saying that. It could be or could not be. It might or might not be the case, but I am definitely not saying that they are low because of competition. There may be competition in terms of what they are thinking about in a given foreign country, and there is certainly competition here. Competition is engendered at the high level by trademark promotion and competing for the doctor's attention, which we are doing here, and spending money to do it. Mr. GORDON. I am talking about. price competition now. Mr. SQUIBB. You can set your price high if you put on your strong sales promotion, and produce large volume. Mr. GORDON. But, that kills competition. Why do you have sales pro- motion? In order to try to differentiate your product, to show that it is a different product from anyone else's. Mr. SQUIBB. That is right. I think we are talking about competition in a different framework. You are right. What you are talking about, our efforts here is to kill competition by taking all of the market your- self. That is true. Mr. GORDON. Now, coming back to the risk business, how many PMA members have gone out of business in the last 10 or 15 years? Do you know of any? Mr. SQUDiB. I do not know of any. Mr. GORDON. Can you tell us how many times during, say the last 10 years, have the profits of the large PMA members fallen below the average for the. industry as a whole? Mr. SQUIBB. Well, the studies I have seen in this connection have always indicated that the pharmaceutical industry represented by these top producers has been higher than average, and among the top one or two industry groups. They have always been in the last 10 years well above what might be considered average or normal return. Mr. GoRDoN. Can you tell us how the drugindustry might be affected by cyclical factors, say depressions? Mr. SQUIBB. Well, certainly in the past, the drug industry has only been slightly affected by this, which can be expected because purchases for medicines are probably the last thing that you cut out when you are hard up. Also in good times you do not go out, rush around and buy a lot of unneeded medicines. But I think that there is going to be even less such effects in the future as the Government steps in buying more and more medicine and reimbursing more and more of the medical bill. Psychological forces are going to have even a lesser effect. They will have very little effect. Mr. GORDON. Now, as to product obsolescence, are there many prod- ucts which have been losses to the drug firms? What is the extent of the losses? Mr. SQUIBB. Well, there are many products which are on a down- ward sales curve, and there are many products at a given moment on any given sales list, I believe, which are losers in themselves and by themselves. They have been winners in the past and now they are losers. Mr. GORDON. Why do they retain them? Mr. SQUIBB. Well, I think that one of your witnesses touched on this very well. He said sales departments find it hard to let these things go. PAGENO="0029" COMPETITIVE PROBLEMS IN THE DRUG INDUSTRY 1579 I think that they do tend to hang on because there are a few physicians who still want to use them and they want to keep, those physician's goodwill. Eventually as these losers grow to a bigger part of the pharmaceutical companies list they are cut out and disappear. Once in a while there is a big washing out of losers. Mr. GORDON. How great is this? What is its extent? Mr. SQUIBB. Every year it is a continuing problem, but it cannot be great in terms of percentage of the total, because the loss is not great enough to damage your profit picture. Any company that lets its losers get above the 5-percent sales level is in certainly bad management trouble. Senator NELSON. If there is no objection on the part of the witness, we will take a leisurely lunch and come back in 30 minutes. Mr. SQUIBB. All right. Senator NELSON. We may have some voting this afternoon. I would rather break now and start again in 30 minutes if you do not have any objection. Mr. SQUIBB. None at all. Thank you, Senator. (Whereupon, at 11 :45 a.m., the hearing was recessed, to reconvene at 12:15 p.m., this day.) AFTERNOON SESSION Senator NELSON. The hearing of the subcommittee will resume. Mr. Squibb, I believe you were on about page 11, is that correct? STATEMENT OF GEORG~E S. SQUIBB, FORMER VICE PRESIDENT, E. R. SQUIBB & SONS-Resumed Mr. SQUIBB. I was going to pick up at the top of page 12, Senator. Senator NELSON. Go ahead. Mr. SQUIBB. As another necessary step to pricing policy reform- the price of the pharmaceutical product should become a function of its cost-a function approximately the magnitude of those in major con- sumer goods industries and varying only in terms of the sales. burden imposed. Prices which are set at nine or 12 or 20 times factory cost only serve as incentives for public attack on the various means and methods that permit such action, patent protection, product monopoly, success- ful sales pressures, or whatever. Exploitation of the value factor of medicines used in life preserving and lifesaving situations, by setting prices far above the cost is what must be deliberately and conscien- tiously avoided, no matter what justification or economic temptation is felt by the manufacturer. For example, if a new drug will largely depopulate State mental institutions by permitting home care of the patients, all of the extraordinary economic gain from this product must be felt by the State, and none by the pharmaceutical company that supplies it. Ordinary profits, yes, but windfalls, no. This may come as a shocking idea to those who set the prices on major volume pharma- ceutical products which are used for treatment of patients in tax- supported situations but it is an idea which must be accepted, or it will be imposed by regulation. As the scope of purchase of medicine by government widens, the idea has greater implication for an ever larger part of drug product distribution and sale. PAGENO="0030" 1580 COMPETITIVE PROBLEMS IN THE DRUG INDUSTRY A product that does the job that modern drugs often do, affords under any comparative value approach, a most unusual temptation for the pricer to set his figure much higher than might otherwise be the case. Classic theories of price established to get a share of a competitive market are applicable only to a very slight degree. In fact, it often might be said that each drug product can be established in its own market by skillful promotion and exploitation of its own particular virtues. Not only has the pharmaceutical industry been successful in maintaining the conviction with many physicians and buyers that not all drugs are alike, but it has even succeeded in persuading them that all products are different, which is a much more effective argument from a sales point of view. Leaving aside at this point the validity of this claim, the mere fact that it has been frequently and effectively established, and continues to be, even under the conditions of current ôontroversy and attack, gives the pricer a unique opportunity to set his figure without relation to any factor except what he believes the market can bear. He, of course, will take into consideration, in a ge1~erah way, the existence and success of comparable products, or products used for the same therapeutic or diagnostic purposes, but not by defini- tion any product exactly like the one to be priced because there is none. It is well to remember here that differences for pricing purposes arise iiot only from varying chemical or molecular. structures, but from differing manufacturing sources or brand names and sales programs as well. Prices set in this way on prescription products lead to the very situation that is so often criticized today; an enormous range between trade or retail prices and those given to institutional or government purchasers. The rigidities of prices at the trade level already de- scribed tend to freeze the prices a.t the top level for at least a con- siderable pei~iod of time during which the disruption of the institu- ti onal pricing structure occurs, and the perplexing and dismaying situation now deplored by retail pharmacists, legislators, and the general public arise. The precise and detailed approach that a pharmaceutical pricer should use to avoid falling into the trap of an eventual "no-price" for his product presents a complex and difficult problem. It amounts to the better utilization of forecasting, and market research tools- including computerization of much basic information-to establish in advance the market utilization of each product and package of product, and the pricing prerequisites of each market segment con- tributing to the product's sales volume. Varying competitive forces and their effect on prices in each market must be recognized and evaluated in advance so determination of product dollar productivity can be measured in realistic terms. By this means, formulas for pric- ing can be set that do not lead in fact and theory-by actual use or by company policy-to anything more than a reasonable price range for a given product. It must be stressed that in setting up any such ap- proach to pricing that due attention must be given to the new hard facts of prescription product distribution-facts which the outside world is imposing on an industry which is reluctant to believe and acknowledge them and even more reluctant to change their policies accordingly. Some such facts are: PAGENO="0031" COMPETITIVE PROBLEMS IN THE DRUG INDUSTRY 1581 (1) That hospital and institutional volume of drug usage will be substantially up in the immediate future. Some of this volume is now due to more patient coverage through government and private insur- ance, but some of it will come out of the retail pharmacy as more niedicines are dispensed in association with hospital and other msti- tutional treatment. As a result, hospital bulk packages of drugs will now be used on outpatient prescription to a substantial degree in~ direct price competition with the retail pharmacist. (2) That the governments, State and Federal, which will now pay much of the Nation's drug bill, are going to insist on paying the gov- ernment price for its purchases-not the much higher retail price. (3) That the governments will see no reason to buy their drug requirements. any way except direct from the manufacturer, thus eliminating wholesalers, commission agents, and other middlemen from this part of the market. (4) The use of various means to permit substitution of brands for those specified, or to require prescribing in generic terms, will con- tinue to be urged by many influential persons and professional groups. This will weaken market positions and set up new price pressures for many products vulnerable to these trends. Facing up to this fact alone is a difficult job and a major challenge for the pharmaceutical pricer. Any new or revised approach to this job by the pharmaceutical pricer will, of course, bring a substantially altered pattern of income to his company. For this reason, this concept cannot be put into effect with immediate or all-inclusive coverage of any company's products. But, it can and should be adopted in new pricing, and also wherever changing circumstances permit on existing products on which there is need for one reason or another of pricing revision. Gradually this pricing policy should be implemented to bring Order from chaos, and public understanding and approval from bewilderment and hostility. Before closing this discussion of the pharmaceutical manufacturer's price schedule, two relatively minor aspects should be touched on- first, the so-called service products and, second, the use of special deals, temporary discounts, and other promotional activities that affect price. This argument is made frequently by pharmaceutical companies that prices are maintained high on some volume products in order to permit the marketing of items which return no volume or profit at all, but provide a needed service for physicians in treating rare ill- nesses or for use in very specialized circumstances. Marketing of prod- ucts at a loss does exist, and to the extent that money is found for these products from the earnings of others, the argument appears sound. However, the actual extent of such support is limited. If volume of the service items is small, both the cost involved and the out-of-pocket loss is small in terms of earnings elsewhere. Overall profits show this. Senator NELSON. Mr. Squibb, may I interrupt a moment ~ That buzzer is a rollcall. I have to go to the Senate. I will be about 10 minutes. We can recess or you could continue to read your statement. I have some questions that I will want to get back to, but we can proceed in whichever way you prefer. Mr. SQUIBB. I can wait for your return, if you wish. Senator NELSON. All right, I willbe back in 10 minutes. (A short recess was taken.) PAGENO="0032" 1582 COMPETITIVE PROBLEMS IN THE DRUG INDUSTRY Senator NELSON. Mr. Squibb, would you mind interrupting your presentation? I would like to ask you about the press release of the Pharmaceutical Manufacturers Association commenting on your statement. Have you had a cha.nce to read it? Mr. SQUIBB. Yes; I just received a copy a few minutes ago. Senator NELSON. Do you have any comment you would like to make about the release? Mr. SQUIBB. I think I look at this more in sorrow than in anger. Senator Nr~soN. I did not understand that. Mr. SQUIBB. I say I am looking at this more in sorrow than in anger. This, I think, just points up what I have said today. I certainly have not commented anywhere that I know of that the prices have not in fact declined. They have declined and they are still declining. That does not change the burden of the argument that the industry is thought to be still high priced. Of course, they have declined. Over on the second page, I certainly do not derogate research. Quite to the contrary. This points up the fact which I bring out later, that it is very difficult to talk rationally about research in the industry with- out somebody getting a little bit overexcited about the emotional im- pact of killer diseases and so forth. And I know of nowhere in the statement where I even mentioned the words "they are cutting equiva- lency of drugs." I do not even discuss this at all. This seems to me to be sort of setting up a straw man and then blowing it down with a lot of wind but it does not really cover the points that I have been trying to make here. Senator NELSON. On the second page it says: "As to the value of the research which Mr. Squibb derogates." In your statement you did not derogate research. Mr. SQUIBB. No, I did not derogate it at all. I indicated there might possibly be within the scope of industry research some which is not wholly productive or wholly important. Just this one slight suspicion or remark ca.st on research brings forth this extraordinarily strong answer which is the point I am making. It is very hard to stand off and objectively talk about research. It is almost a holy man in the industry. I think the research should be examined factually without derogation, without criticism, like every other part of the business of which it is an important part, one of many parts. Senator NELSON. I think that the statement rather dramatically emphasized the comments in the opening of your statement in which you say that the only answer that the pharmaceutical industry has given to date is to, "tell its story," to "improve its image", to "explain its contributions to medical knowledge and the health of the people." I think their statement dramatizes the cogency of your observation about their kind of response to any suggestions or criticism. Mr. SQUIBB. It says I speak for myself. Of course, I speak for my- self. I only hope some day that the words I am saying here will be adopted by t.he industry. That is my wish. Mr. GROSSMAN. Mr. Squibb, I wonder if I might ask you in the same area, do you feel that there is room for dissent in the industry and that there are others who feel as you do and perhaps do not come before us for any reason? Mr. SQUIBB. Oh, surely. PAGENO="0033" COMPETITIVE PROBLEMS IN THE DRUG INDUSTRY 1583 Mr. GROSSMAN. You do feel that? Mr. SQUIBB. I am certain of it, oh, yes. Mr. GROSSMAN. Do you feel that there are any undue pressures placed on any of these people? Mr. SQUIBB. No, I do not think there are undue pressures. It is a mat- ter of judgment, even accepting the accuracies or the truth and the basic reliability of the opinions I am expressing whether it is advisable in tactic to express them at this time. Mr. GROSSMAN. Are any of the dissenters in positions of leadership in their companies? Mr. SQUIBB. I think so, yes. Mr. GROSSMAN. Why have not any of these dissenters who are leaders taken- Mr. SQUIBB. Because I think they feel that it does not prove any- thing or do any good in developing the idea that they have to expose them in any type of a public hearing. Mr. GROSSMAN. For example, even without a public hearing, why, for example, have not they done some of the things that you suggest? Mr. SQUIBB. I think there are discussions and things underway in certain companies all along the lines that I suggest. Mr. GROSSMAN. Yet, there has not been much action. Mr. SQUIBB. There has not been much action yet. The pressures are mounting now and they are sharper than they have been for a large number of years or ever. I `think the pressures are right now on the industry for these reforms for obvious reasons. They are very heavy. I think that there is some careful reexamination company by com- pany going on of many of their pricing policies. I think that this is indicated by the testimony of several of the presidents of these companies you have had here. They say: "WTe are not going `to price it like this any more," and that is almost an exact quotation. Mr. Gnossi~rAN. I think that the heavy-handed nature of these press releases implies that it is going to be very difficult for `others to come up here. I think when `they say here in, the third paragraph that the PMA will present "testimony of distingu'ished economists of unim- peachable integrity and reputation," that certainly would seem to me to reflect that they migtht `think otherwise of you. Mr. SQUIBB. Perhaps they do, but `they do not say so, and it has been interesting to me on my previous statement that there has been agree- ment generally. No one h'as ever `said in `public that they disagree with what I have said. As a matter of fact, I d'o not see any disagreement here with what I have said, because I have not said anything along the lines that they outline here in this press release. They have just missed the point because they do not or will not see it. The fact that wholesale prices have d'eclined 14 percent in 17 years and prices of patent products declined and other products rose, I accept. I do not see what difference it makes to the point we are talking about, which is the public view of the business. If all these things have happened as `outlined in the release, it has not done the pharmaceutical industry any good with its legislative critics or its public critics. And `this is the pickle they are in t'od'ay. I am suggesting there is a way out of this pickle. It is not by simply repeating that price's are going down or' that all i's well. You know God is in his heaven and all is well with the world. 81-280-68-pt. 5-3 PAGENO="0034" 1584 COMPETITIVE PROBLEMS IN THE DRUG INDUSTRY Mr. GROSSMAN. I would just hope. that others of the leadership that you mentioned and I mentioned would be free to come up here and talk to us and let us know what they think, without this kind of press release staring them in the face. Mr. SQUIBB. That is t.heir problem. I do not think you are going to get very many here right now. Mr. GORDON. Mr. Squibb, isn't this a kind of intimidation that this is what anybody can expect from the industry if he comes up to testify along these lines; that PM1A is going to put out a press release attack- ing him? Mr. SQUIBB. I assume so. I am not intimidated and I do not think they figured I would be. But, I think it is too bad that they cannot try to make something good out of what I sa.y, even if only a crumb. I think we were on page 16, talking about the loss of the service products. A better answer to this situation is to try to price the losers more appropriately; they can carry a considerable premium because of the very rarity and urgency of their use, or to dispose of the product to some other firm which beca.use of more suitable facilities, lower over- head, or other advantages can better produce the item. In any case, it does not appear that the practice of providing the medical profession with a few under-priced service items is any real or significant burden on overall earnings even giving no credit at all to the "goodwill" or "good public image" created by the practice and exploited with en- thusiasm by public relations departments. The pharmaceutical industry has historically been partial to special deals of the most extraordinary nature on its products-deals, of course, offered openly to all competitive classes of customers. Perfect.ly legitimate deals as far as Robinson-Patma.n or antitrust are con- cerned, but cleaTs which are greatly disruptive of whatever price structure of a formal nature, and there is not much of it in any case, that may exist. For example, deals often go far beyond the one-free- with-a-dozen offer frequently found in case-good pricing, to such levels as five free with five, and even two free with one. It is not uncommon f or an important pharmaceutical product of a strongly competitive nature priced at, let us say, $10 to the retailer to be offe.red by the salesman for a limited period of time at one free with each purchased, bringing the cost to the retailer for that purchase to $5. The use of this material on a refill prescription for an original which was filled with $10 material opens up all sort of intriguing pricing questions. Diversion, overbuying, substitution, special sales pressuies for price alone-and many other questionable practices-are practically guar- anteed by this sort of pricing. And, behind it all lies the question, what is the real price of the product? Other types of dea.ls offered seasonally for "inventory protection," or as stimulation for salesmen's selling activities, or as devices to move potentially short-dated merchandise, or as lures for improved dealer cooperation and support, all may or may not accomplish the intended purpose, but each without question seriously confuses the price basis for the product and contributes importantly to the feeling that the original off-deal price is somehow inflated and not really necessary. Deals add greatly to the difficulty of outsiders in understanding pha.r- maceutica.l sales operations, and even have the unfortunate result of PAGENO="0035" COMPETITIVE PROBLEMS IN THE DRUG INDUSTRY 1585 adding to the burden of public suspicion unfairly. Sometimes a coin- pany has a published price for a product which is a fine target for "high-price" critics, but rarely, if ever, sells the item at this price because of the frequency and generosity of its deals on the product- deals which the retailers wait for with the certain anticipation of a bargain price. In the meantime, the published price remains for all to see and none to use, and as a perfect~ subject for critical analysis by industry's detractors. Manufacturers should take a new and hard look at their published retail prices compared with their average selling prices based on unit sales divided intO the dollars received. Deals are importantly disruptive of orderly pricing introducing an element of imcertainty into the market that often confuses the manufacturer as well as his customer. They have been: overused in fact, misunderstood in principle, and wrongly analysed in relation to their effect on profits often enough to make them an area for immediate and determined study by any management which believes in pricing refonn as a pre- requisite to the solutions of many of the industry's critical problems. Mr. GoiwoN. When the PMA brought up the problem of generic equivalency a couple of weeks ago, one of its witnesses presented data to indicate that its antibiotic produced higher blood levels than the products of two other manufacturers, perhaps three others, I do not recall. Now, in your opinion, how meaningful is this type of a study? Mr. SQUIBB. It is very meaningful if you are the manufacturer of one of those four products concerned. However, I think the generic equivalency argument which certainly has not yet been developed, is self-limiting, because you are working only with specific products, whether it be 10, 12, or 50, and you prove the equivalency of those prod- ucts. Outside of what you are testing, you have no data whatsoever, and the fact that product A differs or is equivalent to product B, C. and D, or vice versa, does not cover F, G, H, I, and J at all. And in this vast field, whether you are coming in as a: supplier of product A or product D or product F makes a big difference. The situation varies from where you approach it. Mr. GORDON. You would say that no broad generalizations can be made? Mr. SQUIBB. You cannot make generalizations at all. These are specific examinations of specific products for a specific purpose, and you cannot deduce a result in another area because product A and B are different. You cannot say that product A and F are different unless in fact you test product A and F, and so forth. Senator NELSON. Let me inquire about that a moment. The asser- tion continues to be made by some representatives of industry that there is no such thing as generic equivalency. If that statement is to be meaningful at all, it seems to me, it can be argued technically that no two drugs are identical because no two things really are identical. But is it not correct :that the assumption upon which formularies are established in this country is that using the clinical information that a therapeutics committee has, they can select drugs for their formulary, and although there may be four or five brands, brand names or generics, that purport to do the same thing and be the same compound, based upon their experience the formulary committee selects one. PAGENO="0036" 1586 COMPETITIVE PROBLEMS IN THE DRUG INDUSTRY Mr. SQUIBB. That is right. Senator ~ELSOX. And the same day across the Nation another clis- tlnguished forniulary committee selects another company, and on the same day another formulary committee in another part of the Nation selects another. I notice in the Medical Letter of June 2, 1967, they tested 22 preclmsones, and concluded, on the basis of clinical advice and chemical testing, that all of them were equivalent. So, if they are correct, you could have 22 hospitals using 22 pred- nisones, and the assumption all these hospitals are going on is that the one they select is going to produce as good a result as any of the other 21. They then have a chance to test it in the hospital, and if they get the kind of result, the successful result they are seeking, they have a chance to compare the result they are getting with their drug as against another formulary in another part of the country. I conclude from that that there are many drug products made of the same compound by different companies that are, for all practical purposes, therapeutically equivalent in the result they produce. Do you have a comment on that? Mr. SQUIBB. I think you are absolutely right. You have to arrive in the process of making a decision to put one item in inventory that this product has an equivalency. To say it is in fact identical to every other prothict without testing is making an assumption that I do not think has to be made. I do not think it is made by any responsible formulary committee. You outlined it very well. They are assuming that this will do the job for which they want this type of product in their formulary. They made a selection because they have to make a selection, and they make it on the best information that they can get, whether they have to depend upon somebody else's analysis in the Medical Letter or by outside laboratory testing or whether they can make their own analysis. Whether they are simply doing it from the basis of t.heir experience in the practice of medicine and the fact that they have used a given label or a given product or for some other reason, I think that we are beating around too much the question that there might be or might not be identical products. Nothing is identi- cal, as you say. I think that the equivalency has to be determined by responsible men, because they have to in fact, for practical reasons, select one product. They do that, and price is becoming now a factor in this more than it used to be. I think that testimony in hearings such as this, conversations and discussions in the press, and in medical circles all around the country are beginning to stir up this issue more than it has ever been stirred up before. Senator NELSON. Thank you. Mr. SQUIBB. I will proceed on page 18. So much for pricing- Senator NELSON. Are you saying that the cost of the detail man is much greater than all of the other costs of promotion, including ad- vertising in the medical journals, free samples, and all other aspects of legitimate promotion of the drug? Mr. SQUIBB. This varies company by company. Obviously some com- panies believe in large detail forces and others manage to get along with relatively few. But basically speaking I think that in the case of any house that hires detail men, the expense for detail men and their equipment is the largest single factor in their sales budget. If PAGENO="0037" COMPETITIVE PROBLEMS IN THE DRUG INDUSTRY 1587 you added all the other many factors tOgether you may come up with an amount equivalent to the detail man, but this is the largest and most important single item in many budgets. Senator NELSON. How significant is the detail man's role in the sales end? Mr. SQUIBB. Very significant. The detail man is the best and most effective means of producing a sale or prescription on your specialty, no question about it. He can be made much better and that is one of the points I make here, but it still is a fact that the detail man working with problems he did not cr~ate himself and tlia~ are part of the medical climate today, working in a. doctor's office, with a doc- tor whom lie knows from previous calls over the years, has had re- markable results in terms of sales produced out of this personal contact. Senator NELSON. Do you think the effort of the detail man to deter- mine which products the doctor will use is a more significant influence on the doctor than the literature lie will receive? Mr. SQUIBB. Oh. yes, yes, I do. Some physicians, for their own rea- sons, do not like to see detail men or spend much time with them, other physicians perhaps are not contacted Often, for various reasons, by the detail man, `but where there is a strong detail man's pieseiltation to a physician who is known to be a good producer of prescriptions, this physician will be more influenced by the detail man who works with him, sees him once every month or so at least, than by any other promotional force that the drug business can produce. Senator NELSON. I am sure that there are many very intelligent and conscientious detail men, and I do not mean to reflect upon them at all. However, a detail man, no matter how intelligent or conscien- tious lie may be, is not a professional man in this field, insofar as having the qualifications to really judge what drugs a patient should have, although he may be very expert in what his own company's drug is and what they say it will do. So here you have a situation of a doctor relying upon a man who is not a pharmacologist, not a pharma- cist, though some may be, to help him make decisions about what kind of a drug he should use in the treatment of his patient. Mr. SQUmB. Well, perhaps because I have, been a detail man myself and have done this work quite a hit, I feel that the quality of a detail man's work for the reputable companies, although you can always find examples unfortunately that are to the contrary, is excellent. They know the area in which they are working. They know their product. They are well trained. They follow the instructions which are given to them in no uncertain terms as to what they can do, the limits within which they can operate. The detail man is, generally speaking, working in a marketing situation and in an information situation which was not of his own creation. You understand the doctor is there often without information. He has not gotten it, either from his training or back- ground, or he has not any time to get it any other way. The detail man moves into this situation with skill, with discretion generally speak- ing, and is a fine means, although expensive, of transmission of this information. I think he is the best because he is armed with the pro- nlotional material he is going to talk about, he is armed with the ad- vertisements, the samples, the promotional literature, and the refer- ences. He has all of those with them and he can discuss them because PAGENO="0038" 1588 COMiETITIvE PROBLEMS IN THE DRUG INDUSTRY he has been trained to discuss them within the limits of that product. He is not practicing medicine. He is talking about this one particular product and the other products that relate to it. I think that this is the finest practical way that has been developed today to bring informa- tion to the physician. The physician knows that this man is not a physician. He knows or should know the limits that the detail man has, and it might come to a point in every detail conversation where that conversation must stop and the question referred back to the home office for further expansion. I think generally speaking, and I know there are horrible examples in your own record of detail men who have done wrong, this is an ex- cellent method for the transmission of product information which is one of the problems that we are faced with. I think the fact is that sometimes the detail man is criticized for things that are not his fault if he has too many samples, if he has literature which is too flamboyant. He did not create that himself. These are the policies of the house that sends him out. The detail man as a transmitter of information, as a part of the drug world today, is a very excellent though expensive method of doing the job that is needed. Senator NELSON. I come from a profession where you had to try your matters in open court with an adversary who tests every con- ceivable proposition that is advanced. Now, if you had all detail men who handled the same drug before the doctor, and the physician could hear all the arguments about which one is best, and then make an evaluation, it would be quite another matter. But here you have a situation where the doctor has listened to one man from one company with one or several drugs and only one purpose: to sell his firm's product. And even though he may very well know that there is another drug equivalent or better at a lesser price sold by another company, he will never disclose this fact to the doctor. So the doctor is listening to a kind of a one- dimensional presentatioi~, and I observe at this point that this bothers me a little bit. Mr. SQUIBB. Yes. That of course is the limitation. A man is presenting one product with all the skill and knowledge that lie has, and that is his job, to present that one product in that one field. Senator NELSON. I am not critical. That is the job he has. Mr. SQUIBB. That is the job he has, but as a matter of practical operation another man will be along shortly to present his product, a competitive product, and lie will do the best he can with that. I think that these things are not perfect. These are far from perfect ideal situations, but given the diffiei~ilties of a multiproduct market, and the complex and technical problems that are involved here, the detail man and his acceptance by the physicians has evolved, has developed to such an extent that it works, and it works as well or better than anything else now being used. This is where I think physicians look for their information. These are detail men that they know and accept. They will not see some other detail men. Perhaps, they do not like them or they have made mistakes in their office. But generally speaking the teclmique is the best one that the pharmaceutical indus- tries have come up with for the purpose of simple transmission of information on its products to its customers. PAGENO="0039" COMPETITIVE PROBLEMS IN THE DRUG INDUSTRY 1589 Senator N~soN. That method certainly is not the equivalent of an equality method of presentation, using a distinguished therapeutic or formulary committee in a hospital, even though of course detail men do go to hospitals. Mr. SQUIBB. Yes; and sometimes the committee will ask a detail man to come in and describe a given product to the therapeutic committee. Senator NELSON. Yes. Mr. SQUIBB Just within those limits and to make a presentation on his product, they might ask several in. Then they have to make their judgments. Senator NELSON. As a matter of fact, I am familiar with cases where that is the single assignment of the detail man. He spends all day every day at the hospital. However, under that circumstance vis-a-vis a sole practicing physician, you do have in that hospital an expert in blood dyscrasia, an expert in heart disorders, internists, and others plus pharmacologists and pharmacists. So you can benefit from the experi- ence of people who have tried all these various kinds of drugs, or read the literature or have their own expertise in some aspects of medicine in which they know the drugs used. This is quite a different case from the situation where a single detail man is giving some information to a lone practitioner out in the country someplace. Mr. SQUIBB. Yes, I do not mean this disrespectfully but I found from my own personal experience that when you are talking to a given physician in his office and telling him about your product you are talking to the world's greatest expert. This is his attitude. He wants to know for himself. He is not interested in other experts. You can bring in reprints but you have to be careful in dealing with an individual physician. You are talking to a man who is perfectly capable of making judgment on products and does so as a result of your presentation and information. Doctors do depend on skillful, accurate, careful detailing of products. With all of the drawbacks which you have very well pointed out, it has grown up over the years, and I think it has grown up because there is nothing else, not because it itself may be the best thing that ever will be. Under the conditions that have prevailed certainly in medicine in the last generation, this has been the best way to work it~. Now, with the growth of. hospitals, as I have pointed out, it is going to put a new burden on the detail procedures, a need for reexamining these detail procedures. There will be new requirements for the detail man, how he is trained and what he does. This is the point~. I think the burden that the detail man will carry in 3 or 4 years from now will be quite `different from what he carried 3 or 4 years ago, `because the medical world is changing. The companies that see this first will get to work on their detailing studies, the effectiveness of their detail staff, and the money spent for it, and they are going to `be that much ahead in getting the business out of `these therapeutic committees that are springing up and controlling the hospital business. `Senator NELSON. Go ahead. ,Mr. SQUIBB. I am in the middle of page 19. The analysis of this cost, that is the cost of the detail man, in terms of time spent to produce a dollar not of `sale.s but of profit is urgently called for, frightfully difficult to obtain, and very painful to apply. But it must be done, and when, as and if it is done, there c'an only be a large reorganization, reapplication in terms of assignments and resultant substantial PAGENO="0040" 1590 COMPETITIVE PROBLEMS IN THE DRUG INDUSTRY changes in the product mix and customer service patterns of a given company. Efficiencies obtained from such action, however, would off- set much of the loss of income from price reform as already described. Prices are inflated, and by some sort of Parkinsonian law, sales activi- ties are inflated within the limits set by those inflated prices. A better understanding of the precise source of the prescription specification-who wrote it, why he wrote it rather than something else, and where and at what price it was filled is essential to build a Irnowledge of what to do about getting the next specification for your label. The reason it was received yesterday, may not and probably will not, be the reason it comes in tomorrow. The industry does have at the present time much statistical data on the number of prescriptions written, the products specified, and even the prices charged. However, this data is slim on precise motivation behind each selection of drug prescribed, and it is difficult to trace back to the basic cause for the choice. Besides this failing, the shifting trends insofar as numbers of prescriptions written under circumstances other than those of the traditionai private patient in office practice which a few years ago was literally the sole source of drug specialty sales volumes are not properly reflected in data form. Improvement in available information on prescriptions from every aspect will be forthcoming soon as studies now being set up are completed, and fuller utilization of computers is realized. Progress in this area has been slow, for even although a pharmaceutical company is so often quick to spend several hirndred thousands of dollars in research efforts to develop the best possible formula for a drug product, it is more often than not quite unwilling to spend even half that amount to determine the key patterns of the ever-shifting marketplace for the product when it is finally perfected. Even more specifically than this, however, the pharmaceutical manager needs to determine what activities on the part of his sales staff, produce which results in his sales area today. Generally speak- ing, lie is not succeeding in getting this information. Partly this is because it is a very difficult thing to do, but partly it is because there have not as yet been real pressures as yet to require it. Behind it all is the knowledge that if he does acquire it, it will be an overwhelming job to implement his new knowledge with a completely revised and reorganized and reassigned staff. The shifting pattern of medical care centered around the growing importance of the hospital-its accident rooms-outpatient depart- ments, diagnostic facilities, and its expansion to include nursing home and convalescent care in its scope of influence will change the job of the pharmaceutical detailman in the years immediately ahead. This will be true not only to the extent of where he works, but also how he works. It will reduce him in numbers, but no doubt will increase his individual cost to his employer. It will place new burdens on his train- ing, and new strictures of regulation and control on the performance of his duties. All of this calls for immediate reexamination of every facet of this part of the sales procedures of the industry. The phar- maceutical detail man must be, like Caesar's wife, above any suspicion. Economies and readjustments which will result will contribute to the improved climate for pharmaceutical industry operations. PAGENO="0041" COMPETITIVE PROBLEMS IN THE DRUG INDUSTRY 1591 A similar close look must be given to the costs of advertising and product promotion outside the detail man. Criticism of the industry in these areas has been mostly concerned recently with the quality of the advertising material in terms of~ full disclosure and balanced presentation of product information. The results of these regulatory efforts will be, of course, of real significance to the industry, but the question of the effect of advertising on prices is the consideration here. Again we have a difficult problem of balance for the pharmaceutical manager to solve. As always he has a special responsibility to avoid waste, to keep within generally acceptable limits, and to persuade his potential customers that his advertising efforts are useful. To the degree he fails to do these things, and parenthetically he thinks that this is going on, they are failing to do these things today to some extent, to too great an extent, these expenses will be deemed an unnecessary force and contributor to the upward trend of prices. Advertising in an industry is vulnerable to critics who see in it many sinister implications contrary to the public interest, but espe- cially in the drug industry is it suspect of exerting influences which are not only productive of undesirable medical practice, but actually harmful to the heaith of the Nation both in a physical and an economic sense. The pharmaceutical advertising manager finds himself in the same sort of position we have seen before in relation to his counter- parts in other industries. His rules should be different, his responsibili- ties for careful,, accurate. and nonprejudicial work are greater, his field of act~ivity is more restricted by regulation and by public policy, and any deviation will not only be measured in terms of communicative quality, but also in economic effect. Senator NELSON. A week or 10 days aoo the representatives of the Pharmaceutical Manufacturers Association appeared before the com- mittee. We showed them an example of a drug for which the Food and Drug Administration in this country requires in all the advertising of the firm who makes the drug a very speciflc warning to the physician that fatal blood dyscrasias have resulted from the use of this drug and that it should not be used for minor diseases or illnesses. In February of this year the company ran an ad in the AMA Journal in this country, February 20 1 believe, in which they made a full dis- closure as required, in a paragraph with a black line drawn around it. It mentioned all the dangers and warnings very specifically. WTithin the same month, in fact a few days earlier, they ran an ad for the same drug in the British Medical Journal in which they did not include a single word of warning. When I asked them why, they said they simply comply with the laws of the country in which they advertise. I said, "Aren't you worried, concerned about the British people?" I said, "What about the underdeveloped countries where there are no standards, no laws defining required quality. They do not have doctors who are qualified or do not have enough doctors to set up any testing procedures." Well, their only response `to that was "We comply with the laws of the country we are in." So what they were saying was that their stand- ard of quality in `advertising here is as high as it is required to be by law and they really aren't worried if some doctor is misled in England or in some other country where the law does not require full disclosure. PAGENO="0042" 1592 COMPETITIVE PROBLEMS IN THE DRUG INDUSTRY Do you have an observation you would want to make about that? Mr. SQUIBB. I think that is horrible. I think to the extent as you have described it., if that feeling exists in the industry, it negates the special social responsibility which I think anybody dealing in danger- ous drugs has, because all drugs, even the greatest life-saving drug, has dangerous potential. I believe that every pharmaceutical manu- facturer should without limitation or selection at all times disclose completely those things within his own knowledge of his product, making them available to everybody who is a potential customer. The fact that this might be in a country where there is no law as you say whatsoever, and you sell off the back of a wagon, this is no different. If you take any other attitude today, it is just disastrous, it is immoral, it is unethical, it is stupid. I thjnk it is perhaps even worse than being stupid, and a number of other things which you can be, because it is bound to come back on you and bring retribution as soon as your drug is inevitably used by ignorant or uninformed people. In the long term this is not what you can live with. I think it is a horrible story that you recite there. I cannot see how a responsible company can fall back on "the law" to limit its burden for proper disclosure of things it knows about its products, good and bad both. Senator NELSON. Thank you. Mr. SQUIBB. Now we have come to the problem of research, maybe the stickiest of them all. We have seen this point brought up here already today. Research costs are used more often than any other one factor as an explanation for drug prices whenever the 1)harmaceutical industry is replying to its critics-or even when it takes the initiative in one of its "image building" or "storytelling" phases. They are certainly part, and an important part, of the industry's operations. They are in the overall sense essential to keep the medical armamentarium of drugs moving ahead. However, a great deal more light needs to be shed on the general subject of pharmaceutical industry research and the costs thereof-both as to its own internal research organizations, programs, and procedures and as to the research projects it supports outside its own w-alls. It is most difficult in industry circles to criticize "research," or even to attempt to stand off and look at it objectively. To many, drug re- search is almost a religion, and to question it in any practical way is to speak heresy. However, it must be said that there is a poor, waste- ful, extravagant, unproductive, unimaginative, and pointless work being done under the category of "research" by the industry. It would be perfectly astonishing if that were not the case. The degree of this among all of the productive and sound research operations carried on is certainly open to argument. Most probably it varies from time to time and from place to place, but it always repre- sents a sizable part, and a significant one because of the amount of money involved. There are several questions that should be asked by every pharmaceutical manager and in turn by his research director, and answers given before research objectives are set and funds to carry them out are committed. These questions involve such matters as the proportional size of research budgets both minimum and maximum in relation to basic operating costs or even to sales; number of projects PAGENO="0043" COMPETITIVE PROBLEMS IN THE DRUG INDUSTRY 1593 to be supported at any one time and in any one field; relation of basic research budget to product and developmental research expenses; similar or overlapping efforts elsewhere; and last but certainly not least the need actual or potential of the anticipated output of its re- search activities after the forecast costs have been incurred, and after market trends now readily discernible have evolved into significant new requirements for drugs from both the therapeutic and economic standpoint. In other words, I am saying let us look where you are going be- fore you get there. It will be said by every pharmaceutical manager and research di- rector that they do all these things; that they could not do their job if they did not; and that this is simply a restatement of what is clone every time a research dollar is budgeted or a project entered into. 1-Jowever true this claim may be in theory, it leaves much to desire in practice. From the standpoint of cost, the research operations of the industry are not yet properly integrated into the whole pharmaceutical operation, and their relationship with other parts-~articular1y sales, are not clearly worked out for many reasons, most of them internal and peculiar to each individual company. Before acceptiiig the idea that all is well in this particular segment of his responsibility because merely to do research is in itself a creditable and self-rewarding ef- fort, a pharmaceutical manager should take a hard look at his research expenditures in relation to his project accumulative, new prod1uct fore- casts, his competitive position, and the state of his own pocket book and arrive at a private conclusion that he apply the same principles of cost control to his H. & D. programs as he does to manufacturing and sales. Research expenses can and should be closely reexamined by the pharmaceutical industry, and the resulting adjustments need to be ex- plained to, and understood by, the scientific community generally. Objectives, and general routes toward these objectives, should be dis- cussed openly so that better information is available to shape all con- current research programs to make certain of better utilization of the total research dollar. It is not enough for the industry to point with pride at the size of its research budgets, or even to their past research achievements. It is now essential that they explain their research pro- grams in terms of their place in the overall scheme of medical progress, their purposes in terms of public need, and their costs in terms of effi- cient procedures with a minimum of duplication and waste. Such are some of the things that the industry can do to answer the critics now pressing for reform enforced by legislation and bureau- cratic control. There is really no choice as things stand today so far down the road have matters gone. These things must be done by the industry or they will be done for it by others. Senator NELSON. Everybody, of course, favors research and develop- ment of drugs and other commodities, and research-oriented companies have made great contributions to the drug industry and to the health of the Nation, as have other such companies. With the great continuous emphasis placed upon research in the drug industry as a justification for high prices, interestingly enough a substantial portion of the medical profession is convinced that this factor is very significant. In PAGENO="0044" 1594 COMPETITIVE PROBLEMS IN THE DRUG INDUSTRY fact, a good many doctors will say "Well, I will prescribe the higher- priced drug from the research company for my patient because they are entitled to have more money to do more research." It is difficult to find out how much is research and how much is not. Mr. SQUIBB. That is a judgment somebody has to make in the account- ing procedure. Senator NELsoN. We have not had very much good testimony about industry research except the assertion from the industry that they have spent tremendous amounts. Is there any tendency so far as your knowledge is concerned for the industry to inflate the figure as to the amount of money they spend on research, or to include in it, items that really are not genuinely cost or are not genuinely research? Mr. SQUIBB. There is always the possibility, and this is generally the result of internal arguments between the two men involved. The transfer of funds from the sales program to the research budget by the sales manager versus the research director or from manufacturing or fro~n quality control. I think there is always a tendency, when you are talking about research, to want to talk about doing a lot of research, so you want that figure to he high when you are talking to bodies such as this. You want to tell how many millions of dollars of research that you did. So you want to include anything that might conceivably be thought of as research. I do not think that there is any inflation in terms of misrepresenta- tion of research. I think you can judge yourself favorably or unfavor- ably and make the judgment decisions on a perfectly honest basis without anything subversive about it. Senator ~ELSOX. The Harvard Business School did a study called "A Note on the TJnitecl States Drug Industry" and as to research the Harvard report said: In the drug industry research and development is interpreted somewhat broadly to include not only the search for new compounds with medicinal effects but also the costs of elaborate testing procedures designed to determine the safety and efficacy of the drug and elaborate quality control procedures designed to secure purity and uniformity in manufacture. Is there any standard which companies use to determine what should be included as research and what should not `be included as research? Mr. SQtTIBB. No, there is no standard, and I think there is a great dccl of dispute among the various research personnel in the different companies as to what it should be. I know that some companies quot- ing another company's research will answer it by `saying, "Oh, well, they throw a lot of things in there that we don't," but they do not really know this. This is an individual decision in each company, and reflects the philosophies of the management in charge at that moment, and might vary even from administration to administration. I do not think that this is a large enough factor, this judgment variant here, to really be a confusion or a misunderstanding in the job on research that the industry is doing. This is a.n internal matter of accounting proce- clures which would make it vary a little bit, but not really a great `deal. Mr. GORDON. Do you think that the wider use of the formulary sys- tem would have a tendency to increase the quality of research? Mr. SQUIBB. `Well, yes, to the extent that duplicative products exist. Inasmuch as only one of a kind of a product `is all that you can put into a formulary, if you have a line that is entirely duplicative of some- PAGENO="0045" COMPETITIVE PROBLEMS IN THE DRUG INDUSTRY 1595 body else's line, you are going to have trouble getting it into a formu- lary situation. From that point of view the pressure will be on research to come up with products that are `different, either different physically or with different therapeutic purposes. That is quite right. If only one selection of products is `going to be made for a given formulary, the pressure will be in every company to come up with some new, some fresh, some unique products for the line so that they can be justified for formulary inclusion. I think this will be an important factor in directing research in the future. It will be a pressure on research departments to come up with unique and new, not simply imitations of successful products. There are many approaches that a legislative body, State or Federal, can take to attack high prices for pharmaceutical products. In all of them it is suggested that they be studied from a more objective, long'- ranged and balanced view with less emotion, prejudice and p01it~ca'1 consideration. The economics of health care, a potential financial bi~r- den to the Nation that can be and is grossly underestimated, needs the thoughtful attention of the best brains that can be devoted to it. There should be no off-limit areas for investigation all the way from the man- ufacturer through the last distribution point for the product. There should be no preconce~vecl theories of merit based on real or imagined evil pract~ces of the past. There should be no temporizing with the best standards o-f quality and safety for all those w~ho are to benefit in the future from new concepts of public health. At the State level legislatures can reexamine the meaning and intent of their drug substitution laws, in view of current welfare programs, local, State and Federal. These laws passed 10 or 12 years ago served quite a different purpose then, compared to serving a Government-fi- nanced public assistance program. There are ways to permit utilization of the finest drugs in these programs, or any specialties that the pre- scribing physician may desire without freezing into the State welfare budgets retail level prices. The State formulary system carefully and thoughtfully adjusted to the practical needs of the patient can be com- bined and coordinated with pharmacy laws concerned with subst~- tut~ons. Senator NELSON. What do you mean by a State formulary system for welfare? Mr. SQUIBBS. State formualries such as a good many States already have and all States will have I am sure, setting forth the products which that State will pay for under its welfare programs. Senator NELSON. How many States now have a formulary system? Mr. SQUIBB. I do not know exactly. Many of the major States. Cali- fornia hafls been a notable leader in this, getting in first. I think it is more general in the Western States than it is in the Eastern. But again as the States come up with welfare programs and budgets and allocate or appropriate money to spend for drugs, they have got to impose some ground rules, some limits on what they will spend it for. Other- wise the end is off and they are broke inside of 2 or 3 months. We begin to see some of these things developing~ in the medicaid programs that have been launched without careful consideration of the ultimate cost. I think when a State, just like a hospital, has to pay a welfare drug bill and contracts to do this, they are going to have to prescribe and set forth the area in which they will pay. This work is most difficult, and it PAGENO="0046" 1596 coMPETITIVE PROBLEMS IN THE DRUG INDUSTRY has to be done with all kinds of considerations. But it is their judgment, absolutely their own judgment. Mr. GORDON. Am I correct, Mr. Sqthbb, in assuming that what you are saying here is that the original reason for passing the antisubsti- tution laws no longer exists? Mr. SQUIBB. No. The original reason does not exist as the primary reason. Mr. GonioN. Primary reason? Mr. SQUIBB. The primary reason. Of course, you see when the sub- stitution laws were passed 10 or 12 years ago there wasn't any con- sideration that the State itself would be a purchaser of drugs to any substantial degree at all, and the laws were to control individual retail operations in terms of substitution. Now, when you come and set your- self up with a formulary, deciding which drugs are going to be pre- scribed and what price limitations you will recognize, and then you have a substitution law which forbids the retailer to make changes, he is in quite a box. He is in an impossible situation. He has to supply a drug, at a price level at which he cannot buy it, or else break the sub- stitution law. Mr. GROSSMAN. Mr. Squibb, may I ask you along the same line, what would you think of repealing these laws, in view of the position that if a doctor wants to prescribe a certain drug, that he should be entitled and he should lrnow that his patient will get that particular, even if it is a brand name, drug? In other words, don't you think the doctor has every right to make this determination? Mr. SQUIBB. There is no question about it. The doctor should make and must make the determination, and nobody else. Mr. GROsSMAN. And that there should be no substitution, if that is what he decides that he wants? Mr. SQUIBB. There should be no substitution if he decides that is what he wants. Unfortunately there were examples when the sub- stitution laws were passed that this was not always carried out. Mr. GROSSMAN. So then t.here is still a place for antisubstitution laws? Mr. SQUIBB. I think there is still a place for the substitution laws. Unfortunately these are similar to police laws. Any professional man who holds the title and license `of a pharmacist in properly carrying out his profession does not alter the prescribed medicine without per- mission. On the other hand, a law seems to be required because there unfortunately seems to be those who will do it, for reasons of either laziness or greed or other purposes. Senator NELSON. Just one point here. In a hospital it is a clear-cut case where there would not be any substitution, or rather the formulary will be used by the physician as agreed in the hospital. Then as I understand it all of them permit an exception when some physician says, "I want to make an exception for this reason," and they perimt it. The exception is rare because they do use the formulary. But when you have a formulary for all practical purposes the vast majority of the dri.igs that are used have been decided upon by a coin- mittee and the doctor selects from that list? Mr. SQUIBB. That is right. PAGENO="0047" COMPETITIVE PROBLEMS IN THE DRUG INDUSTRY 1597 Senator NELSON. So you cut the size of the formulary and you remove that element of decisioninaking by this particular physician unless he has a strongly held view about some particular brand name? Mr. SQUIBB. That is right.. Senator NELSON. The same would be true in the State formulary for the welfare program, would it not, except in the case where a doctor may decide that he will not select a drug from that formulary although the State will only reimburse for the cost of the particular drug on the formulary, and the patient will have to pay the difference, if there is a difference? Mr. SQUIBB. That is right. I think ~ye should not let consideration of the exception cloud over the real purpose of the formulary. You still put in an opening like that because the practice of medicine is the uncertain thing that it is, or the nonspecific thing that it is; I think is obviously is essential to leave a loophole like that. While these two bodies of drug regulation are now generally separate in the States, they must become parts of the same problem beca.use sheer economic necessity has made it so. Under title XIX of the social security law State welfare and medical programs have been made a most important part of the administration of public meclica.l care in the future. Each State is required to set up its own program in this area for I-JEW approval before Federal funds are allocated to share the cost. States should take this opportunity right now to review all their pharmacy laws to make them suitable for maximum utilization of funds available without lessening of the quality of medical care afforded insofa.r as use of drug products is concerned. At the Federa.l level patent laws ca.n be restudied with realistic and dispassionate attention to the proposals made in the Kefauver report several years ago. Time limitations on drug patents, systems of compulsory licensing at uniform royalties, and the pace of Gov- ernment-paid-for-research in patient procedure all need specific study now for the pharmaceutical industry alone, without relation to other industries who may read into such efforts a threat to their own patent protection laws. Consideration of the importance of medical care from both a social a.nd economic standpoint would seem to set apart the patent rules for prescription medicines from `those for other commercial operat.ions. Patents can provide an umbrella. for high prices and t.o that extent at least their value for medicines should be (luestionecl. Senator NELSON. Once a patent is granted. the company that secures it has the exclusive right in this country to manufacture and sell that drug for 11 years; is that not correct? Mr. `SQUIBB. That is right. Senator NELSON. At the end of the 17 years, there is only one com- pany in America., unless they license somebody else, which has devel- oped complete knowledge of how their Particular drug `is formulated. So there is one expert in `the whole United States. Now the patent runs out and the company has `had the benefit of `the umbrella, so to speak, protection that the law intended apparently to give them in order to recoup their re~earcll costs and to reward the company or the `individual for some creative contributions to society. PAGENO="0048" 1598 COMPETITIVE PROBLEMS IN TUE DRUG INDUSTRY 1-Ioweve.r, when the patent runs out you have a situation as I under- stand it, w'here neither the FDA nor the manufacturer discloses every- thing they know about that drug. So a company that then wanted to go on to the market would not be able `to get the bene.fi~ of all the information that was accumulated by the company that had the legal protection for 17 years, and this second firm could not come onto the market competitively and have the benefit of all the experience and knowledge that is accumulated by the original manufacturer. This system is defended by `industry on the grounds that, "Well, these are our trade production secrets and there is no reason why we should give them to anybody else." On the other hand. the only reason they got the 17 years' protection is that the people of this country gave them the prote~.t.ion. Do you have a viewpoint about what the require- ments ought to be in terms of disclosures to the public of all informa- tion a company accumulates about the manufacture of a. drug? Mr. SQUIBB. You mean the New Drug Application file Oil such a drug after the patent has run out? S~na.tor ~EL5ON. After the patent ha.s run out and a New Drug Application is filed. Mr. SQUIBB. I think there are some very interesting experiences being accunmlated along these lines right, now as these major patents run out in the tetracycline and chloramphenic.ol area. You see these thins going on right now. I think you should not take the NDA bundle of material in any seijarate consideration from the patent itself. You have got to require, if you wani~ to soften the length of the patent by making the drug patents say for only 5 years, that the NDA situ- ation Should `be tied into that. I do not think, having gone through `the patent in 17 years that now you should ride for the next 17 years in effect on your NDA. Here you are adding something again from the Government to the patent law. In other words, it is not a Valid patent any more but has t.he same effect. You have got this vast expensive bundle of NDA which the Government will not disclose to your competitors, so you can ride along on this. Senator NELSON. Just so that the record will be understandable to those who read it, would you explain in a little more detail? At the end of 17 years the New Drug Application is then filed by the company? Mr. SQUIBB. It has been filed, the New Drug Application, and you have been running on your patent. Senator NELSON. Excuse. me, filed at the time they got the patent? Mr. SQUIBB. Perhaps, but not necessarily at the time you got the patent.. Senator NELsoN. And that New Drug Application contains all the material, research results, and everything else that was required? Mr. SQUIBB. At great expense. Senator NELSON. To be supplied to the Government? Mr. SQUIBB. Of course, the Food and Drug can call this a new drug or an old drug. This has been covered in your testimony, particularly I think with Mr. Stetler at some length. But I think we ought to con- sider how long a drug product covered by a patent or a New Drug Ap- plicat.ion in the public interest should be kept exclusively to one com- PAGENO="0049" COMPETITIVE PROBLEMS IN THE DRUG INDUSTRY 1599 pany. I think we should examine this. I think very likely the period, because of the nature again of the industry we are talking about, should be shorter. Compulsory licensing at a reasonable- Senator NEr~soN. You say the patent period should be shorter? Mr. SQUIBB. The patent period, yes, the patent period. Let us talk about the 17 years first. Senator NELSON. Yes. Mr. SQUIBB. That is running first, and there is no point in talking about the NDA disclosure as long as the patent is in effect, so you have got 17 years given by the people for that drug. Traditionally in the Squibb family we think that patents are not desirable in drug lifesav- ing situations like this, but I think that perhaps that is too extreme a position in the current situation. Patents might be shortened. So when you are shortening the patents and giving perhaps a patent for 5 years exclusively plus required compulsory licensing after 5 years or after some determined period of time, I think then you reexamine your NDA in terms of what you do under the patent law, because you have to break the NDA if you are going to ireak the patent in practical ef- fect. There is no point in saying the patent only has 5 years, if you have essentially a patent produced by your New Drug Application, which only applies to your product, and is too expensive for anyone else in a competitive situation to try to duplicate. It is often millions of dollars of work and time. Senator NELSON. If you had comp~ilsory licensing, are you saying that it would require the company holding the patent to disclose every- thing it knows about the product? Mr. SQUIBB. Yes; and give a permission of the licensee to use the documentary protocols and terms that have been put in that NDA. Senator NELSON. How much information is in there? For example, the argument is consistently made, which no doubt is correct, that even if you have the same chemical compound and the same proportions in a tablet, the mechanical formulation of it or the coating of the tablet or what have you may still affect the degree of physiological availa- bility. Does the New Drug Application information disclose all the specific techniques for compounding and making the particular product? Mr. SQUIBB. Yes; but I think that is the minor part of it. That is the least difficult part of it. NDA includes the background of the clinical testing, the toxicology, and the studies that have been made preliminary to the release of the product and the claims that are pre- sented for the product are based upon the material that has been accumulated over a number of years and organized in the NDA. It may be $1 million or $2 million of expenditures involved there. There is a great deal in there which the new licensee coming into the situation would find it impossible to duplicate as he is faced with the oppor- tunity to go into the market. If he had to do that, he just would not do it. I think that the consideration is whether you should or should not treat drug products, professional drug products on which a patent has been given, any differently from any other coi~j~ercial item for which a 17-year patent is granted. If you ~e-go1ng to say it should be shortened, then you have got to effe~ti~ely change your rules on NDA. This is needed so that you make ITt available in not just a token 81-250-OS-pt. ~-4 PAGENO="0050" 1600 COMPETITIVE PROBLEMS IX THE DRUG INDUSTRY gesture, you make the product available, in fact, on a compulsory and fair licensing basis to whoever wishes to get into the market. Senator ~\ELSON. It might be well to shorten the patent period, but in any event, let. us assume for the moment that the patent period covers 17 years and that is not shortened? Mr. SQUIBB. It is long. Senator ~ELSOX. I mean is not shortened by Congress. I suppose it is likely in the near future that this period will not be cut. So you have 17 years. You still have two other things you could deal wit.h. One, compulsory licensing; the other, disclosure by the FDA of all the material in the New Drug Application. As I recall from the testimony, and you may have to correct me on this, Dr. Goddard stated that there is no law now on the statute book that forbids the FDA from disclosing the information, from making public the information in the New Drug Application at the expiration of the patent period. But that since it has been the practice, the administrative decision of the agency over a period of years not to do so, he, personally, would be reluctant to change the policy without some guidance from the Congress; that is, that the Congress pass a statute making it. the law. Is that correct? Mr. SQUIBB. That is exactly correct. I think he feels that that should be the procedure to be followed, because it is obviously a major ma.ttei of giving several million dollars of information to competition. I think he feels that. this is such a major decision, and that it would have such a major impact. upon the industry, that lie should get, and I think quite properly, a specific direction from Congress before doing it. Senator NELSON. If the patent laws continue to give protection for 17 years and no legislation is passed to require compulsory licensing, do you think that. at the expiration of 17 years companies holding patemits have had adequate time to get the benefit that they should get, and that. the law should require that t.lie New Drug Application be made public Mr. SQETIBB. Yes, but. you have to remember it is not always 17 years because a product does not necessarily get on the market the day the patent is issued, and you have to qualify this with some knowledge and some interpretation. Sometimes you oiily have 2 or 3 years actually of sale of a product. before the patent expires. So that that is iiot always the case. But. generally speaking I think that the two things have got to be looked at together, and without the arbitrary preconception of a certain number of years. The market and the lifesaving nature of particular products varies too, amid I think you have probably got to give some judgment amid discretion, wlieii I say "you" I mean the Congress. back to the Food amid Drug Adlministra.tion or the HEW, to make decisions as to how certain products should be handled. Some of them are much more significant than others in terms of their bene- fits. Some are routine. Senator ~\ELsOx. Do you believe that. compulsory licensing ought to be required by law? Mr. SQUIBB. I think compulsory licensing is the fairest amid best ap- iroachi to this problem. Senator NELSON. Amid undler compulsory licensing, t.lien, the owner of the patent would receive a royalty? PAGENO="0051" COMPETITIVE PROBLEMS IN THE DRUG INDUSTRY 1601 Mr. SQUIBB. Would receive a royalty, a fair standard royalty. Senator NELSON. That was to be my next question. Would you estab- lish the royalty by law or how would you do it? Mr. SQUIBB. I would establish the royalty, I think, in the law. Senator NELSON. And would you require compulsory licensing un- mediately after the drug becomes available? Mr. SQUIBB. No. I think probably some type of a period for ex- clusivity is necessary to stimulate or to keep going with research. I would have such a period a good deal shorter than 17 years, maybe 5 years or 3 years of marketing before anybody who asked for a license would be required to be given one. Senator NELSON. So if the law remained as it is, with a 17-year patent period, do I understand you to be saying then once the drug is introduced on the market, some other period of exclusivity, 2 years, 3 years, or 4 years ought to be established, to be followed by com- pulsory licensing? Mr. SQUIBB. That would seem to be the most equitable and fairest approach, serving the public on the one hand and the inventor on the other. Senator NELSON. Thank you. Mr. SQUIBB. The notion that without the inducement of a 17 year patent with no licensing requirements all medical research would die seems quite an extravagant and unrealistic claim. To the extent it can be modified drug prices will come down. Laws that now are concerned with competitive forces in the dis- tribution system for medicine must also be reexamined, particularly the effect of Robinson-Patman. The rigidity of drug prices in retail pharmacy where this law is `applicable, and quite the opposite situa- tion in hosiptals where it is not, have already been discussed. It has been publicly suggested by Dr. Apple of APhA that consideration should be given to enforcing Robinson-Patman on prices to institu- tional purchasers. This would ceratinly tend to bring the same price rigidity to this area of drug distribution now benefiting from unregu- lated price competition among suppliers. It is worth at least as much consideration to suggest the removal of Robinson-Patman from ap- plication to the pricing of prescription drug prices to the retail phar- mnacist. This would undoubtedly bring drug prices to the consumer down, but it would also bring down many retail pharmacists now benefiting from percentage markups of manufacturers prices. To the degree that the professional fee replaces this markup system however- pharmacists would survive. This is `a difficult problem, but one for which solutions can be found if the public welfare is given first con- sideration as certainly it should be. As `a matter of fact and as prelude to any possible revision of Robin- son-Patman or other laws bearing on pricing of drug products, a bet- ter general understanding of the economics of the entire drug dis- tribution system is essential. If real benefits are to be obtained in the form of lower drug prices without loss of product quality or mcd- ical progress in the form of new drugs, `a full realization of the costs that now go into a patient's prescription dollar is obviously called for. Attacking one small piece of this cost which even if completely re- moved would not appreciably `help solve the problem at hand would he PAGENO="0052" 1602 COMPETITIVE PROBLEMS IN THE DRUG INDUSTRY a very minor and even useless victory. Distribution costs of all kinds have to be weighed-their benefits and their burdens. The possibility that we are extending life today to a drug distribution Pattern that should have been made obsolete long ago if not for the fear of outrag- ing special interests should be given real objective consideration in the light of today's requirements, not those of a general or more ago. Among the approaches that could be taken by Government to the pharmaceutical industry is that of the public utility concept by which operations and profits are subject to close scrutiny and regula- tion as are those of electric and telephone companies, and common carriers. Here rates are subject to approval by Government. commission, and profits controlled within generally approved limits. Price ap- proval could similarly be given for drug products with the entire financial operation of the industry subject to review and control to establish levels of performance. This is a. most drastic alternative to possible reforms, but one which is by means unrealistic or incon- ceivable. Thirty-five years ago the~se words were published: The manufacture and distribution of medicines, because of their intimate relation to the health and welfare of a community or nation, partake of the nature of pu~~lic utilities. In view of the shifting of control from professional to financial hands, manifested by recent developments in the drug industry, the public interest may require "regulation" of the industry, through the guarantee of a fair return to investors and the limitation of prices to be charged to consumers.1 They are even more applicable today. Senator NELSON. Who said that and where was that published? Mr. SQUIBB. That was published in a book by the University of Chicago Press, "The Cost of Medicine" by Dr. Fisc.helis and Dr. Rorem in 1932, analyzing the costs of medicine at that time which were certainly an entirely different situation than the bedrock costs. today. Yet they saw this problem of special responsibility for an industry which at that time was just a fraction in terms of the sales dollar of what it is today. Senator NELSON. What is the total drug sales today? Mr. SQUIBB. Well, the prescription sales in 1932 were probably one-twentieth of what they are today. Senator NELSON. What are they today? Mr. SQUIBB. They are $2 billion or some such figure, and the figure back in 1932 might have been $30 million at the most. It is one of the most astonishing growths in industrial history. The impact of this growth of the prescription business in this country since the war, since 1950 really, is the thing that has brought to the industry all of its headaches. It is inevitable when you become very important in terms of dollar volume that new regulatory problems be picked up, and yet the industry does not always appreciate this. I think this is interesting that way back in 1932, 30-some years ago, two men looking at the industry from a philosophical point of view arrived at this particular conclusion. These are some of the principal approaches governments can and should consider to the problem of drug pricing. The extent that any of them are necessary depends upon the immediate reaction of the `"The Costs of Medicines," by C. Rufus Rorem and Robert P. Fischells, pp. 233-234. PAGENO="0053" COMPETITIVE PROBLEMS IN THE DRUG INDUSTRY 1603 industry to the pressures it is now undergoing from pubJic and legis- lative criticism. It is probable, the nature of the modern competitive corporation of no matter what industry being what it is today, that some direct Government action is necessary and desirable now. What- ever is done should be pointed toward tequiring the industry, and drug wholesalers and retailers as well, to reorganize and reform their own procedures, and not he restricted to the mere imposition of new regula- tions and controls of outward and basically superficial parts of the drug distribution system. Major economies are possible in getting our present drug products `into the hands of the public who needs them, changed only in price and not in quality. With the facts of basic costs as they are today, it is clearly false and stupid to say that this cannot be done. It is to be hoped that industry will take the leadership to do it-if not, others will. Senator NELSON. Are you familiar with the book written by Mar- garet Kreig called "Black Market Medicine"? Mr. SQUIBB. Yes, I read that hook. Senator NELSON. We haven't taken any testimony on this question. Do you have any opinion about how big a Problem black market drugs are? Mr. SQUIBB. Black market drugs are a problem. They are a problem, and you cannot say that it is a negligible problem as long as there is even one opportunity for black market drugs to get into normal chan- nels of medicine distribution. rfliis is because of the nature of drugs taken by people who are ill, people who are in no condition to resist an improper drug. I think it always has been an important problem, and always will be an important problem as long as human nature is what it is, because it seems there is always gOing to be somebody who is going to chisel and cheat and make false, dirty, substandard drugs. One of `the burden and one of the `special responsibilities of this industry among others is to fight this particular, battle. This is a more important matter in the case of drugs than it is in several other industries, because more `harm can be clone here. To say that there is only a minus fraction in terms of the number of dosages given that might be substandard or black market medicine begs the point, if it happens to be one of the doses that you swallow, it is a major `problem, and it always will be. Those of us who have been around the drug business for a while and have been in the drugstores and dispensing points know the constant danger of this, and the actual existence also of sedan peddlers and gyp artists of one type or another who purvey `these black market drugs. To the extent that they are resisted by ethical, upstanding pharmacists, and to the extent that they are turned down in their efforts to distribute these products, so much the better, but I `think the industry itself should take a very strong opinion that this is an important problem and always will be an important problem. Senator NELSON. Is there any educated guess about the volume of black market drugs or is it unknown? Mr. SQUIBB. The `only people I have `talked `to who kn'ow much about this are `the people in the Public Health Department of the State of New Jersey, who are very acutely aware of it because in `the industrial areas of the east, the built-up `areas of'the east, it is frequently seen. I think as you go out into the Midwest `and the scattered population cen- ters which are not so large, I think it is' much less `of a problem. I `think PAGENO="0054" 1604 COMPETITIVE PROBLEMS IN THE DRUG INDUSTRY in the New York City area, including New Jersey, with its type of pressures and the type of area we have there, I think it is a major problem. There have been many recent examples of stolen goods and. substandard goods, in and around New York, to indicate a very sub- stantial activity of this sort. But I think it probably diminishes as. you move away from New York City. It is probably a major problem in Chicago. I am sure it is in `Los Angeles, but in small areas it is not. Senator NELSON. Is it more of a problem to a brand name firm than a nonbrancl name company? Mr. SQUIBB. Yes, it is more of a problem to a brand name. Senator NELSON. Why is it more of a problem there? Mr. SQUIBB. Well, in the first place he is probably the target, the brand name man. Senator NELSON. Pardon? Mr. SQUIBB. The brand name man is probably the target of the black market medicines, as they generally are counterfeits of a given brand name in `a pre-established market or it is an imitation of it or it is some stolen goods of the brand name, or it is some kind of a deception practiced on a brand name. Therefore the man with `the brand name product has to take leadership in this. Nobody is trying to imitate a generic product. They could not have the monetary rewards from imitating a generic phenobarbital that `they get from Lluninal, the brand name product. I think the misrepresentation in this area is a direct `attack on a brand name product which is producing a sales and price level which is worth while for the criminal `to submit his own product. Mr. GRoss~L~N. Mr. Squibb, I would like `to ask you first of all `as to your concept of treating the industry as a public utility, do I `take it that you would `advocate `this if `they did not take steps, corrective steps, or do you advocate it a't this time? Mr. `SQUIBB. No, I do no't advocate it at this `time. Obviously this is the last resort to be imposed upon `an irresponsible industry which shows that `it cannot police itself. I do not advocate it. I would be very dismayed if that were necessary. Mr. GROSSMAN. What would be the methods short of using `this public utili'ty concept `that you would advocate now? Mr. SQUIBB. I think that the industry has to restore its reputatiOn among the general public and among the legislative `bodies, Congress, State and Federal, that it is operating within reasonably normal eco- nomic limits. It has a sense of responsibility which I have tried to stress through my presentation here, and I believe it can do th'a't. Mr. GROSSMAN. I do not know, I presume you are aware that the FTC has just published new figures which show that the dlrug mdus- try is now the No. 1 industry in terms of profit.. Mr. SQUIBB. Yes, I saw that. It was in the paper, I believe, yesterday in the New York Times. Mr. GROSSMAN. It now surpasses the automobile industry. Mr. SQUIBB. Yes. In 1966, drugs are the most profitable industry. I think this will probably be an occasion of congratulations internally in the industry, c'\~Te are. really going great and stockholdlers will like this," and all kinds of things and the WTa,ll Street analysts will greet this and the stock will turn up. I think that our problem is what this PAGENO="0055" COMPETITIVE PROBLEMS IN THE DRUG INDUSTRY 1605 sort of thing produces. I think we have a special difficulty with profit when someone takes our product who does not want to take it in the first place. Mr. GRoss~rAN. What were the figures, I think 21 percent? Mr. SQUIBB. 21.1 percent, a rate of return in 1965 had been 21 percent. Mr. GORDON. After taxes? Mr. SQUIBB. This is with an after-tax return of 21.1 percent. Mr. GROsSMAN. Could you suggest to the committee what you think would be a fair return? Mr. SQUIBB. I think a fair return, as I tried to say before, is the average return of comparable consumer goods industries that have a broad base of utilization. As a matter of fact if you want. to pick out a figure, I think it is ~rouncl 12 percent instead of this 21. Mr. GROSSMAN. Do you think that some type of limitations on profit might drive out some of the smaller companies? Mr. SQUIBB. No. Mr. GRossMAN. Leaving the field to the majors? Mr. SQUIBB. NO. Mr. GRoss~rAN. You do not think so. Would you say that any action to reduce profits would necessarily result in lower prices to the consumer? Mr. SQUIBB. Yes. You don't make an action to reduce profits per se. You do these things that I have outlined and some others, and you find what happens then, what you come out with. You plan and you budget ahead a 10- or 12-percent profit.. You reconstruct your indus- try in a believable or a credible fashion for public acceptance rather than leaving the situation as it is now which is out of control. Mr. GROSSMAN. I was just thinking for example of the steel industry which has been hurt very badly by imports and their profit margin has gone down, and we are all aware recently that their prices did not go down, but rather they have just gone up. Mr. SQUIBB. Well, you have got a different situation here. I am say- ing the industry should be concerned with the opinion that. outsiders have of it, because it is such a prime target for regulations. Specific regulation by people who can do it at legislative State and Federal levels. So these people will do it, unless we in the industry take in our hands to correct what are not really. abuses legally, but evidence of good work done in the past. To get yourself to a 21 percent profit return you have to be pretty smart, you have to be creative, you have to have done something right to reach that position. But in reaching that posi- tion, you have opened yourself up, because of the nature of your in- dustry, and this alone, to real serious problems. Mr. GROSSMAN. The social responsibility. Mr. SQUIBB. That is the point, and it is not to the point that your prices have gone down or other things in so doing, or that you have de- veloped any drugs that are for killer diseases and all this kind of thing. That is beside the point, too. We have got to maintain ourselves in an atmosphere of extraordinary goldfish bowl credibility or visibility expected of all serving the health area, not only of the industry which we are talking about here today, but they expect it of everybody who is taking part of the health dollar. We want to be sure that the public's dollar, whatever is spent for medicine or hosipt.al services or medical PAGENO="0056" 1606 COMPETITIVE PROBLEMS IN THE DRUG INDUSTRY care and doctors' offices or diagnostic purposes, that all of this is prop- erly utilized. Mr. GROSSMAN. May I just ask one final thing. You talked about the detail man, saying that it was the best practical method thus far de- vised, and you also talked about advertising. I was wondering what are the chances in terms of practicality of devising some type of ob- jective formulary or compendium with FDA approval, maybe using the FDA, maybe using the Physicians' Desk Reference with the FDA, and would this work? In other words, do you think it would be valuable to the doctor? Could he use it? Is there a more objective way to get this information to the physician without having him rely on the adver- tising and the detail man exclusively? Mr. SQUIBB. That is a tremendous question you have opened up there. WTe could start over again on that. Of course, there are two things in your question. Compendium and forniulary. They are not the same thing, nor for the same purpose. You have to separate them. I think the compendium idea is feasible. YOU can do it. I think it would be a waste. I do not think it is practical in concept of use. Mr. GROSSMAN. You (10 not think that the doctors would really use it? Mr. SQUIBB. 1-le would still have to continue to depend upon his other sources for information even though it is there.. It is hard to take that position because it seems von are sort of against knowledge, not against the compendium. I just do not think in practice it is worth the millions it would cost.. A formulary system for the Government, of course, is something else again. And here, as I have said before, I think a formulary sys- tem is inevitable for the Government, now that it has decided to pay for medicine to the extent it has. I think the Government will pro- duce, if not now in preseilt legisiatlon, in subsequent legislation, a formulary. I think it will be a rather large formulary in terms of scope. It has to be because it covers a large area. It is not just a local problem or even one major hospital problem. But I think it has to be. WTl-iei~ von take money and von decide to spend as much money as the Government decides it is going to spend, you have to set down ground rules, and a formi~ilary is simply the ground rules for spending the money. Mr GORDON. Mr. Squibb, are you acquainted with the PMA adve,r- tisemnent in the November issue of the Reader's Digest? Mr. SQUIBB. I received one. Mr. GORDON. Have you seen it? Mr. SQUIBB. Yes, I have seen it. Mr. GORDON. I wonder if you would give us your comments about it. What is our opinion of it in terms of its technique and its content? Mr. SQUIBB. That is a big question, too. I think the technique is old- fashioned and very disappointing. The one case history, as an adver- tisinsr manager I would not have approved that technique. I think it could have been iunch smarter and more effective in carrying out the message. I think the technique of mailing something out~ without 1rol~er examination, and I have read the testimony, of course, and have heard the arguments about trying to or leaving off the advertis- ing designation, is most unfortunate. I think this is a most unfortu- nate thing, because again the industry should take special precautions PAGENO="0057" COMPETITIVE PROBLEMS IN THE DRUG INDUSTRY 1607 to avoid any of that type of suspicion. Now it may be that this was simply carelessness or oversight or some low-level clerical mcompe- tence. However, it was done and I think this is most unfortunate that that was done. This is a personal opinion only, of course, but I do not think that the advertising that the PMA is getting in that form is worth, what was it Mr. Stetler said, $250,000 a crack. I surely do not. Mr. GORDON. How about the content? Mr. SQUIBB. Well, that is what I mean, the content of the advertis- ing, using a one-case story. It is old hat, an old-fashioned way of doing this. I see what they are trying to say. They are trying to tell the in- dustry's story. I think they could do it in a nlucJI more modern effective way than that particular one. This same type of thing has been done many times in the past. I think, as a personal opinion, that this is rather wasteful. And again I think this is the type of thing that does not impress `the doctors, who say, "Well, you have spent all that money, a million dollars for this, why don't you cut `the prices?" This is one of the nastiest comments an industry man gets to his advertising and I tried to cover that in my advertising discussion. You must be sure that your advertising looks believable and effective so that somebody does not say "nut that out and cut the prices clown." This piece seems to me to be a waste, poorly conceived, and inefficiently presented for the dol- lars spent. Senator NELSON. I want to thank you, Mr. Squibb, for your testi- mony. Your statement `and comments, I think, are among the most thoughtful and refreshing contributions to the dialog about the problems that these hearings are being conducted on of any witness that `we have had. I am sure there will be `some in the industry who will not agree with everything you have said, but your testimony com- ing from a person with long and distinguished experience in the in- dustry, your comments which are obviously intended to `help guide the industry `to improve itself and avoid regulation of some kind or an- other that might come, your testimony with its constructive criticism evinces a dedicated concern for the industry rather than an attack on the industry which the Pharmaceutical Manufacturers Association in- terprets your remarks as being. I think it is unfortunate that the asso- cia'tion would make an attack in `a manner which weakens their position even more. But in any event, you have made a `significant ~ontribution, and if the leadership in the induStry would give serious thought to `the observations you have made, it could be of great value `to the industry itself as well as to the health of the Nation. You have made a great contribution and the committee is appreciative of it, and I am sure that those `~h'o understand what this is all about and who are con- cerned about the problem across the Nation, will be appreciative of your public service in `coming here tod'ay `and expressing honestly, intelligently, and creatively your concern as well as offering your sug- gestions for improvements in an industry which i's a great in'dustry and which can avoid getting itself into serious problems if it is willing to impose some self-discipline as you suggest. Thank you very much. Mr. SQUIBB. Thank you, Senator, very much. Senator NELSON. That will conclude the hearings until Tuesday, the 19th `of December in the Caucus Room. rphank you very much. (Whereupon, at 2 :30 p.m., the subcommittee was adjourned `to re- convene on Tuesday, December 19, 1967.) PAGENO="0058" PAGENO="0059" COMPETITIVE PROBLEMS IN THE DRUG INDUSTRY TUESDAY, DECEMBER. 19, 1967 IJ.S. SENATE, MONOPOLY SUBC0MMIT1~EE OF THE SELECT COMMITTEE ON SMALL BuSINESS, Washington, D.C. The subcommittee met, pursuant to ball, at 10:10 a.m., in room 318, Old Senate Office Building, Senator Gaylord P. Nelson (chairman of the subcommittee) presiding. Present: Senator Nelson. Also present: Benjamin Gordon, staff economist; James H. Gross- man, minority counsel; Susan H. Hewman, research assistant; and Williaiyi B. Cherkasky, legislative director, staff of Senator Nelson. Senator NELSON. The hearings of the Senate Subcommittee on Monopoly will come to order. We will have a presentation today by a group of distinguished economists representing the Pharmaceutical Manufacturers Associa- tion. Since the economists are here today, I want to make a brief state- ment to raise a question `which I would, be happy to have them respond to, sometime during the course of the testimony. Last spring we had testimony put ii~ the record about comparative prices paid by cities and municipalities around the United States for drugs. There was one error in the st.atement with which the witness had nothing to do, but the statement was attacked by the Pharmaceu- tical Manufacturers Association, particularly by Mr. Stetler in a number of speeches around the country saying that the committee had been misled. So I had the subcommittee do its `own survey, and so that there may be an opportunity, while the econOmists are here, for the industry to comment on this survey, I will just read my brief `statement and during the course of the hearing we will be happy to have the com- ments of any of the economists on the statement. Early this year the New York City Department of Purchases con- ducted a survey of the prices paid by 12 municipalities, cities, and counties, for a list of 10 drugs. The survey revealed an extraordinary disparity in the prices paid by these municipalities for the same drug. There appeared to be no rationale to account `for the differential prices paid for them. On May 15, 1967, William F. Haddad, president of the New York Citizens Committee for Metropolitan Affairs, presented the results of the survey in his testimony before this subcommittee. A great deal of criticism of the survey subsequently was heard from the drug industry, and Mr. Joseph Stetler, PMA president in par- 1609 PAGENO="0060" 1610 COMPETITIVE PROBLEMS IN THE DRUG INDUSTRY ticular. In a speech given in Chicago on September 23, 1967, Mr. Stetler stated that Mr. Haddad's testimony was false, and asked, "Why did the Senate witness mislead the Nelson subcommittee ?" Mr. Stetier's criticism ~was directed in particular to the alleged excessively high prices paid by Fulton County (Atlanta), Ga., for various drug products. Although the implication in Mr. Ste'tler's speech was that there was an attempt to intentionally mislead the committee on the part. of Mr. Haddad, it is clear to me that no such deliberate falsifica- tion was intended. Mr. Martin Gersten, acting commissioner, New York City Depart- ment of Purchases, had written t.o Fulton County requesting the prices paid by the county for certain drugs. The data requested was furnished in writing by Mr. F. T. Fa.rran, purchasing agent for Fulton County, to Mr. Gersten. This information was then submitted to the committee by Mr. Haddad exactly as he received it from the New York City Department of Purchases. i wifl submit for the record the letter requesting price data sent by the New York Department of Purchases to Fulton County as well as Fulton County's response. it is clear from this correspondence that there was no intent by Mr. Haddad or any- one else to mislead the subcommittee. in fact., our own cross checking of the New York City survey indicated that it wa.s correct with a few minor exceptions which do not in any way invalidate the general import of the survey. However, since Mr. Stetler's criticisms may have cast doubt upon the general validity of our record in the minds of some people I asked my staff to conduct a similar survey of drug purchasing by a repre- sentative group of cities and counties. Letters were mailed from my office to a total of 77 cities and counties with public hospitals-other than those cities surveyed by New York City-using a list provided by the National League of Cities. In other words, I want to make it clear that none of these cit.ies included in this survey were surveyed by the New York City Department of Purchase, and I used a list provided by the National League of Cities. A total of 38 cities or counties responded. Of these, only 29 purchased a sufficient number of drugs, or in sufficient quantities, to make their price figures meaningful. The Defense Supply Agency and the Ve.ter- ans' Administration are also included, as well as price's charged to the retail pharmacies. I also requested similar data. from Fulton County to verify the in- formation given to the committee. The reply received by me from Fulton County did indicat.e that the most appropriate source for in- formation on Atlanta's drug purchases would have been Grady Mem- orial Hospital, not the county, since it wa~s the largest purchaser in the county. The request for price by Mr. Gersten in New York City was made in good `faith-a misunderstanding clouded the interpreta- tion `of Fulton County's response. I will submit for the record the letter sent to me by the Fulton County Board of Coun'ty Commis- sioners. The 29 cities included in this survey by my office represent a cross sect.ion of cities from all parts of the Nation. It is more extensive than the New York survey, and, furthermore, it strongly reinforces the PAGENO="0061" COMPETITIVE PROBLEMS IN THE DRUG INDUSTRY 1611 evidence of widespread and shocking price differences paid by pub- lic agencies, with taxpayers' funds, for prescription drugs. The dif- ferences cannot be explained by the quantity of the purchase because in some instances the highest price is paid by the largest purchaser and in others the difference in quantity is not sufficient to justify the huge price differential. Neither does the source of the drug explain the differences since the same supplier frequently charges different prices to different buyers of the same drug. In the main, I think the obvious conclusion to be drawn is that the price charged to public agencies is whatever the traffic will bear in each particular instance. For example, while Grand Rapids, Mich. (Kent County), was paying $160 for reserpine, .25 milligram, 5,000's, Chicago (Cook County) was purchasing the same drug for $2.09. That is $160 versus $2.09, a dif- ference of 7,655 percent. Grand Rapids was buying on a trade-name basis, while Cook County's order was for a generic product supplied by a generic house. The quantity purchased does not explain the dif- ference since the Chicago purchase was $338.58 worth and Grand Rapids $160 worth. Another dramatic comparison is the Chicago price of $2.09 for 5,000 tablets as compared with the Defense Supply Agency price of $4.50 for 5,000 tablets-while, DSA paid more than twice as much per unit for reserpine, its purchase totaled $31,410 compared to only $338 for Chicago.1 In the case of dextroamphetamine sulfate tablets-S milligrams, 1,000's-while Des Moines and Newark were paying $22.60 for this drug on a trade name basis, the city of Los Angeles was purchasing the drug generically for $0.53, a striking differential of 4,264 percent. And although the price paid for phenazopyridine-0.1 gram, 1,000's-by the cities of Erie, Indianapolis, and Winston-Salem was $48, Phoenix, Ariz., purchased the drug for only $4.13, less than one- tenth the price. Further, the survey generally indicates that the total dollar value of the contract involved is not a significant factor in price differentials. For example, in the case of chlordiazepoxide-25 milligrams, 5,000's-although Los Angeles bought quantities amounting to $32,- 625.86, they were paying $25.50 per unit while Philadelphia, which bought only $12,210.11, or one-third the quantity, was paying only $18.50. it is interesting to note, in this case, that both cities were supplied by the same firm. In the case of chloramphenicol-250 milligrams, 100's-the high- quantity purchasers are not the cities buying at the lowest prices. Al- buquerque paid $13.50 per unit and bought quantities amounting to only $718.20, while Cleveland, Ohio, paid $24.99-almost 50 percent more-and bought $2,249 worth, which is over three times the dollar value of Albuquerque's purchase. Both cities again purchased from the same firm. An even more shocking example is the case of phenazopyridine tablets-0.1 gram, 1,000's. The Veterans' Administration, a large Fed- eral purchaser, while buying 80 times the quantity of Phoenix-that is, 1 See charts beginning at p. 1745, Infra. PAGENO="0062" 1612 COMPETITIVE PROBLEMS IN THE DRUG INDUSTRY Veterans' Administration bought $31,648 worth as opposed to $413 worth bought in Phoenix-the V.A. is paying eight times as much as Phoenix for this product-$32.16 as opposed to $4.13. Further, although the Veterans' Administration and the Defense Supply Agency contracted for almost identical total dollar amounts for this drug, the VA paid close to three times as much as the DSA-$3'2.16/ $11.60-per unit, and yet both purchases were by the Federal Govern- ment, simply different arms of the Federal Government. Similarly, in the ca~se of chlordia.zepoxide, the VA purchased, in dollars, 35 times as much as the city of Philadelphia-strangely enough, Philadelphia got a better price-VA, $21.75,; Philadelphia $18.50-than the large Federal purchaser. There are numerous other examples of gross price differentials which defy any logical explanation. I have no doubt that any random sample of 29 other municipalities would produce the same results that both this and the New York survey have demonstrated. I doubt whether any amount of explaining can justify this quixotic pricmg structure. Furthermore, I do not see any evidence as of this date that the industry is even concerned about reviewing it~ pricing practices as suggested last week by George Squibb, former vice presi- dent of the Squibb Company. Obviously the cities and other public purchasers are going to have to review their purchasing practices if they expect to get a fair share for the taxpayers' dollars spent on drugs. The committee will be pleased to hear testimony from the PMA at any time on this subject matter. (Letters referred to previously follow:) THE CITY OF NEW YORK, DEPARTMENT OF PURCHASE, New York, N.Y., April 24, 1967. Mr. J. FORREST GEE, Purchasing Agent, Atlanta, Ga. DEAR MR. GEE: The New York City Department of Purchase is currently engaged in an intensive survey of prescription drug prices. We have selected your agency to represent the geographic area where you are located-along with 14 other areas across the United States. Attached to this letter is a list of 25 drugs that have been chosen by my staff as widely used and as pricing barometers. I would greatly appreciate your providing prices for each drug, in the dosage form specified, from your current contracts. If you do not purchase a certain item, or if you purchase in a different dosage form, please substitute an appropriately similar item and indicate that you have done so. Because of our urgent need for this information, we have ~ieen forced to undertake this project on an expedited basis. One of my staff members will be calling your office in a day or two, and we hope that with the list we have sent in front of you, you will be able to supply us with as much information as possible over the telephone. I apologize for the short notice we are giving you. I wish it were possible to collect the pricing information over a longer period of time. I can promise you, however, that whenever you require similar information for price corn- parisons my office will cooperate with yours to the utmost. Thank you in advance for your help. Sincerely, MARVIN GERSTEN, Acting Commissioner. PAGENO="0063" COMPETITIVE PROBLEMS IN THE DRUG INDUSTRY 1613 FULTON COUNTY, PURCHASING DEPARTMENT, Atlattta, Ga., April 28, 1967. Mr. MARVIN GEIOSTEN, Acting Commissioner, Department of Purchase, City of New York, New York, N.Y. DEAR MR. GERSTEN: Mr. J. Forrest Gee, Purchasing Agent for the City of Atlanta, Georgia, has referred to me your letter of April 24, requesting certain drug prices. Enclosed is the list of drugs which you furnished, llllecI in with Fulton County's prices. None of these prices goes pack further than the winter of 1965. Most of the prices are current. The items left blank are drugs not used by the County and for which we cannot substitute an appropriately similar item. I am glad to furnish you this informatiän and hope it will be helpful. Yours, very truly, F. T. FARRAN, Purchasing Agent. Item Your price 1. Chlordiazepaxide HCL (Librium) capsules: 25 mg., 500 in bottle $45.00 2. Diazepam (Valium) tablets: 5 mg., 500 in bottle 38. 00 3. Meprobamate (Miltown or Equanil) tablets: 400 mg., 500 in bottle 31.20 4. Chlorpromazine HCL (Thorazine) tablets: 50 mg., 5,000 in bottle 270. 00 100 mg., 5,000 in bottle `7. 50 5. Prochlarperazine HCL (Cnmpazine): Tablets, 10mg., 500 37.34 Tablets, 5mg., 500 28.79 Injection, 5 mg/cc., 2-cc. amp. 500's 240.00 6. Diphenylhydantoin (Dilantin) in ectable: 250-mg. vial 2. 27 7. Diphenylhydantoin (Dilantin Infantabs): 13.i gr 2 4.00 8. Sulfadiazine tablets: 0.5 mg., 100 in bottle .93 9. Sulfisoxasole tablets: 0.5 mg., 1,000 in bottle 8.75 10. Triple sulfa suspension (trisulfapyrimidines oral susp., U.S.P.) 16-oz. bottle 3. 18 11. Streptomycin sulfate solution: 5 gm./12.5 cc. per vial .88 12. Chloramphenicol (Chloromycetin): Capsules, 250 mg., 100 in bottle 21.00 Succinate, 1-gm. vial 7. 50 Ophth., 25-mg. vial 1.00 13. Tetracycline HCL capsules: 250 mg., 1,000 in bottle 59.00 14. Penicillin G potassium in.: 1,000,000-unit vial . 20 15. Ampicillin: Anhydrous or Trihydrate capsules: 250 mg., 500 in battle `18.50 500 mg., 500 in bottle Oral suspension: 250 mg./cc., 80 cc. in bottle 3.75 Sodium tar in ection: 1-gm. vial 3. 85 16. Nitrofurazone (Furacin): Soluble dressing: 1 lb 6. 12 Vaginal suppositories: 100/box 2. 10 Solution: 16 oz 5.57 17. Tolbutamide (Orinuse) tablets: 0.5 gm., 500/bottle 2 59. 24 18. Trivalent polio vaccine: 10-dose package 7. 20 19. Isoniazid tablets: 100 mg. 100/bottle 1.29 50 mg., 1,000/bottle 20. Sodium Para Aminosalicylate: Tablets, 0.5 gm., 750/battle 3. 00 Powder, 25-lb. drum 21. Prednisone tablets: 5 mg., 1,000/bottle 10. 50 22. Diphenhydramine HCL (Benadryl) capsules: 25 mg./1,000 7.00 50 mg./1,000 3.90 23. Reserpine tablets: 0.1 mg./1,000 in battles 33. 21 0.25 mg./1,000 in battles 13. 14 24. Dextroamphetamine tablets (Dexedrine): 5 mg./1,000 in battles 72. 60 25. Phenuzopyridine HCL(Pyridium): 0.1 gm., 1,000 in bottles 48.00 Per hundred. 2 Per thousand. 2 Box of 12. PAGENO="0064" 1614 cOMPETITIVE PROBLEMS fl~ THE DRUG INDUSTRY FULTON COUNTY, CoMMIssIONERS OF ROADS AND REVENUES, Atlanta, Ga., May 30, 1967. Subject: Prices paid by Fulton County for purchase of certain drugs. Hon. GAYLORD NELSON, Chairman, Monopoly Subcommittee, Committee oa Labor and Public Welfare, U.S. Senate, Washington, D.C. DEAR SENATOR NELSON: This will acknowledge your letter of May 19, 1967, requesting certain information on drug purchases made by the Fulton County government. Before answering your questions, I want to supply some back- ground information about the drug prices presented to your Committee. Fulton County is not a major purchaser of drugs. In 1966, which is a typical year, we spent only $22,282.43 for drugs for five County agencies-Haven Nurs- ing Home, the Department of Public Health, Fulton County Jail, Fulton County Juvenile Court and Detention Home, Fulton County Public Works Camp. We do not purchase drugs for Grady Memorial Hospital, which is the major hospital for indigent persons in Metropolitan Atlanta. The request for information on drug prices came from Mr. Marvin Gersten, Acting Commissioner, Department of Purchase, City of New York. It was sent to the Purchasing Agent of the City of Atlanta. Since the City purchases prac- tically no drugs, the questionnaire was sent to Fulton County. Actually Grady Memorial Hospital should have received this questionnaire since they are large users of drugs and purchase in large amounts. This would have given you a comparison between large drug purchasers. Grady Memorial Hospital is operated by an independent Authority which is financed by Fulton and DeKalb County governments. When our Purchasing Agent, Mr. F. T. Farran, received the questionnaire, he noted the urgent nature of the request. To quote from Mr. Gersten's letter, "Because of our urgent need for this information, we have been forced to under- take this project on an expedited basis. One of my staff members will be calling your office in a day or two, and we hope that with the list we have sent in front of you, you will be able to supply us with as much information as possible over the telephone." A copy of Mr. Gersten's letter is attached. Since it came to us through the City of Atlanta, we received it on the same day that the information was desired. A check of the 25 drugs contained on the questionnaire accompanying Mr. Gersten's letter revealed that Fulton County had purchased only 6 in the last 18 months. We compiled information on prices paid on these 6 drugs from our records and then, in order to complete the questionnaire, the Purchasing Agent called one of our local drug wholesalers and asked for quotations. No bids were taken, no purchases were made, and no public monies were expended as a result of these telephone inquiries. As it turned out, the prices received on the telephone were not, in most cases, competitive and obviously would not have been accepted by the County. Therefore, most of the information presented to you by Mr. Haddad did not represent actual prices paid by Fulton County for drugs purchased and, therefore, should not be used by your Committee as a basis for action. To turn now to the specific information requested in your letter, of the ten drugs listed in the attachment to your letter, Fulton County purchased only two: Tetracycline Hydrochloride 250 mg (1,000) .-On January 31, 1967, Fulton County accepted a low bid price of $35 per 1,000 for 7,000 tablets. The price of $59 per 1.000 shown on your tabulation is incorrect. The drug was supplied by Premo Pharmaceutical Company, which as we understand it manufactures the drug supplied. Bids were requested on a generic basis. D'iphcnhydramine Hydrochloride 25 my (1,000).-On January 31, 1967, Fulton County accepted a low bid price of $7.00 per 1,000 for 5,000 tablets. The drug was manufactured by Parke-Davis. Bids were requested on a trade-name basis (Benadryl). In answer to your other questions: (1) Does your city buy on a generic or a trade name basis? Fulton County buys drugs on both a generic and a trade name basis. A re- view of recent requisitions from County physicians indicates that about a's many drugs are requested on one basis as another. As our institutional physician, PAGENO="0065" COMPETITIVE PROBLEMS IN THE D'RTJG INDUSTRY 1615 Dr. William Reid explains, in certain situations he feels that certain `trade- name drugs are superior to their generic equivalents and in his judgment trade- name bidding is essential. In certain other instances, there is no generic equiva- lent for some of the drugs we purchase. This was obviously not true of the list attached to your letter. (2) What is the trade name and manufacturer of the drugs purchased? See information above. (3) Were sealed bids solicited or was purchasing done through negotiation? All drugs were purchased on the basis of sealed bids requested after public notice and advertisement. (4) What complaints were received regarding any drugs purchased during the past year? We have received no complaints regarding drugs purchased. (5) What was the volume of drugs purchased during the past twelve months? As noted above, during 1966 Fulton County spent $22,282.43 for drugs for County agencies. It is unfortunate that the purpose for which the information was requested was not made known to us. If this had been done, we would not have obtained information over the telbphone but would have used only bid information from our files. In addition, the time element was totally unreasonable. We were re- quested to supply information on the very day that the request was received. I apologize for the length of this reply but unfortunately there has been a great amount of misleading publicity regarding drug prices allegedly paid `by Fulton County and because of this I felt that a full explanation was justified. I would apreciate it very much if you would make this information available to the other members of your Committee and to the representatives of the press who attend the Sessions of your Committee in order that the record may be made clear regarding Fulton County purchasing practices. Sincerely, JAB. H. ALDREDGE, Chairman, Board of County Commissioners. Senator NELSON. As I said in the beginning, the New York City survey was attacked vigorously in various parts of the country by the Pharmaceutical Manufacturers Association. In reviewing the New York City survey, I have concluded that the attacks upon it were not valid with the exception of some minor points and one mistake in Fulton County which was not the fault of New York City and not the fault of the witness before this committee, Mr. 1-Iladdad. The rest of the survey stood up very well in our cross checking directly with the cities involved, but in order to avoid any suspicion of the survey at all, we conducted this survey out of my office with 29 cities and municipalities. I think that since the first survey has come under such attack by the Pharmaceutical Manufacturers Asso- ciation, they are entitled to, and ought to, respond to this survey. And if they are not prepared to respond today, we would be happy to set a date for the Pharmaceutical Manufacturers Association to respond. STATEMENT OP LLOYD N. CUTLER,, SPECIAL COUNSEL, PHARMA- CEUTICAL MANUPACTURERS ASSOCIATION, WASHINGTON, D.C. Mr. CUTLER. May I make a brief comment, Mr. Chairman? \Ve just received this schedule and~ your statement when the hear- ing began, and clearly we cannot comment now on whatever factual issues there may be between the committee study and our own evidence. Perhaps two of our economic witnesses, however, Professor Mark- ham and Professor Firestone, may be able to answer questions you wish to put to them as to the conclusions, some of the economic con- Si-280-68-pt. 5__-_5 PAGENO="0066" 1616 COMPETITIVE PROBLEMS IN THE DRUG INDUSTRY clusions, you draw from this data, assuming that there is no factual dispute between us. I think I should mention though, sir, that the table that you have submitted is not a comparison of prices paid for the same drug in the sense of the same drug product offered by the same manufacturing source to two different buyers. Senator ~\ELSON. Several instances are. Mr. CUTLER. They may be. but. the table does not disclose that. The table simply gives a generic, name of a drug and then lists prices. Senator Nni~sox. Just so the record will be clear, let me remind von that I recited in my statement. those cases where the same seller, the same manufacturer, sold drugs for a dramatically different price to the two different buyers. Mr. CUTLER. I don~t recall any such statement in the 1)1eI)ared state- ment von lust. delivered. Senator Nelson. You make a aeiuerai stale- ment which is doubtless correct that. at the bottom of page 3. neither does the source of the drug explain the difference since the same sup- pont frequently charges different prices to different buyers of the same drug. That may well be. But. the cases you give are cases in whicl~ von identify a generic, name for a drug, which may have been offered by different buyers. In fact you indicate in some of the illustra- tions tha.t they were from different sources. Senator NELSON. That is correct. Several cases are pni~c1iases from different, buyers and in the very dramatic one where it was S0.51 versus $17 or thereabouts. one was generic and the other one was a. tradename drug. But I did read this sentence and I will read it again "For example. in the case of chlordiazepoxide (25 lug., 5,000's) al- though Los Angeles bought quantities amounting to $32,625.86," tli is is on page 5. `worth, they were paying $25.50 per unit while Pliila- delphia, which bought only $12,210.11 worth or one-third the quantity, was paying only $18.50. It is interesting to note in this case that 1)0th cities were supplied by the same firm." Mi'. CUTLER. rfliat is correct, sir. Senator NEIsox. Now. there are several examples of that iii the survey we had. There are limitations on how much information von can put on a chart. What. the survey indicates, among' other thing's, is that in several in- stances the cities have bought from generic companies at a dramatical- lv cheaper price than from brand name companies. It has ln(licatesl that in many instances the same supplier, in the same year, will sell his drug, brand or generic, to different buyers at a dramatically differ- ent. price with no relationship to quantity pui~ciuased. In fact, in the instance I just read, the purchaser of three times as much paid a sub- staiutial percentage more. What it indicates is that there doesn't seem to be any rationale to it. We hadl the Defense Supply Agency buy- ing two or three hundlred thousand dollars worth in less than a year's tIme paying more than a city that bouigltt a very minor amount. The survey confirms what the first. demonstrated: that is, that. there seems to 1)0 no rationale at all in the pricing practices of the dm2 companies. But. I will be happy to have any explanation for these differences from the. drug industr. The economists have addressed themselves in PAGENO="0067" COMPETITIVE PROBLEMS IN THE DRUG INDUSTRY 1617 great detail to the intricacies and economics of this industry, and I would appreciate some explanation from them as to this situation. Mr. CUTLER. Some of today's witnesses may be able to comment on the conclusions you draw but as to any factual dispute between us we would like an opportunity to study this chart and to respond to you at our next session. Senator NELSON. Fine. As I understand it, Mr. Cutler, you desire to have Prof. Paul Coot- ner, Mr. Plot.kin, Mr. Gordon Conrad, and Professor Markham on a panel at the same time? Mr. CUTLER. Yes, sir. Senator NELSON. And would you identify the witnesses for the record? Mr. CUTLER. Yes, I shall. We have three groups of witnesses for you today, Mr. Chairman, to testify as to various economic issues that have arisen iii these hearings. The first group consists of the four gentlemen you have just named who will testify to the relationship they find as economists between the existence of above-average risk and the existence of an above-average rate of return in that industry, and to the presence of above-average risk in the pharmaceutical industry. These gentlemen are, first, on my immediate right, Prof. Paul Coot- ner who is a professor of finance at the Massachusetts Institute of Technology. To his right, Mr. Irving Plotkin, of the Arthur D. Little Co., in Boston, the consulting economist. To my far left, Mr. Gordon Conrad, also of Arthur P. Little, and to my immediate left, Prof. Jesse Markham, professor of economics at Princeton. We would like to begin with a brief statement of the theory of the relationship between risk and return by Professor Cootner. If we may, since all of these initial four statements are quite short, Mr. Chairman, we would like to have them delivered as a group, and then have the entire four submit to questions rather than have ques- tions put to one which could better be answered by another. But we will, of course, abide by your wishes in the matter. Senator NELSON. I have tried this so many times. It means we have to go back and reread and recheck the statements and if I don't have time to write down a question that occurs to me at the time the witness is testifying, I can't remember it later. But, if I ask a question of any one of the witnesses which someone else is better qualified or prepared to answer, I am prefeetly happy to have that person respond to the question, and if that question is more applicable to a subsequent witness I would be glad to set it aside. But I would rather have the record show the questions and answers in the text as I go along, because it makes it easier to read the record and it makes~ it possible for me to ask the questions which subsequently I may forget to ask. Mr. CUTLER. All right, sir. ~Senat9r NELSON. If someone else can answer the question just let the witness identify himself for the record so that the record will show who is answering the question. Mr. CUTLER. Right. We would like to begin then with Professor Cootner, of MIT. PAGENO="0068" 1618 COMPETITIVE PROBLEMS IN THE DRUG INDUSTRY STATEMENT OF PAUL H. COOTNER, PH. D., PROFESSOR OF FINANCE, MASSACHUSETTS INSTITUTE OF TECHNOLOGY, CAMBRIDGE, MASS. Senator NELSON. I am very pleased to welcome this distinguished panel of economists. The committee recognizes these gentlemen as men with distinguished professional credentials and we are pleased to have you before the committee today. Professor, your short biographi- cal sketch will be printed at this point. (The biographical sketch of Dr. Cootner follows:) BIOGRAPHICAL SKETCH OF PAUL H. 000TNER, PH. D., PROFESSOR OF FINANCE Born: May 25, 1930. Degrees: B.S., Florida, 1949; MA., Florida, 1950; and Ph. D., Massachusetts Institute of Technoiogy. Field: Finance, Economic Theory. First appointed: 1959. Professional experience: 1955-56: Ford Teaching Intern, Brown University. 1956: Assistant Professor, Northwestern University. 1956-59: Research Associate, Resources for the Future, Washington, D.C. 1959-62: Asst. Professor of Finance, Sloan School of Management, M.I.T. 1962-66: Assoc. Professor of Finance, Sloan School of Management, M.I.T. 1966 to date: Professor of Finance, Sloan School of Management, M.I.T. Honorary and professional societies: American Economic Association, Phi Beta Kappa, Phi Kappa Phi, Econometric Society, and American Finance Association. Publications: "Speculation in the U.S. Government Securities Market", Commission on Money an4 Credit, 1963 "Competition in U.S. Commercial Banking", Commission on Money and Credit, 1963 The Random Character of Stock Market Prices, Edited by P. H. Cootner, M.I.T. Press, Cambridge, Massachusetts, 1964 Water Demand for Steam Electric Generation (with G. Löf), Resources for the Future, 1966 "Social Overhead Capital and Economic Growth", The Takeoff into Su- stained Economic Growth, 1964 Mr. GROSSMAN. Excuse me, Dr. Cootner, I notice that in looking over the statements of the other witnesses that they all, in their first para- graph or so, mention that they are either on retainer or employed by the PMA and I don't notice this in your statement. Are you appearing independently? Dr. COOTNER. No, I am employed by the PMA. Mr. GROSSMAN. You are retained by the PMA? Dr. COOTNER. Yes, sir. Actually it is not exactly true. I am employed by Arthur D. Little. Mr. GROSSMAN. Who is in turn employed by the PMA? Dr. COOTNER. Yes. Senator NELSON. Does Arthur D. Little have a contract to represent the PMA? Mr. CUTLER. Yes. The PMA has consulted Arthur D. Little, Inc., to prepare economic studies relating to these hearings and in the course of that work with our approval they have in turn retained as consult- ants Professors Cootner and Markham who are about to testify here. Senator NELSON. Is there any objection to informing the committee as to the amolmt of the consultant fees of the witnesses? PAGENO="0069" COMPETITIVE PROBLEMS IN THE DRUG INDUSTRY 1619 Mr. CUTLER. We will be happy to submit to you whatever you would like in that regard; yes. (The information referred to, subsequently received, follows:) PHARMACEUTICAL MANUFACTURERS AssocIATIoN, Washington, D.C., February 9, 1968. Hon. GAYLORD NELSON, Chairman, Monopoly Subcommittee, Senate Small Business Committee, U.S. Senate, Washington, D.C. DEAR SENATOR NELSON: At page 2695 of the transcript of the hearing `before the Subcommittee on Monopoly of the Senate Small Business Committee on De- cember 19, 1967, YOU asked that we provide YOU with infor~mation concerning the consultant fees received by the economic witnesses who presented testimony on that day at the request of the Pharuiaceutical Manufacturers Association. The fact that the PM'A compensated these distinguished economists for con- ducting studies and for preparing and presenting testimony is, of cour~e, legit- imate public information and was fully disclosed by PMA kt the outset of the hearing. We do not, however, see any legislative purpose to be `served by dis- closure of the amounts paid to particular individuals. Nevertheless, in view of the request you m1ade `at the hearings, the information is set forth below. Mr. Gordon Conrad and Mr. Irving H. Plotkin `of Arthur D. Little, `Inc., Pro- fessor Paul H. Cootner of the Massachusetts Institute of Technology, and Pro- fessor Jesse Markham of Princeton University were retained as economic wit- nesses through Arthur D. Little, Inc. We are advised by Arthur D. Little that the fees billed for preparation and `delivery of the testimony, and for the exten- sive `studies prepared and presented by the witnesses were computed at the nor- mal rates generally charged for consultation by these economists. Mr. Conrad and Mr. Plotkin worked for a number of months on the preparation of `the detailed studies they presented at the hearings. Professors Cootner and Markham also spent a silbstantial a:mount of thne in reviewing the `studies and other per- tinent material and in the preparation of their testimony. The a'm'ounts billed for the work described above were as follows: Mr. Gordon Conrad, $13,100; Mr. Irving H. Plotkin, $10,250; Professor Paul H. Cootner, $6,450; and Professor Jesse Markham, $4,500. On December 19, PMA also presented testimony by two kdditional economic consultants, Professor John M. Firestone of City College of New York, and Pro- fessor Simon N. Whitney of New York University. Professor Firestone spent several days and Professor Whitney several weeks in preparing material for the Association and his testimony. The a~noiint's `billed for their `services were as follows: Professor John M. Firestone, $875 and Professor Simon N. Whitney, $4,250. We `are in the process of obtaining certain other information you asked for during the course of the December 19 hearing. We cannot, however, furnish the comments you requested at page 2688 of the transcript on the comparative price study you introduced at the beginning of the hearing until we are permitted to examine the data in your possession from which the study was prepared. I am enclosing a copy of my February 6 letter1 to you on this subject and most earnest- ly request that you reconsider your decision and instruct your staff to make the data available to us. Sincerely yours, C. JOSEPH STETLER. Senator NELSON. GO ahead, Professor Cootner. Dr. COOTNER. Mr. Chairman, and members of the subcommittee. My name is Paul H. Cootner and I am a professor of finance at the Mas- sachusetts Institute of Technology. My, specialty at MIT is the study of the effects of risk and uncertainty on financial decisionmaking of all kinds. I am here to speak generally on research conducted by myself and others on the effect of risk on the rates of return required to induce investment. 1 Retained In committee files. PAGENO="0070" 1620 COMPETITIVE PROBLEMS IN THE DRuG INDUSTRY The relevance of a discussion of risk and rate of return to a discus- sion of pricing in the drug industry may not be immediately apparent, but I think it soon becomes so if we note that any reduction, voluntary or enforced, in the prices charged by the drug industry without simul- taneously reducing its costs will flow through to profits immediately. If prices are reduced across the board by administrative fiat, profits will also be reduced. Now, I do not appear here as an expert on the drug industry, either with regard to its pricing policy or the riskiness of its investments. It requires no such expertise, however, to establish the elementary prop- osition that for an industry selling an item necessary to life and fac- ing an essentially inelastic demand, changes in price lead immediately to changes in profit in the same direction. It is the drug industry's role as a supplier of necessities that apparently occasions this committee's interest in its pricing policy and it is that same role that establishes the close relation between price~s and profits. Senator NELSON. May I raise just a point here? You stated that it requires no expertise to establish the elementary proposition that for an industry selling a product necessary to life and facing an essentially inelastic demand, changes in price lead im- mediately to changes ~ p~t in the same direction. I wonder if that can really stand as an unqualified statement? For example, if a drug is discovered which does something that no other drug does, al- most immediately-that is, after some advertising, some understand- ing on the part of the medical profession-that drug then, if it is a valuable drug, will he in demand in every single case where a patient needs that type of drug. In fact, drugs are more in demand than other products, because health as a general proposition is a prime factor motivating people in pain. So the demand may be inelastic to the extent that every single person who needs a drug is willing to pay for it. It is not true to automobiles or houses or clothes. So I don't quite follow the statement that the demand is inelastic. Dr. COOTNER. Well, as a matter of fact, the description you gave is exactly what I would consider inelastic, that is, that the demand is there regardless of price, and I think we are in perfect agreement on the facts. Senator NELSON. I thought you were using the argument to prove that this was some kind of a handicap to the industry. Whereas, as matter of fact, if there is any other industry that faced this kind of inelastic demand it would be considered a great asset to the industry, wouldn't he, his product? Dr. COOTNER. Well, that depends, if he had an inelastic demand and he had monopoly power lie would certainly like that sort of situation, but inelastic demand in and of itself is no advantage. For example, many people believe, and I think it is probably true, that the American farmer faces essentially an inelastic demand for the same reason, he does not consider that an advantage because it means, as a matter of fact, if the supply increases he is forced to take a substantial cut in price. Senator NELSON. Not if lie had a patent on potatoes. We will get into that some later, but I misunderstood what you were saying al3out it. PAGENO="0071" COMPETITIVE PROBLEMS IN THE DRUG INDUSTRY 1621 An inelastic demand in the case of a valuable drug means a total de- mand by every single person who needs it. That is not true, to my knowledge, of any other single product in the marketplace in America. Dr. COOTXER. But all I am simply saying is that the drug industry, the total sum of all drugs is very much like the total food industry, total sum of all food. Any individual product, of course, any individual food product is not necessary to existence, but all food products taken together are necessary. But I don't think we are in any substantial dis- agreement. Senator NELSON. I see. Dr. COOTNER. Drugs are necessities, and it is precisely because the drug industry is a supplier of necessities that I suspect that this sub- ject is of such great interest to the committee. Senator NELSON. But there are all kinds of alternative choices in food. There is such a great variety, you can be a vegetarian or not, you can consume any number of a thousand different items. However, if a drug is discovered that performs an important health function, and it is patented, there is an inelastic demand in the sense that every single persoi~ in the United States who needs it will want it and will pay for it, and it is shielded from competition for 17 years. So I just want to get that sharp and, I think, significant economic distinction into the record at this time, unless you disagree with me. Dr. COOTNER. No. I will continue: My role in these hearings is to relate this price-profit nexus to its implications for industry's ability to engage in risky, in- novative activity and continue to attraCt capital. I will be establishing the meaningful relationships that empirically exist between industry risk and the rate of return required by investors. It will be left to sub- sequent witnesses to establish the drug industry's role in this process. In an economy such as ours, which is largely directed by a large number of firms and households, all responding to opportunities and hopes for financial gain, it is inevitable that some will prosper and others will fail. When opportunities arise, men will jostle each other to seize them and it is unlikely that they will all share the gains equally. Some, hopefully, due to skill will choose a better or faster route to satisfying a need and earn a better return in the process. In our relatively unregulated capital markets, investors are free to back the enterprises that they judge will be most successful. In doing so, they must inevitably make judgments about the management of the enterprise, the likelihood of success, and the prospective rate of return. They not only evaluate the particular enterprise but they must compare it with all other relevant alternatives. Now clearly, an investor would like to earn a good rate of return on his money. It does not follow, however that he will always put his money to work at the highest rate of return available. The U.S. Treasury is grateful for this "quirk" of investors, for while Govern- ment bonds offer very high yields by historical standards, those interest rates are substantially lower than investors could get from buying the bonds of corporations and still less than those available from bu ving corporate common stocks. Investors, for the most part, do not accept these lower yields out of charity or patriotism. They buy U.S. bonds because those bonds offer PAGENO="0072" 1622 COMPETITIVE PROBLEMS IN THE DRTJG INDUSTRY them something that corporate bonds cannot-immunity from the fear of bankruptcy. In short, such investors are willing to sacrifice gain in order to avoid risk. When an industrial entrepreneur forms a new firm, or finances a new investment, he must attract equity and debt capital from either his own resources or those of the public. He must convince himself, or other investors, that the prospect rate of return on the capital will be adequate. In this industrial investment, as in the purchase of bonds, the adequacy of the rate of return must be judged in the light of the risks involved. A man will not prospect for gold or oil at the same rate of return that suffices for an electric company, nor will he finance the development of some exotic new technology for the same return as on shoe manufacturing. Senator NELSON. May I interrupt here for a moment? Dr. Coo~n. Certainly, sir. Senator NELSON. Your statement that a man will not prospect for gold or oil at the same rate of return that suffices for an electric company; nor will he finance the development of some exotic new technology for the same return as he could get on shoe manufacturing, he would do it for the same return if the risk were the same, wouldn't he? Dr. COOTNER. If the risk were the same, yes, sir; that is right. Senator NELSON. So, do you address yourself to the risk question? Dr. COOTNER. Right. I chose the particular cases because I thought and, perhaps, wrongly, that they were prima facie cases where the risk was greater than in the alternative. That is, I assumed that prospecting for gold was an obviously more risky, alternative than electric utility investment. Senator NELSON. The real test then it seems to me is the consistency of profits, and the consistency of the companies in the industry, their capacity to stay in business and make a profit, is it not? Dr. COOTNER. That is right. It is the question of the relative risk; yes, sir. Senator NELSON. Are you going to give us the statistics for the past few years on the profits in the drug industry and also examples of the drug companies that have gone out of business because the risk was so high? Mr. CUTLER. Mr. Chairman, those sul'jects are going to be covered by the witnesses who follow immediately after Professor Cootner dem- onstrates the relationship between risk and return. Senator NELSON. I read all the statements and I didn't see anything in any of them that related to the speci~c question that I have raised here as to how many companies have gone out of business in the past 10, 15, or 20 years, a.nd how their profits have compared in the past 10, 15, or 20 years with profits in industry in general in the country. I saw the statements, but no specific examples, and I raise the question at this time so that the other economists will have a chance to think about it while Dr. Cootner is testifying. Dr. ~OOTNER. I will proceed. It can be empirically demonstrated that this is so, but it is more than a passing fancy of this year's crop of investors, or even this century's crop. This attitude toward risk is deeply imbedded in our legal code. Thus, while trustees of an estate PAGENO="0073" COMPETITIVE PROBLEMS IN THE DRUG INDUSTRY 1623 are enjoined to invest the funds entrusted to them as profitably as they can, they are judicially warned against seeking return at the expense of taking "imprudent" risk. And the U.S. Supreme Court, in 1909, stated the problem clearly and forthrightly for the first, but not the last, time with respect to regulated corporations. "The less risk, the less right to any unusual returns upon the in- vestments. One who invests his money~ in a business of a somewhat haz- ardous character is very properly held to have a right to a larger return, without legislative interference ~ * As a matter of fact, the U.S. Supreme Court is rather a new- coiner to the discussion of economic risk. More than two centuries ago, Bernoulli, a great student of probability and risk, noted that the ordinary man would usually reject a highly favorable bet if that bet required him to take risks which were large in relation to his capital. Like the genius he was, Bernoulli used this observation to construct the first technical tools for the evaluation of attitudes to risk, tools that have been added to by many of the important econ- omists and statisticians who have concerned themselves with dynamic problems of economic growth. The list of economists who have~ concerned themselves wit1h the question goes back to Adam Smith and reaches forward to embrace almost all of the important economists of this century of yin ually every persuasion: Joseph Schumpeter, Irving Fisher, Frank Knight, John Maynard Keynes and including such diverse contemporaries as Paul Samuelson, Kenneth Arrow, Milton Friedman, and George Stigler. While all of these writers have dealt with this problem, two of them, Keynes and Schumpeter, deserve special note. Keynes' major contribution from an economist's point of view was his theoretical results, but he was also a very practical man, ~ho made (and lost) several fortunes as a practicing investor. As a practical man, Keynes was the first that I know of to make an empirical measurement of atti- tudes toward risk. Like most first attempts, it was a crude attempt, but he concluded that merchants carrying large inventories of raw mate- rials required a premium rate of retUrn of 10 percent of capital, over and above normal rates of return, because of their exposure to the risk of price fluctuations. Surprisingly enough, work that I and others have done in the last few years, using more sophisticated mathematical tools suggests that Keynes' crude estimate is not far off at all. Schumpeter's work was of a much different type but of direct rele- vance to the questions faced by this, and other investigative commit- tees. Schumpeter was deeply concerned with dynamic problems of eco- nomic growth, and the role of risk taking and entrepreneurship on innovation and investment. It was he, more than any other writer, who stressed the fact that the willingness, to innovate played a vital role in the rapid growth and efficient operation of Western economics. He also was quick to draw the concomitant conclusion: that any regulations or legislation that inhibit risk taking, also inhibit innovation and the rate of economic growth. Now I hasten to add, that neither Schumpeter nor I, nor. indeed any responsible economist, will argue that industrial abuses should not be corrected, when found. The important elements are (1) that one should PAGENO="0074" 1624 COMPETITIVE PROBLEMS IN THE DRUG INDUSTRY not be surprised to find large average profits in risky enterprises or to take such profits as evidence of abuse of market power, and (2) if abuses are found, one must take care to eliminate only excessive, and not necessary returns on investment, if one does not wish to strike at the well springs of risk taking and growth. Senator NELSON. I don't know whether this is an abuse or not, but I will give you the facts and ask for your comment. Rhone-Poulenc, a French firm, discovered chlorpromazine. They licensed a company in the United States to produce it, and they licensed a company in Canada to produce it, each of them with the exclusive market in their respective countries. So neither the company in the United States nor the company in Canada spent any money on research. It was just a question of each one of them having an exclu- sive market, both in the same continent with adjoining borders. The price charged by the U.S. licensee for 25 milligram tablets to the Defense Supply Agency was $32.62 a thousand. The price of the Canadian Licensee to Canada's Department of Veterans Affairs was $2.60 a thousand.' Ca.n you explain to me the economics of that? Dr. COOTNER. I obviously can't, sir. I haven't done any study of the drug industry. These figures, of course, are very much `like the figures you presented at the beginning of the statement~, and I trust that the PMA will deal with this subsequently, but I have no expertise in the drug industry and I couldn't comment on it. Senator NELSON. I didn't think I was asking a question that re- quired expertise. This is general economic theory. Under what eco- nomic theory would you justify vast price discrepancies between two licensees of the same product, neither of whom did the research, both on the same continent, with adjoining borders. Why is the Canadian Government paying $2.60 a thousand for tablets, and the American Government $32.60 a thousand. There must be some economic theory under which each of these firms can succeed in charging their respective prices and both still makea profit. * Dr. COOTNER. Well, If I hadto comment I wOuld have to speculate but, of course, there is always the possibility that there were differ- ences in costs of the two firms or differences in quality. If there were no differences in costs or quality you would have to find something else to explain it if you could. But the point is I cannot comment on it without knowing more about the difference in costs or quality. If they are identical products then I think it would require some ex- planation. Senator NELSON. WTell, they are licensees for the same product, and used all the background research and information that Rhone-Poulenc had. Could the differential be explained by the fact that each one of them has exclusive control of the market. There is no competition in either `country, and yet one charges almost 15 times as high a price as the other. The industry keeps saying, among other things, that, "We need money for research." Well, they didn't do the research on this product, and I assume both the Canadian and the American com- panies are making a profit. See chart, p. 1746, Infra. PAGENO="0075" COMPETITIVE PROBLEMS IN THE DRUG INDUSTRY 1625 Mr. CUTLER. Mr. Chairman, Professor Cootner is clearly not quali- fied to answer on this and perhaps I am not either. But I do know some of the factual assumptions put into your question are not correct. I do know, for example, that Smith Kline & French did a very large part of the research needed to prove that Thorazine would be useful as a mental illness drug. That Smith Kline & French did a very large part-in fact all of the U.S. work, necessary to prove the safety' and efficacy of that drug and put it through the new drug approvals of the Food and Drug Administration. There are differences. I am not competent to say how deep they were or how significant they were, but~ this witness isn't either and if you want further explanation on this point, we will be glad to sup- ply it. Senator NELSON. Well, the research you are talking about was the clinical testing and other studies required of them to qualify for a New Drug Application. Mr. CUTLER. I believe it goes deeper than that. I believe, although I am not specifically informed on this subject, that Rhone-Poulenc had a chemical entity which it patented, but that its value in mental illness was largely the discovery; certainly it wits perfected by Smith Kline & French in this country. Mr. GORDON. In a study made for the Pharmaceutical Manufac- turers Association by Paul DeHaen, which was printed in the Amer- ican Professional Pharmacist in July of 1964, Mr. DeHaen listed Thorazine, originator, Rhone-Poulenc, France, marketed by Smith Kline & French in the United States. Also in the Kefauv.er committee's hearings, part 16, pages 9024, 9025, and the following pages, Dr. Heinz Lehmann, of the Verdun Protestant Hospital, Montreal, Canada, clearly states that the new uses for the drug were developed there. Mr. CUTLER. Mr. DeHaen, as I understand it, has another column next to originator called developer. Mr. GORDON. That is right. Mr. CUTLER. And that column has Smith Kline & French, but neither of us really has the facts at hand. Why don't we get the facts? My understanding is that Smith Kline & French did a very large part of the work required to prove the value of this d.rug in mental illness. Mr. GORDON. Incidentally, it may be interesting to know that I called Mr. DeHaen and asked him why lie changed the word "mar- keter" from his earlier study to the words "marketer and developer" in the later study, and he said it was done at' the request of the Pharmaceutical Manufacturers Association. Mr. CUTLER. It may be a more accurate description. Senator NELSON. Well, if there is an explanation of why the same drugs are being sold, one of them for almost 15 times as much as the other, the committee would appreciate having it.1 Go ahead. Dr. COOTNER. Thank you, sir. For, if one reduces the rate of return below that commensurate with the risk involved, that industry will no longer be able to attract the 1 See app. III beginning at p. 2129, infra. PAGENO="0076" 1626 COMPETITIVE PROBLEMS IN THE DRUG INDUSTRY same volume of capital or investors will seek to reduce their exposure to .risk by a greater conservatism or sluggishness in pushing down uncharted paths. But if I tell you that it is important to distinguish between exces- sive rates of return and those necessary to attract venture capital, I do you no service unless I can tell you how to make that distinction. The fact that the U.S. Supreme Court has ruled that return must be com- mensurate with risk in a whole string of important cases has not ended the regulatory debates on rates of return. Federal Judge 1-larold Leventhal has published a masterful, and lengthy summary of the legal problems in the regulatory context in the Yale Law Journal of May 1965. My earlier quotation was from 1Viilcox v. Consolidated Gas Com- pany. In 1923, in a case involving Southwestern Bell Telephone Co., Justice Brandeis explicitly referred to the need for a firm to receive "an allowance for risk incurred" if it were to attract capital. In Blue- field Water Works (1923) and Hope Natural Gas (1944) the Court again insisted that returns are to be concomitant with risks undertaken. To quote from the Blue field Water TVorks case: A public utility is entItled to such rates as will permit it to earn a return on the value of the property which it employs from the convenience of the public equal to that generally being made at the same time and in the same general part of the country on investment in other business undertakings which are attended by corresponding risks and uncertainties- That is emphasized by me- but it has no constitutional right to profits such as are realized or anticipated in highly profitable enterprises or speculative ventures. But despite these injunctions, courts have had difficulty implement- ing them because there was no work done to make the concej~ts opera- tional. To properly address this question requires an expansion of the quantitative approach of Lord Keynes on a more sophisticated basis. It requires, first, the measurement of risk and, second the establish- ment of a relationship between that risk andthe required rate of return. Economists have made attempts to solve these problems although most of these attempts have been of recent origin. First of all, this work was hindered by the slow development of the necessary statistical tools. But even after those tools were developed, further work had to await the general availability of the large amounts of statistical dat.a necessary to establish propositions in which probabil- ity and risk are involved. For example, one of the earliest of the re- cent studie.s was a test by PrOf. George Stigler, of the University of Chicago, which, despite some very imaginative attempts to extract in- formation from inadequate data could not extract enough to get statis- tically significant results. Good data first became available not for manufacturing industry but for financial markets. In 1959, Lawrence Fisher explained interest rate differentials on corporate bonds by risk varia.bles, thus measuring the extra rate of return asked by bond in- vestors for investing in more risky companies. In 1960 1 did related work on holdings of commodity inventories related to the earlier work of Keynes. By the time Prof. Daniel M. Holland and I turned to the ques- tion of industrial risk in 1962, we had unusual access to a large body PAGENO="0077" COMPETITIVE PROBLEMS IN THE DRUG INDUSTRY 1627 of statistical data and could derive such significant results. Since our study the availability of data has radically improved and studies have i'nultiplied. Fred Arditti of Rand Corp., Prof. Shannon Pratt of Indiana University and Portland State College, and Irving Plotkin and Gordon Conrad who you will hear after me, among others, have all completed major investigations into this question since 1964. One of the surprises of all these recent studies-or at least it might surprise those who have heard economists disagree in public testi- mony-is that they all reach agreement on the broad outline of their results. The rate of return does rise as risk increases, and even the quantitative relationships show only minor differences. There are few settled empirical truths in economics, and I am sure that differences will arise and refinements will be made as time goes on and work is reviewed, but the competition is no longer between different "be- liefs" or "feelings," but a comparison, of competing hypotheses and empirical tests. Since the Conrad-Plotkin study is very similar to Professor Hol- land's and mine through quite independently conceived, and will be presented in some detail, I am only going to discuss some of the main problems and implications of this type of work before you hear from Mr. Plotkin. One of the greatest obstacles to a scientific study of the relation between risk and return is the difficulty of establishing a meaningful, objective definition of "risk." Consider an investor planning to pro- duce and sell a particular product. In choosing the size of the invest- ment he must estimate, among other things, the size of the market, his level of costs, and the competitive response of others in the industry. In making his plans he will draw on whatever experience he has or on any market research he can perform. Such analysis may lead him to some "best guess," or it may permit him to place upper or lower limits on the potential demand, but there will always remain a large number of possible outcomes which have at least some probability of occurring. There will be similar uncertainty about the levels of costs that he will experience, or the price or production response of his competitors. In making his final plans, he must weigh all these pos- sibilities in estimating the profitability of the investment; whether lie does so by "intuition" or "judgment," or by mathematical calcula- tion. Finally, after making that estimate he must decide whether the venture is worthwhile. The whole theory about the relation between risk and rate of re- turns on this final decision. In essence, it turns on the belief that if our investor should examine investments in two different industries with identical return, he would prefer the investment which was least likely to result in a substantially reduced return if adversity befell the proj- ect. If this is so we will find that the more risky project will not be undertaken until the need becomes so pressing that the investor can anticipate a significantly higher rate of return than he could on the less risky project. If returns are out of. line with risk, the flow of capital into the riskier industry will be cut back until profits return to the desired relationship. In that case, whenever capital is flowing into both industries, we will observe the more risky industry earning a higher rate of return. Conversely, if the more risky industry fails to PAGENO="0078" 1628 COMPETITIVE PROBLEMS IN THE DRUG INDUSTRY earn a higher rate of return it will not be able to attract new capital and will decline or stagnate. Mr. GORDON. You are not saying that because an industry earns a higher rate of return it is necessarily an indication that it is a riskier industry? Dr. COOTNER. No, sir; I am not saying that. Mr. GORDON. All right. I only wanted to get that clear. Dr. COOTNER. This is only a hypothesis and it could only be tested if we quantify the "risks" I just outlined in words. Now it turns out that words are much less precise than mathematics, but much easier to understand. When we turn to implementing this verbal definition we find that there are a variety of precise formulations of the result. Some of these were used in my study with Professor Holland and others are described in detail in the Conrad-Plot.kin paper you will soon hear about. No one of these acceptable measures is perfect or ideal, but what is encouraging is that the relationships which are observed are all very similar-that is, they are not affected by the particular definition used and all lead to similar conclusions. This basic conclusion is that as risk rises so does the required rate of return. I should stress that this is an average relation-a statistical re- lation that is subject to chance variation. While we develop a mathe- matical expression which relates the two concepts, every industry is not described perfectly by the relation. But subjecting this relationship to all the normal satistical tests confirms the result. The indicated rela- tionship is a real one; it is not spurious or accidental. Finally, I should stress that this is not a complete theory of corporate profit. There are other risks besides the ones I have investigated. Just to name one, the marketability of an investment in Government bonds is substantially greater than one in plant and equipment and this alone would require a greater return for industrial investment even if it was equally risky. Furthermore, there are market imperfections which can influence particular observations. But none of these things affects the validity of the partial relation measured here. 1-laying stated these results, I would like to spell out their implica- tions. First of all, the relationship is an "average" or "long-run" phenomenon. It is particularly true in risky enterprises that things may not turn out as one expects and this accounts for a good deal of the scatter of points about our line of best fit. Low returns which arise from the impact of such risks will not immediately affect investment any more than the outcome of any particular flip of a coin changes the outlook for future flips. However, if one were to decide, by admin- istrative action, to reduce the average rate of return in a risky, competi- tive industry without at the same time reducing the risks, we would find an immediate impact on that industry's investment policy. Mr. GORDON. You say on page 9 that if one were to decide by ad- ministrative action and so on and so forth. First of all, practically all the testimony before our Committee indicates there is very little price competition in the drug industry. Is that your understanding? Dr. COOTNER. I am not qualified to comment on that; Mr. CUTLER. Could that question be addressed to Professor Mark- ham, Mr. Gordon, when he testifies as soon as we finish on this risk- return relationship? PAGENO="0079" COMPETITIVE PROBLEMS IN THE DRUG INDUSTRY 1629 Mr. GORDON. Well, this is important to what Dr. Cootner is say- ing, but let me go a little further then. He says he doesn't know. Let's accept that. Now, the clause "without at the same time reducing the risk," it seems to me misses the whole point. The risks in drug marketing al:e largely derived from existing market institutions and if these are modi- fied the risks will be reduced by the action of the same forces. If prices are lowered, say, by introducing vigOrous price competition or by Government action and the spread between cost and price is reduced, U lien there will be less money spent on advertising and promotion. This will in turn reduce the risk of obsoleting certain drugs. This particular point was discussed more fully by Prof. Henry Steele of the University of Houston in his presentation on February 14, 1964, before the Special Committee on Drug Costs and Prices of the Canadian House of Commons. I ask, Mr. Chairman, that pages 2482 to 2484 of Professor SUeele's brief be included in the record at this point. Senator NELSON. They will be included in the record. (Data referred to above follows:) [Excerpts from Canadian House of Commons, first session, 27th Parliament, Special Committee on Drug Costs and Prices, Feb. 14, 196,7, pp. 2482-2484] (By Henry B. Steele, Associate Professor of Economics, University of Houston, Houston, Tex.) Almost any drug will sell, if promoted intensely enough, at least for a while. Drug firms complain of the high rate of obsolescence of drugs, and argue that such risks justify high profit rates. The argument is not irrelevant under present circumstances, but the risks of obsolescence are not inherent but result from the way in which drugs are developed and promoted. High risks do not justify high profits in this instance because the risks and profits are both symptoms of the same disease: sales promotion rivalry substituting for price competition. The chief reason for the high turnover rate among drugs is, I suspect, to be ex- plained along these lines: advertising alone can sell physicians on a drug, if intensive enough, but any number can play at the advertising game, especially when brand names can be used to obscure the relationship between or even the identical nature of nominally unique substances. The greater the accumulated experience with a given drug, however, the more likely it is that its untoward actions will become known. However, `if the rate at which new products is intro- duced is as great as the rate at which publicity is given to the mischief caused by existing products, the sales of the new products W'jil increase as that of the old products declines, so `that the total cash flow need not `suffer.' On time other hand, as any businessman knows, advertising rivalry e~n sub- stitute-perhaps entirely-for genuine price competition. Price competition is a good servant to the consumer, but a harsh master to the producer. 1-lence sellers tend `to avoid it as much as possible under normal circumstances, and it generally prevails only where it is forced upon them by the structure of the market: numerous small sellers, none dominant; no collusion; no barriers to entry of `new firms or expansion of existing firms. Where sellers are fewer and larger; w-here barriers exist to entry by new firm; where legal devices exist to facilitate a community of interest in price and production policies-under these circum- stances, the forces which compel producers to undertake active price competition will be so weakened that rival firms will attempt to maintain or enlarge their share of various product markets by raising costs instead of lowering prices. Advertising is inherently less destabilizing an arrangement than price corn- petition. Some segments of the market may be loyal to a given brand even in 1 In the United States, brand-name sellers had to be compelled by law to give proper prominence to generic names in advertising. But brand-name sellers do have their uses for generic names. A firm, for example, may advertiseby brand name, but issue warnings under the 5enerie name only. Pfizer and Wyeth adopted this opaque tactic for a triacetyloleando- mycin warning. PAGENO="0080" 1630 COMPETITIVE PROBLEMS IN THE DRUG INDUSTRY the absence of advertising; other segments can be induced to prescribe only by increasingly provocative sales appeals. In general, sales can be increased by increasing advertising coverage, attracting new buyers while retaining the old, and perhaps even reinforcing their allegiance to the product. In the case of price competition, however, even though there may be a substantial segment of the market which is not highly price-sensitive and would buy the product at relatively high prices, in order to attract additional and price-sensitive custom- ers, the prices which all customers pay must be reduced. Under such conditions, the controlling considerations relate to the price sensitivity, or price elasticity, of the total market demand for the product, and the expected price elasticity of the net demand schedule which the prospective price-cutter estimates that lie will face after all his rivals have consummated their reactions to his price reduction. Only if demand promises to be quite sensitive, or relatively elastic, in response to price reductions, will a particular rival feel justified in gambling on a price cut. Even so, one or two moderate price reductions will ordinarily be sufficient to traverse the region of sufficiently elastic demand and hence to exhaust the possibility of further consumer-benefitting price reductions by the rivals. The major difference between the two strategies is therefore that price com- petition benefits consumers through lower prices and higher output, while it reduces the profit levels of producers to competitive rates-an outcome consistent with maximum efficiency of resource allocation in an economy. But rivalry in extravagant marketing campaigns raises costs and prices, benefits of advertising media at the expense of consumers, and possibly also at the expense of company profits, and keeps the total consumption of the products of the industry at rela- tively low levels. In fact, the effects on profits of the two strategies may be the same in the long run: initially high rates of return on investment serve as a stimulus to efforts to increase output and market share; price reduc- tions will directly reduce profits to equilibrium competitive levels; increased advertising budgets, which are mutually offsetting in the same way as com- petitive price reductions, except that they do not reduce costs to consumers and increase quantities produced and consumed, may eventually reduce profits to no more than competitive levels. Hence, monopoly prices may not necessarily mean monopoly profits, but simply excessive sales promotion budgets. A monopo- list does not always make monopoly profits-he does so only to the extent that he is efficient, and one of the great attractions of monopoly is that it reduces or largely eliminates the penalties which a competitive market imposes on in- efficiency. It should be noted in passing that while price competition benefits consumers and while advertising rivalry may benefit no one except to the extent that it attracts more resources into the advertising industry, it is not suggested that monopolistic rivals are motivated by the desire of private gain at the public ex- pense, while competitive producers are motivated solely to serve society. The producers' motivations do not differ. Competition is always a competition in the hope of establishing a monopoly, but where the structure of the industry rules out the possibility of monopoly, the ambitions of competitors must fail of fulfillment. The task of public policy is to adapt market structures in such a way as to pre- serve the vigor of competition while securing the public against the dangers of monopoly power on the part of the too-successful competitor. Mr. GORDON. Let me just read a few sentences from Dr. Steele's statement. He says: Drug firms complain of the high rate of obsolescence of drugs and argue such risks justify high profit rates. The argument is not relevant under present cir- cumstances because the risks of obsolescence are not inherent, but result from the way in which drugs are developed and promoted. High risks do not justify high profits in this instance, because the risks and profits are both symptoms of the same disease, sales promotion rivalry, substituting for price competition. Do you want to comment on that? Dr. COOTNER. Well, first of all, if I can call your attention back again to the sentence which we are discussing, my statement referred spe- cifically to a risky competitive industry. PAGENO="0081" COMPETITIVE PROBLEMS IN THE DRIJG INDUSTRY 1631 Now, the statement that you are making makes all kinds of empiri- cal judgments which I am not qualified to evaluate. As I say, if you can show that you will reduce risk at the same time that you reduce the rate of return that is defensible, but that is an empirical question. Mr. GORDON. That is right. I made an~ empirical statement. You are making an assumption which apparently is not necessarily grounded in reality; isn't that correct? Dr. COOTNER. Well, it is possible that either my statement is not realistic or your statement is not. That is, I can't- Mr. CUTLER. Professor Markham would be glad to take you on right now. Dr. MARKHAM. I am not hankering to take you on, sir. I think I would quibble with you that the statement you have made is an em- pirical one. It is the statement of an opinion. What we are talking about in this presentation largely are matters of fact. I did not see the statement you read, any empirical evidence that says that risks and rate of returns empirically are derived from the same symptom or same disease, whatever the statement was. Mr. GORDON. I didn't say it was an empirical statement. Dr. MARKHAM. He did. Dr. COOTNER. I made the statement that it was a hypothesis, if the hypothesis held then we ought to find evidence from the data and if we can't then it is a different proposition that I am stating. I think the conditions you referred to are clearly met in my statement which says if the industry is risky and competitive then you cannot lower the profits by administrative action without at the same time reducing the risks, if you can propose such a technique for reducing risks and profits at the same time that is a matter that can be discussed empiri- cally and tested and established in connection with evidence. Mr. GolmoN. Let's go a little further. You say here that the high risk-high profit relationship is an aver- age or long-run phenomenon. Do you mean by this that short-run losses or great gains may well occur, but over the long-run the ex- traordinary successes and failures will balance out? Dr. COOTNER. That is correct. Mr. GoiwoN. All right. Now, isn't it true that this balancing out seems to be absent in the high-profit drug industry? Dr. COOTNER. Well, that is an empirical question. Mr. GORDON. Can you tell us, say, for the past 10 years about 1 year in which the average profits of the drug industry fell below the aver- age profits of industry as a whole? Mr. CUTLER. Mr. Gordon, Professor Cootner is trying to testify as to the theory of the relationship between risk and return as an eco- nomic principle. We are then going to present to you, as you have seen in the papers we have here, through the immediate succeeding witnes- ses, the measurements of risk we have tried to make in the drug indus- try, that is empirical data which you may disagree with. As to the conclusions which you would draw or we would draw from that data on risk in the drug industry Professor Markham is prepared to testify about that. But I think we would have a more logical, orderly discussion if we could first establish the theory and then go on to the 81-280-68-pt. 5-6 PAGENO="0082" 1632 COMPETITIVE PROBLEMS IN THE DRUG INDUSTRY empirical demonstration and then the conclusions that each of us may draw as to the reasons and the absence or presence of risk in this industry. Mr. Goirnox. All right. Dr. Cootner, let's go back for a moment to the second paragraph on page 2 of yourprepared statement: 111 our relatively unregulated capital markets, investors are free to back the enterprises that they judge will be most successful. In doing so, they must inevit- ably make judgments about the management of the enterprise, the likelihood of success, and the prospective rate of return. They not only evaluate the particu- lar. enterprise but they must compare it with all other relevant alternatives. Now, these relevant alternatives to the investors you mentioned cover all industries as well as firms within the drug industry, is that correct? Dr. COOTNER. That is correct. Mr. GORDON. In other words, the range of choice involves all firms and all industries. Comparisons as to investment opportunities are not confined within one industry. Am I correct in that? Dr. COOTNER. That is correct. Mr. GORDON. If I were to suggest, Dr. Cootner, that since investors invest in firms, their main concern is with the profitability of the firm as a unit, not with just one item in that firm's catalog. Would you accept that as a reasonably accurate statement? Dr. COOTXER. Sure. Mr. GORDON. If the firm is the relevant business entity, the high- risk firms are those that show more extensive fluctuations in profits than do the rest of the firms in the economy, sometimes doing better than average, sometimes doing worse, but over the long run doing perhaps even better tha.n average to compensate for uncertainties. Is that a reasonable interpretation of your view? Dr. COOTNER. That is a reasonable interpretation of my views if we agree on the yardstick. Let me make the following hypothetical proposition. Say that we set up a company to explore for oil, and on our first try we were to hit a major field. That would establish a very high rate of return on our rather small initial investment. Now, we might have large fluctuations in the return on subsequent investments without any necessity for the company's average return on all of its assets to fluctuate above and below the average rate of return earned by all industry. The additional investments continue to show the same high risks, but the average rate of return remains above the average. It is just as if you have a run of luck in a flip of a coin, you have had that run of luck and you are relatively well off, it is behind you. You will still face the same fluctuations henceforward and yet you may maintain a. higher rate of return. With that proviso, I agree per- fectly with your statement. Mr. GORDON. Now, in the study that was made by Mr. Conrad and. Mr. Plotkin on page 11, the top of the page, you emphasize- Mr. CUTLER. Mr. Gordon, if we can go two more pages Mr. Plotkin is going to testify on that. Wrouldn't it be more logical to bring the question up then? Mr. GORDON. Fine. There are some questions that have been brought up by Dr. Cootner here which have relevance to the drug industry, but. we will bring them up then. PAGENO="0083" COMPETITIVE PROBLEMS IN THE DRUG INDUSTRY 1633 Mr. CUTLER. He is not going to go away. He will be here ready to answer any question, but at that point you will have had this state- ment, this study, put into evidence, and Mr. Plotkin will have had a chance to explain the theory of that study and answer any factual (1uestions you may have about it. Mr. GORDON. In other words, Dr. Cootner is not going to speak about the drug industry at all; is that correct? Mr. CUTLER. If you ask him questions about this study we are about to present on the drug industry as to its theory and methodology he can certainly speak to that as an economist. But as to his knowledge of the drug industry from other economic studies, he doesn't have that. Professor Markham does have that to a considerable degree. Mr. GORDON. All right. Senator NELSON. Your presentation and the division of the labor is so sophisticated that the line between what each of the witnesses is prepared to answer isn't perfectly clear to the committee. Mr. CUTLER. We understand that very well, sir, and this is why we suggested that we have a chance to put in the four papers and then all four can answer questions, whichever is best suited according to his specialty. Senator NELSON. I wouldn't object to the witnesses answering any question addressed to any other witnessif it would help. Mr. CUTLER. Fine. Senator NELSON. Go ahead. Dr. COOTNER. I am on page 9 and we just discussed the effect of administrative action on the profits of an industry without reducing its risks at the same time. Such an impact could evolve in two different, but interrelated direc- tions. If a company had no control over the risks it faced, it would simply find investment less attractive and reduce its level of capacity. If the reduction in return were great enough, investment might virtual- ly cease, except for the outlay of funds necessary to protect existing investment, and the industry would gradually decay. To find the extent, however, that the firm can control the riskiness of its investments, it has another alternative. It will try to reduce its ex- posure to risk. A natural gas exploration company, for example, might shift from looking for new fields to the more intensive development of existing fields. A research-oriented áompany would try to reduce its investment in more risky basic research in favor of the development of variants of existing, well-researched products, or toward improved or increased marketing of existing products. A manufacturing concern might find itself less willing to risk building capacity in anticipation of demand, or to be less quick to introduce new products. If none of these risk-reducing alternatives were available, the lower profits would result in a drying up of capital sources and a gradual reduction of the industry's activities. I would be foolish to claim that profit restrictions would lead to the demise of industry, but there is no question but that such limitation- again without a concomitant reduction in risk-would lead to a dif- ferent kind of economy than Americans have been used to, one that was less aggressive in anticipating demand and less prompt in accepting innovation and change. PAGENO="0084" 1634 COMPETITIVE PROBLEMS IN THE DRTJG INDUSTRY In stating these implications of profit restriction, I wish to empha- size the point with which I introduced my testimony: That one cannot arbitrarily alter the prices of an industry without altering its profit. If excess profits exist, price reduction may do no harm, but the ques- tion of prices cannot be discussed in a vacuum. If present pricing produces profits no larger than those required by the risks of the drug industry price reductions will lead to reduced investment. It is for the witnesses who follow me to discuss that point. Mr. GROSSMAN. May I ask a question? Dr. Cootner, I might ask you in general terms, if risks are not borne out by some losses after a certain period of time, can't we change our estimate of the industry somehow? For example, one of your, one of the gentlemen who will testify today, Dr. Whitney, has said "20 years are too few in any case to deduce absence of risk from continuing prof- its. Look at what happened to the railroads or for that matter meat- packing, sugar, and textiles." Would you go along with that, we have to wait 20 or 40 or 60 years, year after year of high profits, and we're still to believe that this is a high-risk industry? Dr. Cooa~ER. Well- Mr. GROssMAN. I am not asking you in terms of the drug industry, but in general economic terms. Dr. COOTNER. I understand. I would certainly say I would take all new information as it came and use it to adjust my a priori expecta- tions. If you were a classical statistician you always want to wait until all the information is in. I don't think that you have to do that when you are making decisions as time goes on. On the other hand, I do agree with Professor Whitney's implicit statement which is that there are a lot of risks which are virtual- which may not show up for a long time and yet which influence pros- pective action. Let me give you an example which I do know something about. If you look at the rate of return, for example, on IBM, on its assets, it is quite high, and the rate of return on its stock has been also quite high. If an investor at any one time anticipated that the rate of growth of profits of IBM was going to continue at the same rate in the future that it had in the past, he would find that the stock is actually worth a lot more than it is presently selling for. One of the reasons this extraordinary rate of profit continues is because there is always the expectation that its predominant share of the market may not continue. Now that may be an incorrect expectation and if it is anybody within the sound of my voice can profit from that. Mr. GROSSMAN. is there any period of years, for example, he says 20 years are too few, I mean specifically, can you say there is any term of years that we should wait to watch and decide? Dr. COOTNER. No, I don't think there is any particular term. I think-f or example, I used in my study, about 17 years, and Mr. Plot- kin and Mr. Conrad used 16. I do not take the point of view that you cannot ever or you must wait 50 years in order to make a decision that the risks are too high. I think if you could really state that in 20 years there was no evidence of risk then I would be willing to say that that was sufficient to decide that there was no risk. PAGENO="0085" COMPETITIVE PROBLEMS IN ¶]HE DRUG INDUSTRY 1635 Mr. GROSSMAN. Do most industries like, for example, the insurance industry, build up reserves to meet these risks? Don't most industries build these up? Dr. COOTNER. Well, some risks are more insurable, that is predict- able than others, and-it is easier for a company like an insurance company to foresee the potential magnitude of risks and plan for them. As a matter of fact, it is much more difficult for a manufactur- ing company which cannot retain earnings in excess of current needs without running into some difficulty with the Internal Revenue Serv- ice to establish such reserves. It is abstractly possible, but I think in practice that isn't the way it is done, except perhaps in banking and finance, where the risks are a little more predictable. Mr. GROSSMAN. I think you mentiOned once the words excessive profits. Could you explain basically how you interpret excessive profits? If you could apply it to the drug industry it would be helpful. I don't know if you could. Dr. COOTNER. Well, excessive profits in the abstract sense in which I was discussing it would be profits in excess of that required to attract capital, given the existing risks in the industry. That is if you could show, for example, that a company earned very high profits and did not in fact have more risk than any other company then those extra profits would be excessive. Mr. GROSSMAN. If a company was showing a profit over 15 or 20 straight years would you have to look in terms of exceessive profits even if the profits were to, say, one Of the leaders of all industries, if they were the highest? Dr. COOTNER. There is where we have a problem and it is the same problem that I raised in the discussion with Mr. Gordon. For example, suppose that you start with a very high initial rate of return, and subsequently you face risks on new investments. The average rate of return on all your assets is going to remain high. It will not necessarily fall below zero or even below the industry aver- age and yet it will show a continuing diminution or fluctuation because of the varying rates of return on new investment. Now, again if you allow for that phenomenon, then I agree with your statement that you .make, that is if a company shows that it never has any substantial risk on additional investment, that addi- tional investment takes place at a high average rate of return that *is very stable then I would agree that that is not a risky company. Mr. GROSSMAN. One final question: Could you tell me as an econo- mist, do you think it part of your job to assess an industry in terms of its social responsibility at all. Dr. COOTNER. Well, it is a very difficult question. Certainly an industry should be aware of its social responsibilities. The problem is that in the last analysis the industry has to attract capital from investors who are relatively unconcerned and who in fact get no return from accepting the social responsibility. A stockholder cannot bask in the knowledge that he is performing a social service. Now, within the limits imposed by the requirements for raising funds, I think any industry should pay attention to its social responsibilities. Mr. GROSSMAN. You would say though it is a factor? PAGENO="0086" 1636 COMPETITIVE PROBLEMS IN THE DRUG INDUSTRY Dr. COOTNER. Yes, sir, I would agree. Mr. GROSSMAN. Thank you. Mr. GORDoN. Dr. Cootner, I just want to ask you one thing: You talk about prices being lowered as a result of Government flat, Gov- ernment regulations. But you also assume that the costs are not regu- lated or controlled. Isn't it more reasonable to assume that if prices are controlled, costs also would be controlled? Dr. COOTNER. Well, I don't know that that is the practice that exists. Mr. GORDON. Let's take the American Telephone and Telegraph Co. as an example. Dr. COOTNER. Yes. Mr. GORDON. Only certain items approved by the Federal Com- munications Commission can be worked into the rate base. Dr. COOTNER. Oh, yes. Mr. GORDON. Not everything that they spend can be iut into the rate base. Dr. COOTNER. That is true. It is certainly true, for example, that they are not allowed to-certain kinds of advertising are not allowed in the rate base. But whether that is an advantage to the firm or not depends upon what the efficacy of the advertising would have been, and it may be in fact that you actually reduce the profitability of an industry by preventing it from incurring certain costs, for example, during the war, as part of a clear view of the Government about the social priorities we prevented automobile manufacturers from expand- ing capacity to produce private civilian automobiles. Now that didn't help their profits although it certainly reduced the costs they incurred. Mr. GORDON. But isn't it more reasonable to assume, if you are going to make any assumption at all, that if prices are going to be regulated by the Government, costs also would be? Dr. COOTNER. Well, you know, wages are not. I am not saying that you can't regulate some costs, but it would only affect my argument if the regulations of costs in fact resulted in a rise in profits. If you lower the profits by reducing prices and then increase profits by reducing costs then you haven't lowered profits in which case there is no problem with anything I say. If you can reduce prices and costs simultaneously for an industry so its profits are not changed you obviously have influenced the investors not to view the company with any more disfa.vor than you did before. Senator NELSON. Under your definition of a high risk company, that is the part of the definition involving stability on the profit side for 17 or 20 years, a very high percentage of all American industry is high risk industry then, is it not? Dr. COOTNER. Well, all American industry has risk of some degree. Obviously, only roughly half of the industries have above average risk, I mean that is a property of the median. There a.re a lot of risky Amer- ican industries. I am not trying to be fussy, it is just not clear what you mean by high risk. Certainly risk is compa.rative and certainly many industries are more risky than many others. Risk is not a property solely of the drug industry or of any other single industry. Senator NELSON. The reason the question comes before us is that the industry asserts continuously that this is a special situation, a special industry, that is very high risk and, therefore, very high profits are PAGENO="0087" COMPETITIVE PROBLEMS IN THE DRUG INDUSTRY 1637 justifiable. Since the assertion is continuously made, it would be help- ful to the committee if the economists could place before us a genuinely convincing case to prove the claim. Mr. CUTLER. That is just what we hope to try to do, Mr. Chairman. Senator NELSON. Well, I don't know whether you are the one to respond to this. If not I would want to ask one of the later economists. Last week we had as a witness Mr. George Squibb, former vice presi- dent of the E. R-. Squibb Co., now a consultant to the Squibb Co., with 30 years of experience in the company. On this question he states: The industry cannot cite its risk from the uncertainty of research results or from product obsolescence as the reason for the price differential-for some of it no doubt, but not much because year after year the profits of the industry are far above the average for other major industries, and currently appear to be improving rather than worsening. The risks of the pharmaceutical business seems to those outside it to be pretty ephemeral in view of the impressive profit per- formances of the last two decades. Indeed, it is the concensus of the industry's critics that more risk is needed to make it more sensitive to the normal influences of competition. Then I asked a question of Mr. Squibb,: My I interrupt at this moment? Some of the testimony we have had from the industry, statements from the industry and the medical profession, as a matter of fact, have asserted that in order to attract capital for the purposes of expan- sion and for research these high profit rates are necessary because of the great risk and so forth. Do you have any comment about that? Mr. SQUIBB. In terms of risk, I commented on risk earlier in my earlier paper. I do not feel that the pharmaceutical industry itself under the definition of the word "risk" as I understand it-is a risky business. It is a profitable business, one in which companies have long engaged, and most of the companies have many years of profit history behind them. And while it is true a given product in their structure may come and go, or be relatively short-lived, however the risk, the financial risk, it seems to me in my experience, has been very little in the pharma- ceutical business. A well-managed pharmaceutical company has as good an investment, is as good a place to put your hard-earned dollars, as safe a spot as any I know. I do not think there is a risk involved here. In terms of what is com- monly thought of as a risky speculative sort of business, the drug business is not. Here is a rather clear, sharp statement about a di~tinguished mein- ber of the pharmaceutical manufacturing industry. - Dr. COOTNER. That is certainly an empirical proposition and I am sure it is really going to be a matter of whether Mr. Squibb or the witnesses who are going to discuss this question here are more con- vincing to you. - I don't necessarily think just because somebody has been in an industry he necessarily is more convincing than an economist, but that is perhaps because I am an economist and not a member of industry. Senator NELSON. I will buy that because a lot of people in the in- dustry don't think anybody outside of it is qualified to comment on it-. If we will both agree that those outside are qualified to comment as well as those inside I will accept that. )ou do not wish to respond to that? - Dr. COOTNER. I think, I am in no-~ - Senator NELSON. Fine. I didn't know whether somebody else would. Just out of curiosity, as a former member of the Henry George Society, what is the Econometric Society? Is that a group of econo- mists, professional economists of a special kind? PAGENO="0088" 1638 COMPETITIVE PROBLEMS IN THE DRUG INDUSTRY Dr. COOTNER. It is an association of economists whose particular specialty is in mathematical and statistical approaches to economics. Senator NELSON. That is the new school as contrasted with the classical economists? Dr. COOTNER. That is one way to define it. Senator NELSON. Who is the next witness? Mr. CiTTLER. The next witness, Mr. Chairman, would be Mr. Plotkin. Mr. Plotkin and Mr. Conrad of Arthur D. Little have tried to make some empirical, if I can use an economist's word, measurements of risk in the drug industry, and have tried to correlate the relationship be- tween their measurement of risk and return in that industry with an array of other industries across the entire spectrum. We would like to have Mr. Plotkin present that to you right now. Senator NELSON. Both of these? Mr. CirTr~Eu. That is the booklet entitled "Risk and Return in Amer- ican Industry-An Econometric Analysis." Senator NELSON. That is the case in which charts are used by both witnesse~s? Mr. CUTLER. That is correct. Senator NELSON. All right. We are glad to have you here, Mr. Plot- kin. Your biographical sketch will be printed at this point. You may proceed to give your statement. (The biographical sketch of Mr. Plotkin follows:) BIOGRAPHICAL SKETCH OF IRVING H. PLOTKIN I am an economist specializing in econometrics and operations research. My academic training includes work at the Wharton School of Finance and Com- merce (University of Pennsylvania) and Massachusetts Institute of Technology Sloan School of Management, Department of Economics (for which I am present- ly com~i1eting my doctoral dissertation). I have served as a consultant in prep- aration of econometric Studies for several government agencies (Department of Justice, AEC, FDIO) as well as some of the nation's largest corporations. My work has been primarily concerned with decision making under uncertainty and the measurethent of risk. STATEMENT OF IRVING H. PLOTXIN, CONSULTANT IN ECONO- METRICS AND OPERATIONS RESEARCH, ARTHUR D. LITTLE, INC., CAMBRIDGE, MASS. Mr. PLOTXIN. Thank you, Mr. Chairman and members of the com- mittee. I am Irving H. Plotkin, an MIT economist and a consultant in econometrics and operations research. As a part of their research for the Pharmaceutical Manufacturers Association, Arthur D. Little, Inc., has retained me to perform econometric studies of the risk/return pattern of the American economy. Our full statement of the findings of this research and the rele- vance to the pharmaceutical industry is contained in the report "Risk and Return in American Industry-An Econometric Analysis",1 just referred to by Mr. Cutler. I request that it be printed in the record as part of my testimony. 1 The report "Risk and Return in American Industry-An Econometric Analysis," begins at p. 1746, infra. PAGENO="0089" COMPETITIVE PROBLEMS IN THE DRUG INDUSTRY 1639 I propose, with the aid of a few charts, to summarize here, by the use of simplifying charts I have brought along, the key methodological and quantitative results of this study. However, we must first be con- cerned with the question of relevance. Professor Cootner has shown that one cannot meaningfully discuss prices without considering profitability. Likewise, one must consider the risk/return pattern formed by the economy as a whole before any particular industry's profitability can properly be evaluated. There- fore, the body of research I am here to discuss today is a general in- vestigation of risks and returns in all American industry. I would like to underscore this point, that the first part of our pre- sentation is addressed to the overall pattern of risk and return in the economy as a whole. It defines the measures that can be applied to vari- ous industries. The results and implications of this analysis have been accepted for publication in the Harvard Business Review. I, too, as did Professor Cootner, do not appear before you as an expert on the drug industry. My expertise concerns the measurement of relations between risk and rate of return. Other witnesses will dis- cuss the specific nature of the risks faced by the drug industry. I can, however, say that our research has established a valid measure of the industrial risk/return pattern for the :American industry. Further, based on our investigation, there is no reason to conclude that the drug industry's performance in the statistical sense is out of line with the statistical relationships we have measured. Attached to my statement are three charts which illustrate the phi- losophy and methodology we have followed in measuring the risk/ return relationship. They show how our risk measure concentrates on the uncertainty or lack of predictability of marginal profits and, fur- ther, how the economy's actual risk/return pattern can be econometri- cally quantified and the position of individual industries isolated and evaluated. I, therefore, with the aid of these charts, propose now to summarize the findings which are spelled out in detail in the written document. As Professor Cootner indicated, it is my task to show that one can meaningfully use statistical theory and econometric analysis to quantify the risk/return pattern; that is, to raise the argument from one of mere beliefs and contentions to quantified results and compet- ing hypotheses. But before comparisons between competing industries can be per- formed, it is necessary to specify definitions of the terms upon which we want to compare industries. Therefore, we have to define for this study two important concepts. The first is a definition of rate of return, which can be applied across many diverse industries: industrial as well as financial; those engaged in manufacturing basic necessities as well as those engaged in manu- facturing of luxuries. More difficult, however, is the construction of some measure of the risk environment which can measure the impact of, without isolating the individual courses of uncertainty arid lack of predictability within the wide sprectrum of American industries. May I emphasize that point, Senator. We do not at this stage in the study try to say why each industry, one by one, for the 59 industries we have studied, is risky. We construct a statistical surrogate, a measure which corresponds through the force of our theory and measurements PAGENO="0090" 16 ~O COMPETITIVE PROBLEMS IN THE DRUG INDTJSTRY we have taken to the lack of predictability or uncertainty that faces persons making investments in the various industries. We then see what type of pattern is established, and whether the drug industry shows evidence of being out of line with the risk return pattern. Senator NELSON. May I interrupt a. moment? Is that the chart which you show with the clots on it, and a. circle around the drug industry? Mr. PLOTKIN. They finally end up on the chart with the clots. But my first charts go to show the definitions behind each clot, the definition of risk and the definition of the return. It is a. very straightforward demonstration without mathematics. Senator ~\ELSOX. And the graph line on that chart. represents the average for all industries? Mr. PLOTKIN. No, sir. That represents a statistical relationship be- tween risk and return which, I think, can be most easily explained after the other definitions are put forward. The basic unit of concern in this analysis, as any risk analysis must be, must be the individual company. Mr. Gordon has properly pointed out that that is the unit of decisionmaking. It is within the individual company that the balance between ex- pected returns and expected risks is struck. However, this does not imply that the. only or best source of information about risk ex~ecta- tions is historical company information. Let us pause here, it is the expected risk on a venture that we ary trying to measure, not some- thing a.bout the past, but from the past gain information about in- vestors, opinions about how risky are particular ventures. Senat.or NELSON. In context, what do you mean by the word "ven- ture"? Mr. PLOTKIN. A venture, I am using it in it.s general sense, can be viewed by an individual as a. gambling proposition, the rolling of dice in the game of "craps" or the drawing of cards in the game of black- jack; it ca.n be viewed in the industrial undertaking as being a new biological active agent in a research and development program, or the design of such as a new car, which could be either a Mustang or an Edsel. Senator NELSON. You are talking about a risk venture, a high-risk venture? Mr. PLOTKIN. I am talking about any venture, and I want now~- Senator NELSON. Any investment venture by anybody? Mr. PLOTKIN. Yes, sir. I want to see whether I can find a pa.ttern among industries that show some of the ventures undertaken by certain industries seem to be less risky than the ventures undertaken by other i ndlustries. I believe this goes to the heart of the question you are asking before, to show the results that have been claimed by various witnesses. Management, it is our contention, forms its risk expectations not solely in the past profit history, of its own company, but also on the diversity of histories that other companies undertaking similar "ven- tures," to use the term again, other compames within its industry undertake. Likewise, entrepreneurs considering entry into an industry, will assess the general riskiness of the industry by examining the range of corporations active, and most importantly or previously active, in that PAGENO="0091" COMPETITIVE PROBLEMS IN THE DRUG INDUSTRY 1641 industry. It is, therefore, not unreasonable to ask for a measure of expected risks based on historical industrial performance. I would now call your attention to the fact that the noted economists Professor Cootner has referred to have offered very numerous defini- tions of uncertainty or risk. The common element in all these defini- tions, because they were all looking to the future as it unfolds before us, was the uncertainty or lack of preclictabihty of the outcome of the venture. The uncertainty involved in predicting the outcome of a particular investment or a particular investor's venture is, in general, what makes the process risky. The greater the uncertainty about final outcome of the event, the greater the risk for those investing in that venture, that future event. With that rather simple proposition in mind, consider the two alter- native propositions as though you are being offered two gambling propositions, one, the game of blackjack, the other a game involving dice; alternatively the choice may be between an offer to invest in the shares of companies making computers and the shares of an electric utility company. May I call your attention to the chart, sir? Senator NELSON. You are not talking just about high-risk ventures now? Mr. PLOTKIN. No, sir. We considered the spectrum of American industry. I might at this point make a statement about the data source. I think it is appropriate. We did a pilot study where the data were very, very carefully looked at to make sure that industries were closely defined. There were only a small number of industries and a small number of companies, about 13 industries and 70 companies. The pilot study was just to test the hypothesis where we understood something about the risks of these industries to see whether we would get the line you referred to before, the upward sloping relationship between risk and return. However, for a public and general presentation, the presentation to the scholarly and academic community to whom this study is also addressed and, therefore, being published in scholarly journals, we have taken the Standard & Poor's data~ source, something available to everyone, so everybody can repeat it, the data source, and Professor Cootner referred to this in his testimOny, is the Compustat Annual Industrial Tape prepared by Standard Statistics. We have not changed one iota of their data, including their industry classifications as to which companies belong to which industries. In one case we have some reasons to believe that the data might be less than fully accurate, although not having a great effect on the outcome of the study, I would just note, because we are particularizing on the drug industry, S. & P. have included in that industry the Gil- lotte Corp. When I called them they admitted there was an error on their part, but that was in the tape, alid it only went to, in fact, it greatly went to, increase the average rate of return in the drug industry as so reported, I have left that point in there, and we have done so in our analysis so as in no way to bias the standard authoritative source, the Standard & Poor organization. PAGENO="0092" 1642 COMPETITIVE PROBLEMS IN THE DRUG INDUSTRY Well, let us stay with the chart. Let us consider the distribution of possible outcomes of two ventures, venture A and venture B. (The chart referred to follows:) DISTRIBUTION OF POSSIBLE OUTCOMES OF TWO VENTURES Mr. PL0TxIN. Here-the "a," axis-is the value of the outcome of either venture, and here-the "y" axis-is what is known as the proba- bility of occurrence of that value. So as you see, the higher you are here, that means this is the most likely event to occur, having most of the probability. Any other value has somewhat lower probability. Now, my claim is quite a simple one. It is not highly sophisticated in its mathematics. It merely says that faced with venture A and venture B, one is more certain of how he will wind up if he says, "I will ride curve A," than if he says "I will ride curve B", because curve A is the least dispersed, less diverse in its nature. Curve B is more spread out. This characterization is a very good characterization of various games of chance, some of which, like roulette, have very widely dif- fering outcomes on very small probabilities, and very high gains. Others of them-I am not a good enough gambler, so I am not sure- are more concentrated as to what the value of the outcome will be. I would, however, make one more point. The measure of risk does not depend merely-the measure of uncertainty is not and I use these words interchangeably, because that is the common thread through all definitions-on the extremes of these distributions because theoreti- Value of Outcome PAGENO="0093" COMPETITIVE PROBLEMS IN THE DRUG INDUSTRY 1643 cally, and practice has shown this-the~r run out ~tll the way to the end of the positive line and to the end of the negative line with very, very small probabilities. It is a measure of the weighted distribution, so that you get a picture of the center of gravity, where most of the events occur; not just what is the highest and the lowest. The initial scholarly investigation which tried to rank industries by the highs and lows got those atypical com- panies which would make the industry look very high or very low, and did not find any answers that correspond to commonsense. Here we are talking about gambling ventures or stock market ac- tivity as viewed by an investor. How does this help us qualify indus- tries, individual industries? (The chart follows:) DISTRIBUTION OF RATES OF RETURN EARNED BY FIRMS IN INDUSTRiES A AND B IN A PARfiOULAR YEAR Rates of Return Mr. PLOTKIN. Consider now, and I have not done any fancy foot- work, these are precisely the same curves, I just relabeled the axis, because I wanted to use the analogies, these are the distribution of rates of return in two indusi~rie,s. Here we standardize for the different size that companies might retain by relating the return, the amount of money you get back, to the amount invested. The chart shows rates of amount of returns in hypothetical industries A and B in a particular year. Here we have A .~ B PAGENO="0094" 1644 COMPETITIVE PROBLEMS IN THE DRUG INDUSTRY the percent of the firms which earned a given rate of return, and here we have the rate of return which they earned. Distribution A might be considered typical of the utilities industry where, because of regulation the final rate of return earned is very similar from firm to firm. It is narrowly distributed, centered around 6 percent, and if you earned 8 percent one year, the next year at a rate hearing, you are pulled back down. In any one year, given that you are making a physical or stock market investment in utilities, you can be fairly certain of what your rate of return will be. In another industry, like the computer industry, the rates of return are indeed very dispersed. They have a high average, but there are some earning a good deal of return, and there are some earning less return. In the same sense that venture A was less risky than venture B9 I would like for you to think of these two industries here char- acterizeci as industry A being less risky than industry B. Industry B, if you are selecting a company at random, is more difficult to predict than is industry A under these terms. \Ve have now defined the basis of our measure of risk; that is, the dispersion of the results of individual companies based on some measure of return about the industry average. I would like, returning to the desk there, to concentrate for a moment on the problem you asked us to keep in mind, the definition of the rate of return for that, too, must be specified before we can measure the two quantities, the relationship between risk and return. However, I do not wish to underestimate the difficulties with this approach to risk measurement. For example, the indicated amount of risk in the particular industry will depend on how carefully the industry is defined or for some paiticular industry,investorsmay not have foreseen the actual degree of risk that really did occur. But if we define our variables carefully and utilize a large number of industrial classifications, and they standardized outside source, we can find a stable relationship with a high degree of statistical confidence. Furthermore, a number of different studies were undertaken by us at the Arthur D. Little organization, as well as other scholars, to compare the effect of using a number of different periods and different industrial classifications. It is encouraging, as Professor Cootner said before, and I reiterate here, that the results were not found to be highly sensitive to the precise definition or the precise number of years or indlustriail classification. Therefore, as a scholar and technician, we can be sure what we are seeing unfolding here is not a fluke due to chance and a particular set of definitions, but is objective and verifiable and can be constructed by any interested scholar. As with the measure of risk, we have followed the same approach in the measure of rate of return. The main variable used, and the one we believe to have the most significance for this purpose, is simply the rate of return earned by companies on their long-term investment at book value, since it is this rate of return which is the most commonly usedi and is the one considered by most investors and economists when they are at all concerned with the overall profitability of the industry. When the economists is charged with answering the question: "Is a sufficient amount of capital flowing into that mdlustry to providle the PAGENO="0095" COMPETITIVE PROBLEMS IN THE DRTJG INDUSTRY 1645 public with the services it needs?" The answer turns on the total capital and the total earnings of that capital. But, however, to placate common practice, we have used a number of other measures in order to assure that our results are not sensitive to the particular definition, all these measures are reported in detni 1 in the study which has been presented to you. Having defined that the rate of return will be the rate of return on the total asets invested, we must now state in precise terms the statistical hypothesis we seek to test. Formally, we can say we are constructing an equation. that was the way of describing the line that you have referred to that says, "indus- try return," the rate of return earned by an industry, is some linear function-as a first approximation-of "a" plus "b" times industry average risk: Industry return a + b (industry risk). We expect the coefficient "b" to be positive such that higher returns are associated with higher risks. One should be cautious here that we are not just wrapping around and meeting our own tail in the end. We are not saying that because a firm has higher returns it is, therefore, risky. Now, using two separate measures, one on the riskiness of the industry and another on its return, we test to see whether there is any such relationship. Once again the chart can help us. (The chart referred to follows:) RISK/RETURN PATTERN ~ MB VA Industry Risk (Variance) VB PAGENO="0096" 1646 COMPETITIVE PROBLEMS IN THE DRUG INDUSTRY Mr. PLOTKIN. In this particular hypothesized two industries here, it happens that the average rate of return in industry B, is higher than the rate of return for industry A. This is not ~ mathematical necessity because this curve could have been centered here, and curve A centered there. But it also turns out in this hypothesized case, not necessarily true in reality, the spread of risk of B is greater than A. Once again, this need not be true. We can summarize-and here, sir, is where I get into a definition of that line, I think, that will be most clearly understood-the fact, that the mean, the average, return of A is below the average return of B, and also the fact that the variance of A is below the variance of B on a two-dimensional graph where we plot the industry risk or variance here-x axis-and the industry average return here-y axis. Now, we see where the risk or variance of A and variance of B come, and we put dots at their intersection. WTe have traced out a straight line connecting these points. It is the hypothesized relationship, that an industry that has a higher return will also have a higher risk and vice versa. It may not be true in fact. Here is the way it happened in this hypothesized case. The statistical question, when we put a similar dot for each of our 59 industries, comprising 783 firms, will go through a scatter of dots, by which we put the line of best fit and, therefore, it will not hit every dot. It will show an upward sloping relationship, a downward sloping relationship or no relationship at all. Given we have established what the slope of the relationship is, we may then ask where does any one industry fall. Does it have evi- dence of being above or below the normal pattern? Now, in measuring the rate of return to protect ourselves from the sensitive of using any one year, which may be economically abnormal, we have considered the period 1950-65, a 16-year period, which cer- tainly allows for all post-World War II adjustment and conversion. In each year we constructed the industry average return and indus- try variance, so we had 16 yearly returns and 16 variances for each industry. PAGENO="0097" COMPETITIVE PROBLEMS IN THE DRUG INDUSTRY 1647 We then further averaged those numbers to take out any sensitivity to a particular year and, therefore, the "average variance" and the "average return" are numbers that contain 16 years' worth of information. From a statistical sense, and for Mr. Gordon's benefit, I say this is a highly rich statistical body, so that if we have a significant statistic, we can be quite confident in it. We have tested besides the book value rate of return another very important subsidiary rate of return, that earned by the stock market investors should they decide to invest in the industry through the mechanism of this stock and bond market, rather than investing hard, cold cash in machinery, or research and development. We know in the stock market there is certainly a continuum of returns that are r&ated to risk. Government bonds yield the lowest of most taxable bonds. Then come AAA corporations, Ak corporations, going up to mort- gages, and third mortgages, each with higher and higher yields, and more and more risk. The same with common stock. The blue chips have low yields as opposed to the speculative shares and computer shares which have on average, for an extended period of time as an industry group, high yields. Therefore, we felt it necessary to test the validity of our definition of variance, our definition of risk, in the stock market first. The chart presented in the study on page 20, I believe, is a test of the relationship measured at market value. The fact that it has a high statistical significance, the fit is very tight, and forms an upward sloping line, is to us a true validation of our measure of risk, that we have not quantified a stable measure of risk. We may then proceed to ask the second question which concerns the real economy, the second chart found on page 19; that is, what is the pattern really dependent on, do rate of returns and risks go hand in hand, and if they do, what can we say about individual corporations. I must correct myself, not individual corporations, but individual industries. (The chart follows:) 81-280-68-pt. 5-7 PAGENO="0098" Industry Return (Net Income + Fixed Char~es)/ToiaI Capitalization Industry Risk Average Inter-company Dispersion Cl 0 ENDISTRY RATE of 20 RETURN ~rj 0 02 8*' t:rj Li Li 2 ci 02 I I I I I _.__~ .-~ 10 20 30 so 50 60 70 80 90 100 110 120 130 ~ INDWTRy RISK MODEL I: ThIDUSTRY RISK/RETURN PATTERN MEASURED AT BOOK VALUE NOTE: (~)REPRESENTS POSITION OF PHARMACEUTICAL INDUSTRY PAGENO="0099" COMPETITIVE PROBLEMS IN THE DRUG ~DUSTRY 1649 Mr. PLOTKIN. This is a reproduction Of the chart which appears on page 19 of the report you have on hand. It has a striking similarity to the chart for market value. You have this upward sloping relationship. Higher rates of return are associated with higher risks in fact, not in hypothesis, and not because it is a necessary mathematical relation, but because it seems to be a validation of an economic truth. This is, in essence-you asked me before about the econometric~ society-this is in essence what an econórnetrician does, takes a hypo- thesis where he has an expectation of a certain relationship, goes to the real world, measures the data and tries to determine whether or not that is a valid hypothesis. This is on upward sloping relationship with a good fit, the dots are for this purpose tightly clustered around the line, tightly relative to normal statistical tests to which this rela- tionship has been exposed. Senator NELSON. This chart is through 1965? Mr. PLOTKIN. Through 1965; 1950 through 1965, each point repre- sents one industry, an average of over 15 companies per industry, but this number varied because some industries were smaller and some larger. So there are 783 companies, 59 industries, for 16 years. Senator NELSON. How many companies did you use in the red dot, in the drug industry? Mr. CONRAD. Twenty-nine. Mr. PLOTKIN. Twenty-nine, as classified by the Standard & Poor Corp. It included Gillette with its fantastic return on total assets, of about 30 to 40 percent. We included precisely all those coiñpanies included by Standard & Poor's in every industrial classification. Senator NELSON. Is my eye correct-is the red dot representing the industry with the third highest rate of return on that chart, or the second? Mr. PLOTKIN. It is the fourth highest in risk, and it is the third highest in rate of return. Senator NELSON. If it were brought up to date for this year then, would not the drug industry be first on rate of return? Mr. PLOTKIN. No, sir. I realize what the FTC report said, but that only puts in 1 more year, so it would then be a 17-year study. Senator NELSON. This is for a period? Mr. PLOTKIN. This is for the 16-year period. It is unfair to ask a statistical test to establish a relationship based on just 1 year's observa- tion. That is why before one can have confidence in an econometric analysis, it must be very rich in data drawn on many years' observa- tions, so as not to be unduly influenced- Senator NELSON. This past year's return would just raise it, but you do not know by how much? Mr. PLOTKIN. It would just raise it, but I can tell you that its effect, this return is about 3 percentage points greater than its 16-year aver- age, and would count one seventeenth of 3 percent-perhaps Pro- fessor C'ootner can help to figure out mental arithmetic here. Senator NELSON. I cannot tell what the difference, percentagewise, between the three which are above it is, so I do not know how much it would change it.. Would it change its position at all? Mr. PLOTKIN. It would change it, but not perceptibly. PAGENO="0100" 1650 COMPETITIVE PROBLEMS IN THE DRUG INDUSTRY Senator NELSON. It would change it to third or second? Mr. PLOTKIN. It could change it to be equal to third or to second, but it would not materially affect these results. Senator NELSON. What a.re the three industries which are higher? Mr. CUTLER. Mr. Chairman, you can see all those figures on page 78, industry by industry. Senator NELSON. Seventy? Mr. CUTLER. Seventy-eight. Senator NELSON. That risky gold-mining business is right up there in third place, I see. Mr. PLOTKIN. I am very glad you asked about it, and perhaps this is a subject of many senatorial and congressional investigations, these days. Gold mining was one of the industries which allowed us to have a good deal of fit in this relationship. We know gold mining does not operate under the normal economic pressures that confront companies competing for the consumer's dollar in the marketplace. The price of gold is artificially established by administration fiat, independent of the demand for gold. Some question may be raised whether that fiat will be maintained, despite the demand for gold. However, as an industry, it does not return a rate of return com- mensurate with the risks involved, which simply means that there is a. distortion of the amount of capital that has flown into that industry. This is something to be expected when the Government is primarily controlling the market. It is not functioning along, and competing along, with other industries. Senator NELSON. Did you operate on the assumption in your evalu- ation here that there are really no significant differences between the drug industry and the other ones that you have compared it with? Mr. PLOTKIN. No, sir. Senator NELSON. You say there is a difference in the gold-mining industry because its price is regulated. Did you take into consideration the fact that there is a high degree of control by the individual seller of the drug product in the market, and that he may very well be the highest priced with a product that is no better than one that sells for one-tenth of the price, and that there is no other industry that I know of in which that is true? Mr. PLOTKIN. Let me trace through my logic of the gold industry much more carefully. I, in no way, tried to determine the peculiarities of the individual industries ahead of time. I set up a standard of measurement, a defini- tion of return, and a definition of risk which was reasonable-reason- able to a large body of scholarly economists-through which definition I looked at all industries with the same set of glasses, and then I looked at where they fall on this pattern. After the analysis is done, I noticed that gold mining is seemingly out of the pattern, is one of the outstanding ones to be out of the pattern. I then see whether I know of any reason which could cause it to be out of the pattern. The statistical question, and this is why they call economics the dismal science, but perhaps statistics is equally dismal, and the statis- PAGENO="0101" COMPETITIVE PROBLEMS IN THE DRUG INDUSTRY 1631 tician always starts out with a null or negative statement, a null hypotheses. We are examining into the discussion-although it is not what the purpose of this entire investigation was-to see whether or not there is evidence, based on these definitions, alone, sir, that the drug industry is outstanding, is out of the pattern, that is, for the risks encompassed. So, measured by our definitions, it does not yield a nonnormal return to its investors. The answer comes from a statistical sense, not from a judgmental sense, by saying how far is that versus how far is something down there, but applying standard statistics, and the tests, there is nothing to suggest in fact, that leads us to reject he hypothesis that point 4 is outside of the distribution traced out by the rest of the points here. But this hypothesis can be applied to any of the points, and this test can be applied to any of the points. When applied to point 3, there is doubt that it is within the normal pattern traced out, this upward rise in pattern. That is the limit of our abilities as statatisticians and econometricians to make the statement. I would not want to seem that my testimony is making any other statement but that on a measure of risk, which has shown to have economic viability, making good commonsense and good practical sense when viewed against the stock market, and applicable across all industries, no matter what the cause or the source of risk, the drug industry does not seem to violate that pattern. This does not say that everything is rosy and perfect, and it does not even say such an analysis cannot be pushed to say how much should the rate of return be. When an economist catches a solid repression line he really means a gray area within which he cannot distinguish, and above or below which he says points fall outside of the pattern. There is no evidence that the drug industry is not in this hazy, gray area. This is not my assertion, but any econometrician or any statistician can do the tests for himself and, using standard confidence intervals, 95 to 99 percent would be led to the same conclusion to, reject the hypothesis that the drug industry is out of the normal pattern. Senator NELSON. In your judgment, what are the most important factors that go into your conclusion that the drug industry is a high- risk industry? Mr. PLOTKIN. The most important factors that go into my conclu- sion are that I believe that I have a good and workable hypothesis of risk and return in the American economy, that higher returns are given with higher risks. I believe, and we have submitted this to arbitration of scholarly members of my profession, in MIT, Chicago, Stanford, the Rand Corp., that there is a workable definition of return and risk proposed, and, fitting that pattern, there is no evidence of a violation of the nor- mal pattern. I do not know, have no expertise, have no prior beliefs, about why this point should be there and not down here or not here. I am speak- ing only from the evidence. I am raising this to a contending hypothesis which someone else might come up with the alternative quantifiable measures. PAGENO="0102" 1652 CO~ETITIVE PROBLEMS IN THE DRUG INDUSTRY If economics has made any progress in the last 20 years, it is our ability to take things out of the field of judgment, out of the field of authority and whose textbook you read, and into objective, verifiable, repeatable experiments. Senator NELSON. The first students of aerodynamics came to the conclusion that the bumble bee could not fly because in proportion to the body the wing size was not big enough. But he flies pretty well. You used 29 companies to represent the drug industry? Mr. PLOTRIN. Twenty-nine companies. Senator NELSON. How many of these companies have lost money in this period; is it 17 years? Mr. PLOTKIN. Sixteen years, sir. Senator NELSON. In the 16 years that you used, did any one of those 29 companies have a loss year? Mr. PLOTJIIN. Yes. The answer is yes. These facts are in the report. May I once again-I submit that. we do have a rather subtle division of labor, but I think that the committee benefits by the expertise of this division. Mr. Conrad, of the Arthur D. Little organization, is primarily in charge of the relevant information with respect to the pharmaceutical industry, which firms were comprised in it, how it reflects on it, and how his additional studies, together with those of Professor Markham, go to sa.y why this point is to be expected, why the risk should be high; and I would respectfully ask that these questions be deferred and be integrated into his testimony. Here I am just prepared, at this point, without any expertise other than a personal opinion which I have eschewed, discussed the general validity of such measurement. Mr. GORDON. Mr. Plotkin- Mr. PLOTKIN. Yes, sir. Mr. GORDON (continuing). In the study that you have presented to the committee on the proposed theory of risk and return, on page 12 you say: In order to perform these statistical tests, it was necessary to construct a quantitative measure of the industry risk (or uncertainty). We selected measures of the dispersion of individual companies' rates of return about their industry's average rate of return for a given year. Mr. PLOTKIN. That is correct. It is so worded. Mr. GORDON. Now, it seems to follow from your definition that if there were two industries-let us take two industries, A and B, each with two firms, A-i and A-2, B-i and B-2. Now, A-i earned 25 per- cent on its investment, or its assets, after taxes. A-2 earned 15 percent on its investment, on its assets, after taxes. B-i earned 9 percent. B-2 earned 7 percent. Maybe you ought to write that clown. Mr. PL0TKIN. Fifteen, twenty-five, seven, nine. Mr. GORDON. A-i earns 25 percent, A-2 earned 15 percent, B--i earned 9 percent, B-2 earned 7 percent. Now then, industry A, according to your definition, would be the high-risk industry; is that not correct? Mr. PLOTKIN. That is not correct, because you would have to further hypothesize that this was the case year-in and year-out., and also- Mr. GORDON. All right. Let us take it on a year-in and year-out basis. PAGENO="0103" COMPETITIVE PROBLEMS IN THE DRUG INDUSTRY 1653 Mr. PLOTKIN. And also that those two firms-I am wondering if this is part of your hypothesis-were the entirety of the industry, not a sample thereof, but the entirety of the industry. And if you have only two industries in your economy, the entirety of the economy~ Mr. GORDON. All right. Let us assume that I said there were two industries. Mr. PLOTKIN. I wanted to have~- Mr. GORDON. And two firms in each industry. Mr. PLOTKIN. I wanted that clarified. Mr. GORDON. Let us say, year-in and year-out they earn 25 percent, 15 percent, 9 percent, and 7 percent. Industry A, according to your definition, would be the high-risk industry; isn't that correct? Mr. PLOTKIN. Sir, this is correct. This is a case, and this is typical of cases we considered in drawing up this definition. Mr. GORDON. All right. Let me go just a little further. It would also mean that the better firm A performed, all other profit rates remaining the same, the riskier industry A is compared to B. That is also correct, is it not? Mr. PLOTRIN. Defining risk as the uncertainty of taking a blind choice of getting firm A-i or A-2, and knowing \~hat your outcome will be, that is correct. Mr. GORDON. It also means, does it not, if A-i and A-2 maintained their profit rates, but B-i and B-2 kept changing back and forth, say, from seven to nine, industry A is still riskier, is it not? Mr. PLOTKIN. Yes; under these hypothetical situations, that is true. Mr. GORDON. All right. Now, Mr. Plotkin, are you going to tell us if you are assured of profits of between i5 and 25 percent, even though your exact rate of return may be uncertain, that this is a riskier use of your money than putting it where the return may be between 7 and 9 percent? Mr. PLOTRIN. I am going to tell you that, to the extent that one can give me that assurance. In the hypothetical economy which you have drawn, reminiscent of the Senator's hypothetical bumble bee of the pre- vious example which could not fly, that would be true. Our definitions were not meant to cover cases not relevant to the economy as we know it. This is not a textbook exercise. It is true they break down like any `other econometric definitions when carried to the extreme, to reductio ad absurdum. We performed specific tests against the robustness of these de~ uctions, against just the type of hypotheses you mentioned, to see whether they hold true in the real world, and whether they would affect those definitions. We have found they would not. I would like, though, to ask Professor C'ootner's comment especially on the point, or Professor Markham, that there is no possibility in industry A of earning less than industry B. The possibility of that describing a real world phenomenon, and what the requirements of that then would `be, I would like to ask them to comment. Mr. GORDON. Mr. Plotkin, it seems to me that the record of the high and stable profitability of the drug industry will speak for itself on the matter of evidence of risk. Mr. CUTLER. Mr. Gordon, we have `asked ourselves these very same questions that you are putting now, to see whether the range, the dis- PAGENO="0104" 1654 COMPETITIVE PROBLEMS IN THE DRUG INDUSTRY persion, of rates of return in the drug industry data that are included within this study, goes frequently below either the drug industry average, or indeed, the all-industry average, and Mr. Conrad has made a statistical analysis of that that he would be glad to tell you about. I think we will show you that your hypothesis A is not the hypothesis of the drug industry. Mr. GORDON. I would like to continue with Mr. Plotkin. He is the one who is making the statements, and he is the one who presented the study. We will get to Mr. Conrad in just a few- moments. Mr. CUTLER. I believe we will let Mr. Conrad-I mean, Mr. Plotkin, continue, but I believe the chairman said that if we thought a par- ticular witness can answer a particular question better, we can have that witness do so. Senator NELSON. That is correct. Anybody who wants to answer, may answer. Mr. CONRAD. My name is Gordon Conrad, from Arthur ID. Little, Inc. In looking at the data bank that we have for the drug industry, and attempting to see what kind of fluctuations actually occurred in the data over the 16-year period, we ran an analysis w-hich took each data point for each company, for each year in the study, and analyzed the distribution of profit performance that relates to that chart for the total number of companies, and related this to the average re- turn on investment for all the industries shown on that chart. The average return for all those industries comes out to about 10.75 percent for that 16-year period. Now, for the 16-year period, about 24 percent of the drug industry points fell below that average. Four percent of those- Mr. GORDON. Excuse me. Would you repeat that once again? Mr. CONRAD. Twenty-four percent of all drug industry data points out of the total number of data points for the 16-year period again 24 percent fell below the all-industry average return on investment level of 10.75 percent. Mr. GoRDoN. Are you saying that 25 percent in the last 16 3-ears, the average profit for the drug industry fell below the- Mr. CONRAD. No. Mr. GORDON. What are you saying, precisely? Mr. CONRAD. If you take each company and each year and give it a weight of one, and add up all the company years, 400 points, it comes out to-in other words, you have 400 actual data points over the 16-year period, and of that 400 data points, 24 percent fall below the all-industry average rate of return on investment of 10.75 percent. Saying this another way, in 24 percent of the years by company there were returns that were lower than the industry average, the all-industry average. We are not relating this at all to the drug industry's average return value which is shown on the chart. So, yes, then we are down- Mr. GoRDON. Can we get the raw material on which you have based this study? Mr. PLOTKIN. You have the raw material in the report. The num- bers are spelled out. Each company year value is given. There are a number of negative values given, also, we would point out. PAGENO="0105" COMPETITIVE PROBLEMS IN THE DRUG INDUSTRY 1655 So that is why I answered your question as I did, that it is true, one can posit situations under which the situations would break down. Whether there is such gain in such positing depends on how relevant both situations are to the real world. W7e submit not through judgment or opinion, but through analysis, they do not represent real world situations and do not affect the va- lidity of the results. Senator NELSON. Risk is a relative thing, is it not? Mr. PLOTKIN. We say the risk/return relationship is relative, one industry versus another. It is not an absolute quantification, but it is a measure such that when it is higher, the risk of investing in that in- dustry is higher than the risk of investing in an industry when that is low-er. It does not say that it is worth $10 or $20, or something like that. Senator NELSON. But you could project your chart which now cov- ers 16 years to 50 years, and if you came out with exactly the same result, would you come to exactly the same conclusions? Mr. PLOTKIN. I would, sir, because what the chart seeks to do, does not matter as long as you have some minimal number of years so you are certain that it is not a perturbation that you are measuring, the chart seeks to look at the ex ante, if Imay use the technical word, the risk that appears to a person before he takes the plunge, he makes that decision based on past information. Let us see where is the relevance. Before a drug company president can commit capital to research and developing a bioactive agent, he must ask himself whether that is the most profitable use of his capital for the risks involved in that venture. He repeatedly asks himself that question, or would he be better investing in IBM stock, buying the telephone company, or, as some of our insurance companies have come to conclude, buying out credit card companies, rather than in- vesting in insurance. If he feels that the prospective risks of the individual ventures are not justified by their rate of return, he will not commit capital to those ventures. This analysis addresses itself to quantifying the prospective risks based on the simple hypotheses, given those two ventures, which one is more risky, which one is more uncertain. This is not a complete theory of profitability. It is not-I would be both naive and boastful to claim that I have quantified here the entire American economy. I have taken two important dimensions of it and shown its relation- ship. It is based not on mathematical concepts that a businessman w-ould not be operating under such as-~- Senator NELSON. Of course, when he is making that decision as to where to put his capital, he may very well put it into a company in whmhi the profit is lower than it is in his own business, perhaps for purposes of diversification, or he may do it for a tax loss. He may do it for any number of reasons-noneconomic reasons. Mr. PLOTKIN. The reason you mentioned yourself, the diversifica- tion, diversification is the other side of the coin of risk. It is a way of getting away from risk. Now, what we have to do here is isolate the risk of the particular industry, the drug industry, to see whether capital will flow into there PAGENO="0106" 1656 COMPETITIVE PROBLEMS IN THE DRUG INDUSTRY or will the drug companies become like the fire and property insurance companies, merely holding companies of capital, not actively expand- ing their own markets, but going into other ventures which are, may be, less profitable on the average, but have less risk and which they feel more confident in. This is the thing which worries economists when there is regula- tion proposed. If there is excess profit, profit above the risk level, then there is no harm to the normal flow of capital in reducing the profit. But if the profit is concomitant with the risk level, the harm is a potentia.l threat to capacity in that industry, and that is the only rea- son we commend such a study to your attention, pointing it out not as proponents of specific points of view, but a proponent of a more general point of view that regulation must take into effect and into account, not only the rate of return but the risk in the industry, and asking only that that judgment be made. I am critical, personally, and admit this is a personal point of view-, about the FTC year-in and year-out ranking industries by rate of return, and indicating by implication that ones at the top of the list are bad for the economy. This is a false statement, and cannot be justified by even the most elemental of the textbooks. They should also have another column ranking the riskiness. Senator NELSON. I have not followed it, obviously, as closely as you have. Did the FTC issue a report that those industries which ranked so high are somehow doing damage to the economy, or is that your opinion? Mr. PLOTKIN. It is my interpretation, not from reading only the entire report but from the news release, the way they point things out, there is an implication. In discussing it with my students at MIT, they have read the news release, I have received questions as to why is it. bad to be profitable. Senator NELSON. Did you read the actual report of the Federal Trade Commission yourself? Mr. PLOTKIN. I read the whole, the entirety of the report. But I feel as far as influencing Iublic opinion, the news release is sometimes more potent, and I concentrate heavily on that as well, because I know that not many people are professional economists and will not read the whole report. That is always t.lie danger of epit.imization and summary. Senator NELSON. But I am just trying to get this straight in my mind. Are you saying that the Federal Trade Commission says some- thing derogatory about high profit companies, or are you just saying the very fact that they released the high profit companies somehow is derogatory by implication? Mr. PL0TKIN. I am saying that they release information in such a w-a.y that it has been interpreted, to my knowledge, by not very so- phisticated students of economics as implying some derogatory function. I am not saying that the FTC does imply it, but I would think it would be useful if in addition to listing the companies in order of re- turn, they also listed the companies in order of riskiness. Senator NELSON. Do they have sufficient. information to list them in order of riskiness? Mr. PLOTKIN. They have the finest economists who certainly could create a measure of risk the way they have created it at MIT and PAGENO="0107" COMPETITIVE PROBLEMS IN THE DRUG INDUSTRY 1657 Arthur D. Little in their research, aiid any other university in the country. Senator NELSON. You will concede, however, that there is a large judgment factor in your chart and your own theorem, won't you? Mr. PLOTKIN. I readily concede that and refer back to Professor Cootner's point that no matter how one changes the definitions of risk or the precise definition of rate of return, the basic relationship of higher profits with higher risks, cannOt be killed. This is one of the few times we see economists in agreement of any persuasion, Paul Samuelson and Milton Friedman, for example. Senator NELSON. What I am getting at is, you are asking the Fed- eral Trade Commission to come up with their own theory about what high risk is, and then if a group of you economists disagree, you will start pounding the Federal Government for making that conclusion. Mr. PL0TKIN. I am sorry if my remarks have led to that interpre- tation. I only meant it is sometimes implied that just because one has high returns, he may be doing a disservice to the economy. I was hop- ing this analysis disproves the implication. Mr. CUTLER. Mr. Chairman, unless you get any impression that the Federal Trade Commission's economists would differ from the meth- odology that was followed here, both Professor Markham and Profes- sor Whitney, who follows in turn, were both chief economists at the Federal Trade Commission at earlier stages of their profession. [Laughter.] Mr. PLOTKIN. I certainly have no quibble with my colleagues. Senator NELSON. Being one who majored in economics when I was in school, and having listened to the arguments in those days, I just learned enough to discover that you could find any point of view you wanted from a set of economists any place in the TJnited States, and from my discussions with my friends tOday, I find that is still true. Dr. MARKHAM. If I may intrude on the dialog, Senator, I think you are quite correct when you went to school, and that was true when I went to school, too. If this analysis means anything, it means, I think, what precisely Professor Cootner and the rest who have testified, have said here, and that is as Professor Plotkin has said, this takes the issue really out of the area of opinion, it puts it in the area of quantitative fact. One may have some difference of opinion as to what is the very best measure of risk or, indeed, for that matter, what is the very best meas- ure of the rate of return. But it, at least, reduces the dialog to the meaningful definitional questions. The factual analysis stands on its own feet, and, as has been pointed out, when you have picked someone who may be extraordinary highly sympathetic to Government regulation, of which we may name some economists, or someone who is very much opposed to Government regulation of the economy, and here I can be specific, Prof. Milton Friedman, those economists of different views, political views and dif- ferent opinions, about this sort of thing would still have to, I think, indeed already have, agreed that this establishes the statistical rela- tionship between risk as measured and rate of return as measured, and both the rate of return index and the risk index are very sensible in- dexes that most people would agree measure these respective aspects of the economy. PAGENO="0108" 1658 COMPETJ~fIVE PROBLEMS IN THE DRUG INDUSTRY Senator NELSON. Before we recess for noon, I would like to make just one point. This is an industry evaluation of risk and profit, is it not? Mr. PLOTKIN. That is true. Senator NELSON. By industry. Mr. PL0TKIN. The unit of observation on the chart is the industry. Senator NELSON. So it is no doubt true that. you can reach into one of your industries on the chart, which appears to be a relatively low-rate return industry, and within that industry select, out a very fine com- pany that would rank very high in rate of return. Mr. PLo~I~IN. You could not do that with any assurance, blind- folded. If you looked in with a lot of foresight and particular exper- tise, you might have a chance of doing that. In fact, that is part and parcel of the variance measure that all the companies within the in- dustry, because of the undertaking that they do pursue, do not have the same results. You are perfectly correct, Senator. Senator NELSON. So what I am saying is that what we are looking at here is a chart that measures, according to your theory, the rate of re- turn and the risk of an industry. But within one of the industries which will rank relatively low in rate of return and high in risk, you can select out a company that. has a long record of profit and that would rank high in rate of return and yet would rank low in risk; is that not. correct? Mr. PLOTKIN. It is possible to do that. It does not necessarily mean that there is one in the particular sample which we have used here. Senator NELSON. Well, can't you reach into one of those industries that ranks relatively low and select out one of the companies that does rank high in profit and low in risk? Mr. PLOTKIN. I believe- Senator NELSON. These are averages that you have given. Mr. PLOTKIN. Yes. The answer is- Senator NELSON. Is it true, or is it not? Who knows the companies, who knows the answer? Mr. PLOTKIN. Well, your assertion is right in respect to company years. Senator ~\`ELSON. With the knowledge that you have, since you put the cl~art together and selected the industries from Standard & Poor's, can't you say there is within a higher risk, low rate of return industry, a high-profit, low-risk company? Mr. CONRAD. One point here.: the industry risk factor is an overall factor for that industry, and is part of the riskiness that any company exposes itself to. It is not the total riskiness that a firm exposes itself to. So while it is true that. you can pick from any one of these imiclustry points, a firm that may have high profits for an extended period of tinae, this does not speak to the validity of the average industry risk factor that we have calculated. Again that is only OIIC component of the total riskiness that firm experiences. Now, there will be others. We have not addressed ourselves to the individual firm risk in this particular part of the analysis. There are other analyses that could be `conducted which would then assign par- ticular company risks to each firm. But that does not relate to what we have done here. PAGENO="0109" COMPETITIVE PROBLEMS IN THE DRUG INDUSTRY 1659 Senator NELSON. So you measured each individual company, did you not? Mr. CONRAD. Yes. The whole analysis is put together from indi- vidual company measurements. Senator NELSON. And then from the individual companies, you came to a conclusion about the industry as a whole which represented the averaging out of those who were higher risk, those who were lower profit, lower return; did you not? Mr. CONRAD. Well, the actual mathematical measurement is to measure the distribution of performances-rates of return or total investment-of the companies around their own industry average return. So each firm~s weight is taken into account in calculating a total risk factor. Senator NELSON. Let me get back to my question. Perhaps I am not being clear. All I am saying is that., based upon the statistics that you gathered, you took 29 drug companies. So, if you took the drug industry, couldn't you select the poorest rate of return company in that industry and nome up with one that is below the average? Mr. PLOTKIN. Yes, sir. But it is true not of the average return over the history of the company. This is something that addresses itself to the year-by-year return. We can select a company in particular years whose return is below the drug industry average and is below the economywide average as well. Senator NELSON. Well, let me go back to my original question again. Take one of your companies on the chart which ranks at about the average, 10.5, for industry in rate of return, and there sits the industry ranking right in `the middle. Now, within that industry you have, say, listed 30 companies. Can't you take, examine, those 30 companies, and over the past 16 years pick one out that over the past 16 years has consistently had a high rate of return that would put it, if you used that sole company, put it up at the top of the rate of return companies; isn't that correct? Mr. PLOTKIN. No; that is not the case. You can find One 111 the past 1(~ years whose rate of return has been both above and below. About tIre company's 16-year average, we `can make no stateme.nt~, sir. Senator NELSON. Why not? Mr. PLOTKIN. Because that is not what the analysis is addressed to. it is addressed to the profitability of ensuing ventures, the marginal profit, what will next year's return be.. Senator NELSON. What you have given us is an industry average. All I am saying to you i's that the average is made up of X number of companies, maybe 30. In that grouping. you are going to have a corn- Iany with strong management, strong sales, goodi promotion, good research, good recognition in the market place, and regularly high profits. In the same group you are going to have one with average manage-. ment, average sales, average research, average everything else, and below average return `in the market; are you not? And you are going to have one that has a 50-year history of good returns and no losses, or very few losses, and one with very lOw average rate of return; are you not? It is true of every other industry. PAGENO="0110" 1660 COMPETITIVE PROBLEMS IN THE DRUG INDUSTRY Mr. CUTLER. You can undoubtedly find companies, Senator, in the data we have given you that had a consistently high rate of return, drug companies, year after year. Senator N~soN. Yes. Mr. CUTLER. But I think the fallacy of what you are trying to say is best illustrated by the New York Yankees, who won the pennant 9 years out of 10. On your theory, they became a low-risk enterprise. It was easy for them to win the pennant. But in the 11th year, they finished ninth. A company which has a consistently high rate of return in a high-risk industry has the same risks on its new invest- ments every year, even though it has batted out, let us say, a 25-percent rate of return in each of the 10 preceding years. That is the point these economists are trying to make. Senator NELSON. I am not trying to work out a theory. I am just saying that when you collect 30 companies that have been in business for 10, 20, 30, 40 years, you can find companies that have consistently clone very well, and you can find companies that have been consistently below your curve; isn't that correct? Mr. PLOTKIX. No, sir. I do not, depending how it is- Senator NELSON. Let me give you General Motors as an example. Compared to the rest of the auto industry, they have done well. If you took the years when Chrysler was falling apart and went from 23 percent of the market to 12, and American Motors went from four or five down to less than one, and Ford dropped off from 23, and General Motors occupied about 50 percent of the market, can't you have a case that General Motors consistently, compared with all the rest of the industry, has done well in the marketplace? Dr. MARKHAM. I think, Senator, the answer is, yes. Could you, in looking at General Motors now, their investments, infer that it is goingto make 20 percent as a rate of return from that new investment? The answer is, you cannot infer from the past history, given from these past indexes, that General Motors will make 20 percent on, say, a new model coining out the next year. Senator NELSON. You cannot do that in any industry? Dr. MARKHAM. You cannot do that in any industry. Senator NELSON. Then it is irrelevant. But my point is, what you are talking about here, and it may be a very fine theory, I am not qualified to judge it, but I am just a layman listening to your testimony. I am saying you come in and take an industry in which there is good and bad, and then you plot it out on a curve. Let us assume that you are correct, and that all knowledge- able people will agree with you. All I am saying is that within the industry selected there are in- dividual companies that have a record that is good enough so that investors are going to say, "That is where I will put my money", and then when you talk about the drug industry, if that is the kind of company you have, a firm that is doing all right in research and development and is making a good rate of return. I have never heard of anybody investing in an industry. I have heard of people investing in companies within an industry. That is the only point I am trying to make. This theory may have validity as to what the whole industry represents, but there is some questionable validity about it as applied to a particular company within an industry. PAGENO="0111" COMPETITIVE PROBLEMS IN THE DRUG INDUSTRY 1661 Mr. CONRAD. I think we certainly have been cognizant of this problem and we are sensitive to it, and I think the following testimony that we have, both the report that I am putting in and the joint work that Professor Markham and I are doing, speak directly to this issue, and will take a bit of time to explain. But I think we will clarify even in an industry where you have good performers and poor performers, the good performers may well be subjected to the same risk influences that the poor performers will be subjected to. I think we are going to try to illustrate that point for you here. Mr. CUTLER. Mr. Chairman, Mr. Conrad's and Professor Mark- ham's statement is just two and a half pages long. If we got that in before lunch, we would be through with this group except for your questions. Senator NELSON. I will have a few more questions. Why don't we just break for lunch right now. We will adjourn until a quarter to 2. (Whereupon, at 12 :50 p.m., the subcommittee recessed, to resume at 1 :45 p.m., this same day.) AFTERNOON SESSION Senator NELSON. Come to order, please. Mr. Plotkin, your report does list the companies involved in each one of the industry selections, is that correct? STATEMENT OP IRVING H. PLOTKIN, CONSULTANT IN ECONO- METRICS AND OPERATIONS RESEARCH, ARTHUR D. LITTLE, INC.-Resuined Mr. PLOTKIN. That is correct, sir; and the number of companies in each industry. Senator NELSON. What was the basis of the selection? Mr. PLOTKIN. It was derived from the Standard Statistic Corp., a subsidiary of Standard & Poor, Compustat Industrial Tape. This is a magnetic computer tape which contains, in essence, balance sheet, income statement, and stock market data on approximately 800 cor- porations in 59 industries. The period was 1950 through 1965. Senator NELSON. Did you use the same number of companies in each industry? Mr. PLOTKIN. No, sir; I used in every point all the data that were available to me on the tape. This gave a different number of companies in each industry. In general, these companies are the larger, more successful, actively traded corporations because it is in the interest of Standard & Poor to collect them. Senator NELSON. So in the auto industry, you used all four of the auto companies, is that it? Mr. PLOTKIN. I think since the period went back to 1950, we may have as many as five in one field. Mr. CUTLER. The number of companies in each industry, Mr. Chair- man, appears on pages 34,35, and 36. Senator NELSON. On what basis does Standard & Poor list the companies? PAGENO="0112" 1662 COMPETITIVE PROBLEMS IN THE DRUG INDUSTRY Mr. PLOTKIN. They have experts in their organization in each indus- try using the Department of Commerce SIC, standard industrial clas- sification numbers, based on the type of product sales of the individual companies, who will assign, admittedly arbitrarily, companies to various industries. Because this is a standard reference source, we have elected to use it rather than make up our own industries, which could be somewhat objectionable or at least questionable as to whether the results were influenced by our own selection. As I admitted earlier in my statement, I have some serious qualms about particular items-should one company be here or there. But we have done careful studies to show that the results are not sensitive to any one company moving in or out of an industry. Sir, if this would help, I might add that these data sources have been used for similar studies in a number of university campuses-MIT, Chicago, Stanford-by diverse scholars for many purposes, always relying on these classifications as a general, widely accepted usage. The First National City Bank, in their monthly letter, often lists in- dustry by industrial classification, and it is quite similar to the ones appearing on Compustat tape. Senator NELSON. You used 29 companies in the drug industry? Mr. PLOTKIN. Yes, 29, sir. Senator NELSON. Is that all the drug companies listed in Standard & Poor? Mr. PLOTKIN. To the best of my knowledge, it is all the drug com- panies listed in Standard & Poor. It even contains a company of exceedingly high profitability which, by their own admission, Standard & Poor now say should not be in there. That is the Gillette Co. Rather than take the chance of influencing the outcome by my personal selec- tion, and Arthur P. Little take a chance by its exercising judgment, we decided to go by what is an industry standard and used in all academic circles. Senator NELSON. In what groirnds does Standard and Poor's select. them? Eliminating Gillette, are these the 29 largest companies? Mr. PLOTKIN. To some extent, they are the largest in that these would be the actively traded, the ones that information is most readily ob- tainable on. If I can paraphrase the Standard and Poor's description, they try to select that group of companies actively traded on the New York and American Stock Exchange which represent the overwhelm- ing majority of assets or sales in any given industry. Obviously, this necessitates a different number of companies as industries are more or less concentrated. In the autos, with four, you have 100 percent of them. Senator NELSON. Are any of the large drug companies closely held and not actively traded? Mr. PLOTKIN. Squibb, I believe, is closely held. It certainly is not. actively traded, and for the Standard and Poor's standard, it is not included in here. Mr. Conrad, who is an expert in the chemical field, in trading in that field, might be better able to elucide on this point. Mr. CONRAD. Standard and Poor's data here includes whatever they could gather that was publicly held and also what they could back up in discussions with the companies and through SEC registration state- ments, proxy statements, and so forth. There are eliminated from this PAGENO="0113" COMPETITIVE PROBLEMS IN THE DRUG INDUSTRY 1663 a number of larger closely held pharmaceutical firms where there is no data available. Senator NELSON. Which are the large, closely held pharmaceutical firms? Mr. CONRAD. Well, E. R. Squibb is part of Olin Mathieson. Hoffman LaRoche. Senator NELSON. Ciba? Mr. CONRAD. Yes, Ciba.. They are not in the sample in the report. There are many that are not. We could not take other than what the data source allowed. Senator NELSON. Schering? I don't see that. Mr. CONRAD. Yes, Schering is included. Senator NELSON. Well, do you have a. list of the substantial com- panies that are not actively traded and therefore are not included in the list? Mr. CONRAD. I do not have tha.t at my fingertips, but we could put that together for you. Senator NELSON. Could you furnish that to the committee? Mr. CONRAD. Surely.' Senator NELSON. If my memory is correct, the PMA states that their members produce about 95 percent of the prescriptions used in this country. Is that correct, Mr. Cutler? Mr. CUTLER. Yes, sir. Senator NELSON. What percent of the drugs produced is represented by the companies listed here in your report? Mr. CUTLER. I do not have that information, sir, but if we took a guess, say, in the range of 75 to 80 percent., I am sure we would not be more than 3 percent off. Senator NELSON. Would you have the figure? Mr. CUTLER. No. Senator NELSON. I mean do you have it available? Mr. CUTLER. No, we have not added the figure. I think we could com- pile a figure as to these 29 companies. Senator NELSON. Would you supply that for the record? Mr. CUTLER. Yes, we will.' I think the only unincluded companies are the American subsidiaries of Swiss companies and a few others, like Squibb, which are part of a much larger conglomerate, whose overall results do not really reflect the earnings of Squibb. Squibb's earnings are just not available to us. But they were given to this coin- mittee, as I recall, by `Squibb's president. Mr. PLOTKIN. An important reason why we did not go beyond the confines of the Standard & Poor's tape just for the drug industry was our desire to be evenhanded in treating all industries which appear on the chart. Not having access to nonpublic information would mean an unfair comparison. Therefore, whatever `bias appears in the drug company, the same bias is present in all other companies, so the rela- tive distribution does not change. Senator NELSON. To get back to the question we were discussing be- fore recess, if a company were included in an industry grouping that was high profit and high risk, but that company itself had a long, 1 See Appendix III beginning at p. 2129, Infra. S1-280---OS-pt. 5-S PAGENO="0114" 1664 COMPETITIVE PROBLEMS IN THE DRUG INDUSTRY long record of high profitability, without any significantly bad periods, is that still called a high-risk company? Mr. PLOTKIN. That is a company still participating in a high-risk industry whose new investments and future investments are subject to the same vicissitudes as those undertaken by any other company in the industry, sir. Senator NELSON. As I recall the testimony this morning, whoever w-as responding to the question, if a company had a 20-year record of stability and good profitability, the conclusion by one of the witnesses was that then at some period, say after 20 years, it would be con- sidered a low-risk company. Mr. PLOTKIN. I think it is Professor Cootner's statement to which you refer and I will turn the mike over to him at this point. Dr. COOTNEP~. The statement I made was that a 20-year period was sufficient for discussing the riskiness of an industry. But I did not make reference to the return of any individual company, I was saying in other words, the 20-year record of the drug industry was a sufficient basis on which to make your decision. Senator NELSON. But you do not say that about an individual company? Dr. COOTNER. I think I can give you Senator NELSON. You would not draw that conclusion about an indi- vidual company within a high-risk industry after 20 years of stability and high profitability? Dr. COOTNER. Well, again, I want to refer back to my previous testimony, wthere I pointed out that a company could have a high rate of return for 20 years but still show substantial volatility in changes reflecting the very variable rate of return it had on its marginal in- vestments. Just looking through Mr. Plotkin's study, I can give you a specific reference-for example, the Kendall Co., which has a greater return, which goes from 14 percent steadily down to-I am afraid the number is not completely clear, but it looks like 7 or 8 percent, year after year. Now, the rate, if you looked at the variability of rate of return from year to year, `it is rather small. But the sharpness of `that decline over the entire period suggests that the marginal investments were relatively unprofitable. Senator NELSON. What is a marginal investment? Dr. COOTNER. That is an additional investment beyond the invest- inent you have at the beginning of the period. In other words, these companies will be substantial companies at the beginning of the period. The new investment that they undertook- Senator NELSON. In the same product line? Dr. COOTNER. In the same total product line. That is, I am not talk- ing about any specific product, but I mean their total investment in the drug industry. Senator NELSON. Well, then, supposing you have a company with a 20-year record of stability and good profits, consistently above the average in all industries, maybe a fair degree above it, you still would not conclude after 50 years that that company was a low-risk com- pany? Is that what you are saying? PAGENO="0115" COMPETITIVE PROBLEMS IN THE DRUG INDUSTRY 1665 It would seem to me if you look at some company over a period of time, perhaps a 50-year period, it would be difficult to conclude that either the investor or anybody else can assert that a stable company consistently making profits above average for all industries is not a low-risk company, apart from the industry as a whole. Mr. CONRAD. One answer to-Professor Cootner, do you want to take that first? Mr. PLOTKIN. I lust have a point as it relates to the industries and then to the companies. The statement about 20 years being sufficient time in an industry, and we are not quibbling about the precise number, 16 or 20, was not by looking at the time pattern for 20 years, but if we average 20 individual years' variances that would be sufficient to get a risk idea. We never claimed that we could look at whether or not the industry average fluctuated over 20 years and learned anything about risk. In fact, referring back to my testimony, we explicitly rejected the hypothesis that the best place to look~ for risk information is in the time pattern of industry or company results. This is something that is very commonplace and has been used sometimes in economics, but without good results. It seems to not be able to explain in the sense that this line explains in rates of return the true phenomena in economics. Now, if you would like an additipnal memorandum on this, we could submit it to you why in economic terms the temporal variance, as it is called, the time variance, is not a good measure of uncertainty, does not let us extrapolate into the future with any certainty. Mr. Conrad can address himself to the individual industry now. Mr. CONRAD. To answer the point you raised for this company that has had 50 years of high profit, high and steady profitability, I think the correct statement that one might make about that would be that the company in its past history was in a relatively low-risk situation. But that does not necessarily allow you to say anything about that company's future possibilities. That company may today be in a condition where it faces extremely high risks, even though for 50 years, it has not. You cannot, a priori, say that because those 50 went by and the company earned a high profitability, that tomorrow it can't turn around and either lose money or suffer a significant decline in profitability. Senator NELSON. Let's take an industry which does not require a high rate of technical skill to enter. Now, I will mention an industry in a moment. I do not mean to say that those who stay in it and are successful do not have lots of skill, but anybody can get into it and tens of thousands of people do-that is, highway construction. Now, everybody who can buy a truck and a grader or a ditchdigger can enter the industry. As you probably know, tens of thousands of people are in business who have just one piece of equipment. It is a ditch- digger. It costs $25,000. You can borrow the money to get into it. They may not have capital behind it; they may not have adequate bidding knowhow, or perhaps they do and they go into big business. But most of the construction people I am familiar with, and I am famil- iar with quite a few, started out on their own, with their own truck, or if it is an old company, went in with their fathers. Now, there are PAGENO="0116" 1666 COMPETITIVE PROBLEMS IN THE DRUG INDUSTRY loads of very fine highway construction outfits with a long, long stable history in every State in this Union-highly profitable, highly suc- cessful. But if you took a. look at that industry as a whole, you would have to, I would think, by your measurement or any other measure- ment, say it is a. high-risk industry. A lot of people get into it who do not belong there.. But it is not. a high risk for the people who have the capital and the knowhow and get into the. business. Now, how do you evaluate that factor? Mr. CONRAD. One point I would like to make is that we hope to show by the testimony that will follow now that in the drug industry, at the point we are at, regardless of performance in the past., the finns, the highly successful firms, are subject to a set of risks which could put them into a great deal of trouble. Senator NELSON. A set of risks that. are not common to or comparable with other types of risks, in many other industries? Mr. CONRAD. Correct. They are somewhat, in some sense, peculiar to the pharmaceutical industry. Those firms, large or small, successful or unsuccessful, are all subjec.t to these same risks in the pharmaceu- tical industry. Senator NELSON. I read the list of factors-some seven of them, I believe, that you list- Mr. CONRAD. Five, I think. Senator NELSON. Well, whatever it is. But anyway, I did not see anything in your evaluation. You listed what you considered to be peculiar risks, some of which, when we get to them, I have doubts about whether they are peculiar risks. I think they are a security factor in the industry. But did you consider other factors of special insurance, so to speak, in the marketplace that no other industry has when you made this evaluation? Such as the fact t.hat a drug established after 17 years in the marketplace, with its patents, still coiitinues to hold a high volume selling at several times the cost of comparable drugs? Did you consider that protective factor which is not found, to my knowledge, in any other product area in the marketplace in this country? Did you consider that? Mr. CONRAD. We did and we will address ourselves to that. Mr. CUTLER. Sir, if we could go to Mr. Conrad's and Professor Markham's paper right now, I think these questions are germane to that paper. Senator NELSON. All right, but be.fore we do, the biographical sketches of these gentlemen will be inserted in the record at this point.. (The biographical material follows:) BIOGRAPHICAL SKETCH OF JESSE W. MARKHAM, PH. D. Present position: Professor of Economies, Princeton University. Education: A.B., University of Richmond, 1941; AM., Harvard University,. 1947; and Ph. D. Harvard University, 1947. Professional experience: Has served on the faculty of Vanderbilt University, Harvard University and Princeton University. Acting Director and then director of the Federal Trade Commission, Bureau of Economies, 19~3-55. Has acted as consultant to the Federal Communications Commission, the Brookings Institution, the U.S. Department of Justice, the Tennessee Valley Authority and the Office of the Special Representative on Trade Negotiations. Participant in a Ford Foundation inter-research program on the economics. of technological change and economic growth. PAGENO="0117" COMPETITIVE PROBLEMS IN THE DRUG INDUSTRY 1667 Author of numerous publications, articles and reviews in the area of public policy towards business. Among others, his books include: The Fertil- izer Industry: Study of an Imperfect Mar1~et (1958) and The American Economy (1963). _______ BIOGRAPHICAL SKETCH OF GORDON H. CONRAD I took my academic training in chemical engineering at New York University and later continued with business administration at Columbia Graduate School of Business. Since 1956 I have been employed in the Industrial Economics Section of Arthur D. Little, Inc. conducting market research, economic feasibility and long-range economic planning studies in the chemical and allied process indus- tries. During the last two years I have been in charge of a number of economic studies relating to the operations of the phamaceutical industry which have been sponsored by the Pharmaceutical Manufacturers Association. Dr. MARKHAM. Mr. Chairman, and members of the subcommittee, the short statement I have is to report on the study that Mr. Gordon `Conrad of Arthur D. Little and I have been conducting for the past month or two, one of us has to read it, since it is not very appropriate to do this in the form of a duet. So we tossed, and I am going to put it in the record. STATEMENT OP JESSE W. MARKHAM~ PH. D., PROFESSOR OP ECO- NOMICS, PRINCETON UNIVERSITY, PRINCETON, N.J.; AND GOR- DON R. CONRAD, SENIOR STAFF ASSOCIATE, ARTHUR D. LITTLE, INC., CAMBRIDc+E, MASS. Dr. MARKHAM. My name is Jesse W. Markham and I am here largely as a consultant to Arthur D. Little Co., Inc., a company with which I do some consulting from time to time. We appreciate the opportunity to appear before your committee and report on some additional studies we are making at Arthur D. Little, Inc., concerning the relationship between risks and profits in the ethical drug industry. The study just presented to you indicates that there is a strong positive relationship between risks and profits in American manufacturing industries generally. Since this committee is inquiring into the economic functioning of the pharmaceutical in- dustry, the question naturally arises as to what specifically are the elements of risk the pharmaceutical industry confronts. We have been engaged in a study `designed to shed light on this question. While the study is far from complete, some of the preliminary findings should be of interest to this committee. * The risk/return study just presented shows that the pharmaceutical industry ranks high among manufacturing industries in terms of risks. Just what are the components of these risks? Our current study indi- cates that among the major risks a pharmaceutical firm confronts are those unforeseeable events beyond its control that could, in a relatively short period of time, wipe out a substantial portion of its total profits and total revenues. We have, with reasOnable certainty, identified the following as major risks somewhat peculiar in this regard to the pharmaceutical industry: 1. The development of a complete product superior to one of the company's major products which causes virtually complete replace- ment of it in a short time. PAGENO="0118" 1668 COMPETITIVE PROBLEMS IN THE DRUG INDUSTRY Senator NELSON. May I interrupt a moment? Do you consider this risk higher in the drug industry than any other industry? Dr. MARKHAM. Not higher than in all other industries; no, sir. I elaborate on this a little bit later on. I can anticipate that or `go on to it, but I do have a paragraph that tries to tell whether or not this is fairly unique of the drug industry in terms of the incidence of the impact of a new product. Senator NELSON. Do you consider this risk of t'he development of a competing product superior to one of another company a poor risk factor in the drug industry? Dr. MARKHAM. Yes, I do. Senator NELSON. Can you give me a list for the last 15 years of the drugs that have been replaced by a superior competing product.? Dr. MARKHAM. I do not irnow whether I could do this or not over the past 15 years. We can certainly give you some fairly striking examples. I do not even have an exhaustive list of those with me now, Senator.' Senator NELSON. I understand you to be talking about the develop- ment of a competing product superior to one already marketed by another company, as one of its major products, which causes virtually complete replacement of it in a. short time. I do not `think an example here and there proves anything. I am assuming that you came to the conclusion that this was an important risk factor after a detailed study of examples of where that happened and where the company marketing the original product as a conse- quence suffers substantially. What I am interested in for the record, in addition to your conclusion, are the factors which caused you to reach the conclusion. For the record, I would like to see the name of the drug, the company involved, and I would like to see the impact of this particular case upon the earnings, the profit of that particular company. Because if it did not have any impact or a substantial impact on their profits, I cannot see that it means very much. Dr. MARKHAM. I am delighted that you want it and so do we. Let me elaborate. We have not finished this study yet. We have some preliminary results. That particular aspect of the study has been formulated. We have not gotten yet the amount of data that would make it worth while to submit it at this time. What we do have, how- ever, are the data that you will find in "Trends in Market Shares for Ethical Products," which, in an overall sense, documents the point pretty well, that products are knocked off the leading list very frequently. If you will take a look at what firm ranks No. 1 in a given therapeu- tic class, 2, 3, 4, and 5, and then look at where these firms rank 5 or 6 years from now~ Senator NELSON. When you say a. given therapeutic class, you are talking about antibiotics or~ Dr. MARKHAM. Diuretics, things of that type, yes. Senator NELSON. But you again reach your conclusion as to item 1 without having an example of a series of compelling specific cases that is large enough to prove the case. 1 See Appendix III beginning at p. 2129, infra. PAGENO="0119" COMPETITIVE PROBLEMS IN THE DRUG INDUSTRY 1669 Dr. MARKHAM. No, we have some examples, Senator. For example, Merck in production of Diuril hit very hard all the producers of diuretics and it hit it instantaneously at the time it was introduced. We have not, let me make it clear, built a complete catalog of all such instances. I referred, however, to the interchange in range of firms which largely goes at the rank of their product in that particular therapeutic class. You could-evidence presented here and contained in the hearing before the late Senator Kefauver's committee as well shows a good deal of turbulence among the various products. The only way you can, for example, be dropped from one to where you do not even appear in the top five is that some product came and knocked you out of that list. When you see firm B, in other words, displacing firms A, C, D, E, and F, then one has to infer from this that a product has knocked another product off of the list of the top five. Senator NELSON. Well, when they have been knocked off the list of the top five, you are talking in term of total sales? Dr. MARKHAM. No, I am talking in terms of sales of that particular product, therapeutic class, yes. Senator NELSON. Are you talking about amount or dollar sales? Dr. MARKHAM. I am talking only about rank. Senator NELSON. Rank in total amount of products sold? Did you consider when you used this factor that you can have less than 50 percent of the marketplace and substantially more than 50 percent of the dollar value of the sales in this~ business, this industry? Did you consider that? Dr. MARKHAM. I have not considered it. There is a study-I do not know whether it is in evidence or not, Senator Nelson-that puts this, I believe, in terms of dollar value and in terms of- Mr. CUTLER. Mr. Chairman, perhaps I should offer now this study entitled "Trends in Market Shares for Ethical Pharmaceutical Prod- ucts," and the accompanying statement of Mr. Gordon Conrad, which you have before you, which we did not plan to read but which we planned to testify about. For each of the pharmaceutical classes, we do break down the sales by product on a prescription basis and on a dollar basis so you may make those comparisons.' Senator NELSON. On the number of prescriptions and the number of dollars involved? Mr. CUTLER. That is right; for the sale price. Mr. GonioN. But this deals only with products. It does not deal with firms; is that not correct? Dr. MARKHAM. I think we have some additional evidence also on firms in terms of how they rank among the leading pharmaceutical manufacturers. I assume this is put together in terms of total revenue as well. Mr. GolmoN. Just a second. It does not tell us, for example, how much investment was put into these. Suppose a drug is made obsolete. The firm producing the drug may have recouped their money in 2 or 3 years. It may have had very small investment. Now, it seems to me that what we ought to have is an exhaustive-not just a sample, but an 1 The prepared statement of Mr. Conrad and accompanying study, "Trends in Market Shares for Ethical Pharmaceutical Products," begin at p. 1784, infra. PAGENO="0120" 1670 COMPETITIVE PROBLEMS IN THE DRUG INDUSTRY exhaustive-history of those drugs which have been superior to one of a company's major products and caused virtually complete replacement in a short time. These are the conditions given in your statement. Dr. MARKHAM. Well, one slight modification to that "causes virtu- ally complete replacement." I am aware of the fact that many drugs that lose their commercial position for the firm are still held on. They appear among the drugs that are offered for sale. The point that I am trying to go into here is that this is an industry in which you get rapid product innovation. I do not think anyone denies this. Mr. Goiwox. Dr. Markham, we are talking al)out this particular one risk that you have mentioned as risk No. 1. Now, in addition to the exhaustive list which I hope you will supply us with, as the chairman suggested, that you will supply us with the investment for that particular product which has been made obsolete; what the investment for that particular product recouped; how much money was lost, if any; how much money was spent on research and development by the firm for the product; as the chairman asked you, what happened to the profits as a percentage of net worth or on invest- ment of the firm; whose work was made obsolete. Dr. MARKHAM. I think that last we can give you.1 I acknowledge the relevance of what you are asking for. I think I can categorically state for you you can never get that information. The amount that is spent on research and development on a particular product is philosophically impossible to determine as well as from an accounting point of view impossible. The number of times you make an experiment and you come back with the answer no, for example, the number of compounds that ou test and find that they are of no immediate therapeutic value, are nevertheless a part of the research and development cost that in a sense should be charged to some of the successful products. I do not think that any set of books will reveal for you or to me in the terms that I would like to have them all of the investment, including resea.rch and develooment, technical research and development, including clinical testing and all the other costs that go-I just do not believe that you will find, in other words, profit-and-loss statement by product line. The companies have at least convinced me they do not `have these data. I am going to, in a moment, come to some possible estimates that might be relevant to this point, but you are just not going to get those data and I do not think-I would be less than honest if I said I would try to get them, implying that I could get them for you. Mr. GORDON. In my opinion, Dr. Markham, that is very essential data, because a drug can be made obsolete after the company has made a considerable amount of money on it., even after its investment has been rec.overed. I would not call that a. risk. Dr. MARKhAM. I think you would concede-let us take this last set of data you asked for. If he can show, for example, that a particular product by a rival, contributed to the effect, I think you would take this, although it is not specific data, as meaning it would have at least nn impact on the drug company's revenues. Senator NELSON. That is true, but you are using a 17-year period to test the risk of your company. These companies are all in the same 1 See Appendix III beginning at p. 2129. infra. PAGENO="0121" COMPETITIVE PROBLEMS IN THE DRUG INDUSTRY 1671 business. Doe.s this knock them out of bus;iness? Does this knock them way down? Every company that is any competition has this risk. I was just looking at the reports on the manufacturers of sugar pops and Wheaties and corn flakes, and so forth. Almost every month, General Mills or Kellogg or somebody else is coming out with a new cereal and somebody's frosty flakes are knocked out by somebody's sugar pops. It is going on all the time. When you loOk at the stuff they are putting out now as compared with what they did when we were kids, you can see that they are doing it right along. It is all in the same business and going up and down just like the drug companies. But the company, General Mills, is doing right well; so is Kellogg. Dr. MARKHAM. Let me make it clear that I am not making any presentation here, Senator Nelson, whatsoever, that the drug industry is the only industry that confronts this kind of risk. Later on, I try to show the unique magnitude, or rather the magnitude of this risk and that it does have a certain element of uniqueness when applied to the drug industry, but it certainly is not a risk that the drug industry alone confronts. No one is making that claim here. Senator NELSON. That is what I want; to make clear for the record. Dr. MARKHAM. Nor am I trying to assert here that the drug industry is the only risk industry in the economy. The chart you have just had given to you there will document in my mind fairly conclusively that other industries have high risk as welL The point that has been docu- mented is that, in general, the rate of return and the industry risk tend to be strongly correlated. But that is not only the drug industry. There are 59 points on that chart from which that particular line was made. And there are many other industries that fall on the risk scale as welL Senator NELSON. I was listening to your statement. That is why 1 raised the question about your point that "we have, with reasonable certainty, identified the following as major risks somewhat peculiar in this regard to the pharmaceutical industry." The word "peculiar" itself means unique. It means somewhat rare in my opinion. Dr. MARKHAM. It is somewhat rare in that all these five, plus some others, converge at the same time in this industry. I would not make any presentation to you that the drug industry is the only industry that suffers from product competition. Senator NELSON. Then you will submit evidence showing the im- pact of the introduction of new drugs upon the profits of firms manu- facturing less effective but comparable older drugs. I do not think select handful mean anything. If it is a factor that produces a high risk industry, it must be a factor that is burdening them all the time and has been in the past with considerable frequency. If it is in the range of frequency for a number of other industries, I do not see that it is peculiar to this industry. Dr. MARKHAM. I am arguing that it has peculiar impact, and that is one of the points I am going to try to document, Senator. Again, I can simply repeat that I am not trying to urge on you here that the drug industry is the only industry that suffers from or en- counters severe competition from new products. I agree with you about the cereal industry; occasionally also in new models in automobiles. In response to your request that I submit to you additional informa- PAGENO="0122" 1672 COMPETITIVE PROBLEMS IN THE DRUG INDUSTRY tion or cases where the introduction of new products has significantly affected the profits or the revenues of competing firms, I have to defer to Mr. Conrad, because this would be done at Arthur D. Little's. I am just a consultant there. I do not frame the projects, nor do I execute them myself. Perhaps he would like to answer that now. Mr. CoxRAi. One point. \Ve, in constructing this analysis that Pro- fessor Markham is presenting now, had the belief that the drug in- dustry is subjected to the risk whether or not one can list over the last 15 years, 20 or 30 or three or four instances where this type of risk took place. The fact that logically, it can take place, that the industry is set up, the medical profession is set up iii such a way that it is pre- pared to replace a product virtually overnight if it finds a superior agent to treat a particular illness, the fact that this characteristic exists in the pharmaceutical industry we see as an element of its spe- cial riskiness. There are other industries that have product obsoles- cence, but they do not have it with the peculiar character of product obsolescence that the drug industry has. Now-, we can cite examples. We have talked to some of the companies in starting this study and we will be able to cite some examples where product obsolescence has hit the company and had some impact on its earnings. A list of 30 or 40 may not be possible. We do not know yet just how many companies can respond. But the point is that, regardless of how many, the risk from the economist's point of view is there and it is real and the companies are also facing it. Senator NELSON. I suppose there is risk there, but it would seem to me that if it is a. risk that is special to this industry, you ought to pro- cluce evidence of its frequency and then also show the impact upon the particular company within the industry in each case. If it is a tem- porary impact, that is one thing. If it is a permanent impact, that is another thing. But the fact is that the leading manufacturers in this industry have been the leading ones for a long, long time. They have identification in the marketplace, they have identification with the physician. They are producing more than one product. They are all doing research and so they may have one product knocked out by an- other firm. It would seem to me that you would have to show some impact upon a number of companies as a consequence of this type of situation; otherwise you could not conclude that it is a very serious factor in risk, it would seem to me. Dr. MARKHAM. May I just cite one example, Senator Nelson, to il- lustrate at least the point I am trying to make? I did use diuretics as an example. If you were to take a look at the market share of product No. 1 on the chart on page 42 in this docu- ment that was just handed you, Table 11-A, Diuretics, if you take a look at 1957 it had 53.5 percent of the market. In 1958, 1 year later, it had dropped to 17 percent and by the next year, had dropped to 7.8. Go to the bottom of the numbers in that table. Product No. 5 in 1957 w-as eight-tenths of 1 percent of the market.. By 1958, it had 72.1 per- cent of the market. I only submit that in spite of the fact that other industries encounter product competition, this is an unusual shift in market share in that product line. I would find it very difficult, and I know something PAGENO="0123" COMPETITIVE PROBLEMS IN THE DRUG INDUSTRY 1673 about markets in manufacturing, I would find it difficult to find other examples that come close in these magnitudes. Senator NELSON. What were the names of the two companies? Dr. MARKHAM. I do not know, I suspect No. 5 was Merck. It would have to be. These are products and that is, I suspect, Diuril. But I have no idea. of what the others are. Senator NELSON. You say Merck is the one that had the 35 percent in 1957? Dr. MARKHAM. No, sir; it is one that had practically nothing in 1957 and 27 percent in 1958. The others, however, had their shares of the market decline very perceptively. Senator NELSON. If you will give me the name of the company that had 35 percent rate of return in 1957~ and dropped to 16 percent in 1958. Does that column for the company run vertically? Dr. MARKHAM. No, sir; it goes across the page. It dropped to 16 in 158 and to 7.8 in 1959, 6.6 in 1960, and by 1965, that same product had dropped to 3.7 percent. Senator NELSON. Now, then, it is company No. 5 here that went from ~o percent to 72 percent of the market in that product, is that it? Dr. MARKHAM. Yes, product No. 5. Senator NELSON. Now, could you give us the names of those two companies? Merck is No. 1, is that it? Dr. MARKHAM. No, sir; I am virtually certain that product No. 5 is Merck's Diuril. I do not know, but we can, I assume, supply data. Again, I have to defer to Arthur D. Little, because I do not know the particular conditions or safeguards that surround these data. But I assume they can be gotten. Se.nator NELSON. Do you have the names of the companies? Mr. CONRAD. Not with me. Senator NELSON. I must say that does not help very much. But we have Merck Co. here in 1957-58. They went from 8/io percent to 72 percent of the market in this product. The earning assets column shows that they went from 13.79 in 1958, 14.060 in 1959, and down to 12.393 in 1960. So the fluctuation based upon the one who got 72 percent of the market was very tiny. If it were the same on the company that lost the market, I do not call that much of a risk factor, either, do you? Dr. MARKHAM. Yes; I do consider it a significant factor. First of all, I think it ought `to be clearly understood that, take any group of firms, particularly looking tOward the top, you are talking about firms who have somehow met risks, somehow or another you would expect that, by definition, the leading firms over a long period of time will be leading firms in almost any given year. That does not in any sense indicate that the risk is not there. Some firms undertake satisfactorily and successfully risk. There also are some winners. There are a few winners at Monte Carlo who win consistently. But that does not alter the probabhistic outcome of the gambling game. Senator NELSON. Well, maybe so far as your mathematics and eco- nomic theory are concerned, it does not, but the cold, hard fact of the matter is that you can look at the earnings statement of General Motors and it will fluctuate 10, 15 times that much in a single year without any additional competition, holding the same share of the market. So now you say that a company loses a substantial share, a major share, PAGENO="0124" 1674 COMPETITIVE PROBLEMS IN THE DRUG INDUSTRY of the market on one product and there is a fluctuation of a half, a third of a percent and that, therefore, this is a. risk factor. I say to you I have looked at General Motors over the years and they will fluctuate 10 I)ercent or more without any change in the marketplace at all. That is true all the time. You are saying here that obsolescence of a drug constitutes a risk factor. I am just saying that you have areas of industry where there is no obsolescence at all, a.nd just the ordinary play of the marketplace causes fluctuations. Dr. MARKHAM. You are only confirming the theory that has been given you, Senator. The automobile industry ra.nks very high. It hap- pens that they have a different kind of risk. Given a collapse in con- sumer expenditures for durable goods, especially for automobiles, they encounter risk, too. No one denies the automobile business is not highly risky. You are quite right in singling that one out. General Motors is an excellent case of a company that encounters risks, but different from these. You would be hard pre.ssed-at least I would-t.o find a single General Motors model that went from zero on to 70 percent. in even a model in a wide price range in the market, or to find another one that had been reduced from 35 to 11. What you are saying is that the automobile industry has certain kinds of risks; they are a cyclical nature. And indeed they do. And I believe they show up in the three or four top industries in terms of risk and the automobile industry is about the same as the drug industry. Senator NELSON. 1ou say you would be hard put t.o suggest a com- parable instance in the auto industry. I might point out that Corvair sales in 1 year-if I am not mistaken-fell about 50 percent. I recall that the drop was very, very substantial. I think around 50 percent.. So that is a peculiarity of the circumstances involving a.n automobile. Dr. MARKHAM. I am sorry, sir, I did not hear the percentage you were using. Senator NELSON. I say what sticks in my memory is 50 percent. I may be wrong. I will check it and correct the record if necessary. Does Mr. Cutler know? Mr. CUTLER. YOU are asking about the drop of Corvair's share. of the market? Senator NELSON. I do not know about the share, because I do not irnow who they compete with. Mr. CUTLER. It is true that the Corvair's sales dropped over a year by a factor of about 50 percent, but all of Corvair's sales could not have amonnted to 1 percent of 2 percent. or 5 percent. of the sales of General Motors. Sena.tor NELSON. We have not in the record here what percentage of the sales were represented by the dramatic drug instance Dr. Markham has suggested Mr. CUTLER. Dr. Markham is about to give you some data or some theory he has on this subject. Senator NELSON. Before we get. to the theory, I would like to ask a question on the point we were talking about a while back. If it is a risk wh1ch ultimately affected a company~s earnings, and its general sit- uat.ion as a. competing company, not. for 6 months or a year, but for a 16-year period, that would be much more. impressive to me than simply PAGENO="0125" COMPETITIVE PROBLEMS IN THE DRUG INDUSTRY 1675 to say there are lots of products coming on the market in the drug in- dustry. You are bound to have that. But it is `the actual impact that makes it a risky business. You cite a case in which one company picked up almost all of the business but their earnings assets column barely changed. I would like to have the name of the company that lost the business to see what that earnings assets column looked like. Dr. MARKHAM. I think we can furnish you some information on that, but we are not prepared to give you a detailed list of that today. Senator NELSON. All right, if you would submit them for the recoid. Dr. MARKHAM. Fine.' Senator NELSON. We will put it in the record at this point along with table ha on page 42 of the document entitled "Trends in Market Shares for Ethical Pharmaceutical Products." (Table ha referred to follows:) TABLE 11-A--DIURETICS [Percent share of market based on numbers of prescriptions written, 1956-651 Product 1956 1957 1958 1959 1960 1961 1962 1963 1964 1965 1 52.6 53.5 17.0 7.8 6.6 5.6 4.7 5.1 5.0 3.7 2 13.3 12.9 3.3 1.4 .9 .4 .2 .2 .4 .3 3 6.3 2.4 .2 .1 4 5.6 3.1 .6 .3 .1 .1 .1 .1 5 . 8 72. 1 61. 8 42. 3 33. 3 32. 7 23. 7 23. 1 25. 9 6 14.3 20.6 19.6 17.3 18.8 14.4 20.0 7 7.8 11.3 7.8 6.9 4.5 3.3 4.9 8 5.4 5.0 5.4 6.3 7.1 4.1 9 1.0 3.0 4.9 5.7 6.6 5.9 10 .2 2.2 3.2 5.7 8.7 3.3 11 , .1 3.5 4.1 5.0 5.1 4.7 12 .4 6.1 4.9 4.4 4.9 13 .2 2.6 Cumulative market share of above products 77. 8 72. 7 93. 2 93. 5 88. 5 80. 9 85. 6 79. 9 78. 4 80. 3 All others 22. 2 27. 3 6. 8 6. 5 11. 5 19. 1 14. 4 20. 1 21. 6 19. 7 Category total 100. 0 100. 0 100. 0 100. 0 100. 0 100. 0 100. 0 100. 0 100. 0 100. 0 Senator NELSON. All right. I `did not have any more questions on the first item that you listed. You were going on to No. 2. Dr. MARKHAM. Yes, No. 2. And we are enumerating the various risks. Back again to the heading in the statement, the major risks, somewhat peculiar in this regard to the pharmaceutical industry. 2. The discovery of unanticipated side effects of a drug which lead to an immediate limiting of the physical `indications for which it may be prescribed. Again, an example of this might be Chloromycetin, which was with- drawn because it was found to have certain-it gave rise to some blood disorders. Senator NELSON. You are talking about a situation back in 1950-something? Dr. MARKHAM. I am sorry, I used the~ wrong word. It was limited in its use and in a sense, withdrawn by prescribing physicians from some uses because it was found to generate or be conducive to certain blood disorders. 1 See Appendix III beginning at p. 2129, infra. PAGENO="0126" 1676 COMPETITIVE PROBLEMS IN THE DRUG INDUSTRY Mr. Go1i~oN. But what happened to it, Dr. Markham? Dr. MARKHAM. Well, now we can be a little bit more specific in the risk and return of American industry you have. That was the volume we were discussing before lunch. I think I am correct on this. If you turn to page 60 of the Risk and Return in American Industry, the disorder that that drug created was discovered in, I believe, 1952. If you will take a look at Parke, Davis & Co., the manufacturers of the drug- Senator NELSON. Just a minute. We do not have the right one before us. Dr. MARKHAM. Its profits were reduced almost by 50 percent. Now, they went from a 20 percent rate of return in 1952 to 11.4 percent in 1953. Let me make it clear, I have not, Senator, made the precise analysis that can establish the cause and effect on this, but following your own line of reasoning earlier, it looks as though that was a sig- nificant blow to Parke, Davis & Co. in terms of what happened to profits. Mr. GORDON. Dr. Markham, even though, as I understand it, Parke, Davis' profits had fallen, it was still 1.1 percent of net worth after taxes or less than 3 percentage points below the average for all large drug firms for that year. Dr. MARKHAM. Well, that may be; I do not know where you get the 12 percent. My table does not show that. If you go to earnings on assets- Mr. GoRDoN. Well, I went some other place. I can put. that source in the record, too. But even then, when the profits went down, it was still less than 3 percentage points below the average for all large drug firms for that year. Dr. MARKHAM. Oh, sure. What one then has to ask himself is, sup- pose two product lines had been hit. The probability may not. be high, but suppose it had. Mr. GORDON. And incidentally, do you know what happened to Chloromycetin last year? Sales were $45 million in the United States and over $70 million worldwide. Dr. MARKHAM. It may be, but I see by the morning pa.per, that has not solved all of Parke, Davis' financial problems. I do not want to read this into the record. Mr. GoRDoN. We are talking a.bout Chloroinycetin now. Dr. MARKHAM. Yes. All I am saying is that even t.hough it may have come back, disorders do get cleared up, I assume, the point is really a very simple one. A single product that gets hit because of this particular risk drops the profits by 50 percent. I do not think anyone would argue that that is not a. significant risk. Mr. GORDON. Even then they could still be making a high return on their investment. Senator NELSON. So the record will be clear, do you have the total sales of Chloromycetin for the year 1952 versus total sales by the com- pany? I would like to have that projected. We do have the figures on the sales of Chloromycetin, I believe, for last year7 which were $45 million in the United States and $70 million worldwide and it repre- sented a major factor in Parke, Davis' total sales. PAGENO="0127" COMPETITIVE PROBLEMS IN THE DRUG INDUSTRY 1677 Dr. MARKHAM. I think we can supply this. You want the total sales of Parke, Davis & Co. in 1952 and 1953 and also the sales of Chloromycetin? Senator NELSON. Yes, and for each ~year as a proportion of total sales, to see what that factor was. Dr. MARKHAM. Yes. I can only report we will all say we will try to get it.' Senator NELSON. All right. In any event, discussing the risk factOr, that is, a risk, if that were the factor that did cause the decline-I do not know. They were back up within a matter of 3 years to within~ 3 percentage points and then they jumped in 1957, more than 3 percentage points on earnings above the 1952 figure. And, they stayed 4 years in a row above the 1950 figure. All I am saying is that this may be one risk factor. There are all kinds of risk factors and I would just point out that you could prob- ably project a whole series of companies~that did not face any product replacement, obsolesecence, competition, and still experienced the same fluctuation. Dr. MARKHAM. I will concede it did not drive Parke, Davis out of business and that it did subsequently recoup. I do not think, however, that that would lead one to the conclusion that this is not a risk that confronts pharmaceutical manufacturers. I shall continue on, sir. Senator NELSON. Go ahead. Dr. MARKHAM (reading). 3. The discovery that a drug can be mis-. used in ways that create a significant social problem which then leads to limitations on its marketability or possible removal from the market.. I am told that meprobromate might be a good example of that. Mr. GORDON. Has that been removed from the market? Dr. MARKHAM. No, just its sales are subject to greater control. 4. Withdrawal or restriction of a drug by the U.S. Food and Drug Administration until additional evidence of safety or efficiency is produced. I am told that MER-29 might be an example of that. Senator NELSON. Are you going to b:e able to give us a series of examples? Again as with the first question, if it is an important risk factor, it seems to me we ought to see the number of examples of drugs that either had to be limited in their use because of discovered undesir- able side effects not known at the time of the New Drug Application, or for some other reason. But I would think in order to prove it is a risk factor, peculiar or somewhat peculiar, as you put it, to the indus- try, we would need a series of examples to demonstrate that fact. If you have concluded that that is the case, you must have some examples. Dr. MARKHAM. Yes, we are in the process of compiling such a list, Senator.2 Senator NELSON. Let me ask you a question: you referred to unantic- ipated side effects or withdrawal or restriction of the drug on the market. Now, how is that risk, in terms of imposition of expense, more im- portant, for example, than the risk of~ the automobile industry on auto recalls? 1 The information had not been supplied at time of going to press. 2 See Appendix III beginning at p. 2129, infra. PAGENO="0128" 1678 COMPETITIVE PROBLEMS IN THE DRUG INDUSTRY Dr. MARKHAM. Oh, considerably greater. I happen to have followed rather closely the steering defect in Mustang. It does not affect the market at all. What you do do is you incur some, the dealer or the manufacturer-I am not so sure how they share it-you call in to have this particular problem corrected right away. This does not happen in the case of drugs. You cannot just call in the thousand bottles of pills that you sold and say, we will try to correct this. It just does not happen that way. It is not a mechanical defect, in other words, Senator, that you can take the current output in inventory and say, well, we incur some slight expense to eliminate this particular risk. Senator NELSON. Do you know how many cars were recalled last year? Dr. MARKHAM. No, sir; I do not. A substantial number. But I would say that is not the same kind of risk at all. I happened to take one back myself. Senator NELSON. I do not have the figure on costs, because I do not know if it is available. But there were over 4 million cars recalled in 1 year's time. I assume-maybe Mr. Cutler would have an answer to that, too-I assume the manufacturer pays the dealer for handling that aspect. Mr. CUTLER. Just from memory on this, I am not sure of the 4 mil- lion figure, but in any event, a figure of that magnitude is the number of cars that had to be inspected to determine whether they might. con- tain the particular defect. The number that contained the particular defect that had to be repaired was far smaller. Senator NELSON. Correct. Mr. CUTLER. That is a part of the warranty cost that every company incurs and factors into its pricing. Compared to the removal from the market of a product such as MER-29, or a decision such as in the case of meprobromate that it will be subjected to drug abuse, the effect of the latter restrictions would be far greater. In any event, we concede that the automobile is a very high risk industry, with almost the identical rate of return as this industry. Senator NELSON. Go ahead. Dr. MARKHAM (reading). 5. The development of quality control problem which necessitates withdrawal of a product from the market until the problem is traced and alleviated. I offer as just an example, Cutter's polio vaccine. This list of risks is only illustrative and we make 110 pretense that it is complete. Moreover, we do not mean to suggest that other indus- tries do not confront risks. The study just presented to you provides abundant evidence that American manufacturing industries confront varying degrees of risks and that these risks are strongly correlated with profitability. Our study of indivithial pharmaceutical manufacturer's experience reveals that these risks are not simply potential but that most have in fact materialized in recent years. We can provide illustrations in suffi- cient numbers to support this point should the committee wish us to cite them. The most important issue to which Mr. Conrad's and my study is directed, however, is not the frequency with which such risks actually materialize, but the magnitude of the loss in profits and revenues tile PAGENO="0129" COMPETITIVE PROBLEMS IN THE DRUG INDUSTRY 1679 typical firm would suffer should one of these potential risks material- ize. There may be several approaches to answering this question, but the one we are pursuing involves the measurement of the extent to which the major pharmaceutical firms' revenues and profits are con- centrated in relatively few product lines. The logic underlying our particular approach is quite simple. If the typical firm sells numerous pharmaceutical products, no one of which accounts for more than a small fraction of total sales, the above risks may not be significantly larger than the usual risks American business firms encounter in their day-to-day operations. The materialization of any one of these risks would simply mean that a relatively small share of a firm's profits and revenues would very likely evaporate-an event that almost any business firm, confronting average risks, might experience, and ex- pect to experience, in a dynamic economy such as ours. On the other hand, if profits and revenues are highly concentrated in two or three product lines the potential loss from the materialization of any one of the enumerated risks might be quite substantial. The potential col- lapse of a single product's sales accounting for a large percentage of a finn's profits is obviously a much more serious risk than the collapse of the sales of a product accounting for a very small fraction of a firm's profits. Such a potential situation becomes even more uncertain and risky in the case of the pharmaceutical firms when account is taken of what they may have in prospect after the possible collapse of a major product. The types of collapse we refer to do not offer hope, in most cases, of subsequent recovery of the product's market position. Senator NELSON. Do you have some examples of cases where there has been such a product collapse and the company has not made a sub- stantial recovery? Dr. MARKHAM. Oh, yes, we have numerous products. Senator NELSON. I mean companies. Dr. MARKHAM. No, we are not saying that a company collapses, although I think perhaps there may be data on that, too. I am saying that a firm's product line, if indeed it does turn out that product lines can evaporate and do evaporate, then the mere fact that a large per- centage of the firm's sales or profits is concentrated in relatively few product lines makes that potential risk much more of a risk than if you produce, say, a hundred products and 1 percent of your profits were just undercut by the introduction of a product. As I have said, that may be a normal risk that any bilsiness firm encounters. Senator NELSON. Well, I certainly appreciate that you have a risk here but the likelihood of it happening and having serious consequences does not seem to be too great. Mr. CUTLER. As an example, you might look at the Cutter Labora- tory figure on page 50 of the risk and return tables. Cutter's problem with its polio vaccine developed seriously in 1955 and now, in these 1965 figures, 10 years later, had not rea~ched the rate of return which it had enjoyed before that particular event occurred. So in 10 years Cutter has not yet fully recovered from the effect of this particular incident relating to its polio vaccine. Senator NELSON. But if you backed up to the year 1952 you would see that in the year 1965, they are approaching four times the 1952 earnings. 81-280-68-pt. 5-9 PAGENO="0130" 1680 COMPETITIVE PROBLEMS IN THE DRUG INDUSTRY Dr. MARKHAM. But you cannot read that tahle without concluding that it has taken Cutter a very long time indeed to recover from that particular `incident and they are not back yet to where they were just before it happened. Senator NELSON. I am not sure that this is the real factor involved. The first year after it happened they went down from 16.8 to 5.5, but the next year they were up to 13.7. Dr. MARKHAM. You are omitting 1 year. They went from plus 16 to minus 17. Senator NELSON. Oh, I see. Then they went up to a plus 13 2 years later. Figures are figures. I do not know what that proves. I do not even know how much they had invested in polio vaccine and what other factors might have `been involved. All I am saying is that sure, it is true that if a company has a lot of products and they each repre- sent only a small percenta.ge of sales and one gets knocked out, that is not going to be a. substantial loss. Then you go on to say if a firm has only one or two products and one gets knocked out, it is catastrophic. You have presei~tecl 29 companies here. They are manufacturing, according to Mr. Cutler's educated guess, 75 to 80 percent of all the perscription drugs in America. So they are the ones who are the `big- gest in the marketplace. Can you show me any of those, anything sig- nificant among several of those 29 that indicates the kind of risks you are talking about that has seriously damaged them? Dr. M~uKIr~r. If you are asking me as a professional economist, I have a difficult time putting myself in the position of president of Cutter in the year 1955. But as an economist standing on the sideline, I have no difficulty whatsoever being impressed with a firm who goes from plus 16 percent rate of return on his equity to minus 17 percent because a product has encountered this risk. It seems to me fairly evi- dent on its face. I do not need to look much beyond that. Senator NELSON. Let me ask you the question, who are you im- pressed as an economist with what happened between 1950 and 2 years later, 1952, when they went from 19 percent to 3.2 and the drop did not involve polio vaccine at all? Dr. MARKHAM. That may be very true. I have not examined to see what caused that drop in rate of return. I am impressed also with that decline in the rate of return. Senator NELSON. How do you know that the decline to minus 16.55 had anything to do with polio vaccine if you cannot explain to me what caused them to go from 9.5 to whatever in a 2-year period? Dr. MARKHAM. I made this statement earlier, Senator, I make it again: I have not traced through the sales of any one of these products into the revenue accounts of these firms. The timing here is so close with the problem of the polio vaccine, I think the reasonable inference is that that. was one of the risks that the firm encountered in the year 1955. I have also made it clear I have not tried to catalog every single risk that Cutter or any other pharmaceutical firm might encounter. it is perfectly possible that a confluence of all five, seven, or 10 major risks tended to drive its rate of return from 20 down to 3 percent over the previous time that you have indicated. I have not studied that. Senat.or NELSON. So as a scholar, maybe you would recognize that the same fa.ctors were at work in 1955, when t.hey dropped to minus PAGENO="0131" COMPETITIVE PROBLEMS IN THE DRTJG INDUSTRY 1681 16, as were at work from 1950 to 1952, when they dropped a total of 15 percent. And maybe it did not have anything to do with product replacement or obsolescence at all. Dr. MARKHAM. It may possibly be, and I am making no assertion that it is. .Just to be again understood, ,I only say I am impressed with the magnitude of the change, the timing with the polio vaccine prob- lem is anmistakable. That is as far as I would care to go. Senator NELSON. Well, you have given us financial data about 29 companies, which we all have studied very carefully. The drug indus- try ranks as one of the highest in earnings of all industry. Can you name for the committee among the 29, then, for which you have shown dramatic losses, one that suffered any kind of a detrimental effect as a consequence of drug obsolescence? T)r. MARKI-IAM. I have not gone over these tables, Senator Nelson~ Senator NELSON. Oh, all right. Mr. CONRAD. We have not gone over these tables in the detail, as you have suggested, that would be required to say with firmness what happened in the Cutter situation. We have not talked to the financial people and the marketing people about the specific impacts on their earnings from any of the drug changes that Professor Markham has been suggesting. But if we can just quickly read through the list and see where there were a number of instances of very significant change in profit level for these 29 companies. Senator NELSON. But if they substantially restored their position, do you call that a risky company? Mr. CONRAD. As long as the company is subject to significant fluctu- ation in its profit levels over time, the fact that it restored profit levels does not change the fact that it was a risky situation w~hen it oc- curred. It was subject to the risk, the risk took place, the profitability dropped, the profitability then went back up. The `company, through diligence and through research and development and good market- ing, was able to return itself to the better position. But it is still sub- ject, even after returning to this better position, to the same set of risks. Senator NELSON. I just point out tli~tt even Cutter, which you recite as the sole example thus far, has met the industry average 7 out of the last 9 years, or even shown profits above it. I)r. MARKHAM. No; I think you are possibly reading the wrong column. The colunm that we used in the study and that we used when we set the 10.75 percent all-industry ~verage this morning, that is re- turn on total invested capital, the second column, earnings plus interest over book capitalization. Senator NELSON. You used column No. 2? Dr. MA1uicIJAM. Column No. 2 is the column that the study uses and the ones we referred to this morning when we calculated an all-indus- try average return over the 16-year period of 10.75 percent. Using that column, Cutter has been below that level in all except one year~ Senator NELSON. As a matter of fact, that is Cutter's position in the marketplace. They have. only been above it t wiee in the last 1G years. So you have a company that does not have an earnings record that is at the average anyway. Right? They are low. Dr. MARKHAM. And for 3 of the years prior to the polio thing, they were 9.9, 9.2, and 2.9. PAGENO="0132" 1682 COMPETITIVE PROBLEMS IN THE DRUG INDUSTRY So I do not think that proves an awful lot in terms of this big risk you are talking about. Mr. CUTLER. Sir, if we can get this information from Cutter and if we can demonstrate to you that Cutter's decline in rate of return was primarily attributable to the incident that occurred with its polio vaccine and that it has not yet. recovered, would you agree that that is evidence of a substantial risk of t.he type that. Professor Markham is talking about? Senator NELSON. I do not know how well they manage the com- pany or anything else. Mr. CUTLER. We will submit that.' Senator NELSON. I~t may very well be a. factor. All I am saying is that it does not impress me, quite frankly, very much. You have the whole manufacturing mdustry and all the best economists in the country available and we are quibbling a.bout one case which may or may not support your thesis. All I am saying is that you are proving tha~t these factors, such as obsolescence in the marketplace, constitute ~a significant risk factor beyond what normal industry would face. Then let's take a look at the major companies and see how many of them went into a decline and have not recovered a reasonable spot in ~the marketplace because of obsolescence of a drug. That is all. You ~would not go into court with less than that. You cannot go in with one example to prove a case or use the exceptions to prove a rule. Mr. CUTLER. It simply depends on the court, Mr. Chairman. The question is will the judges of the court be economists or who will they be? Senator NELSON. I think any judge, economist or not, is going to say, give me a pattern of evidence that proves the case. I am asking for a pattern; you have given me one case. And you do not have available before you the facts that prove whether or not that is a case that proves your thesis. You said you will find the information. Fine. I am just saying I do not want the record to stand with these assertions that product ob- solescence is a significant risk when you do not have the evidence to prove a single case. Dr. MARKHAM. No one has said, Senator-at least I have not. I think I ha.ve made it quite clear from the beginning that we have not over time traced these particular events through the revenue accounts of the companies that a.re concerned. I thought some time. ago, we did promise to give them to you, to the extent that we could get them to put them together. The focus of our study, I just finished saying, was on a different aspect of this. We have not tried to catalog the total number of times this occurred or any of these risks occurred. We have, however, assembled some da.ta on the extent to which the firm is vulnerable by Virtue of the fact that its sales a.nd its estimated profitability seemed to be fairly highly concentrated in a few, at this point in time, successful products. We are making the very simple thesis that when you have this kind of all your eggs in one basket, or two baskets, so to speak, that firm is more vulnerthle to these risks, however frequent or infrequently they may occur, than would be the case if a firm went on all the time with its sales dispersed over eight or so or 100 products, and fairly evenly. 1 See Appendix III beginning at p. 2129, infra. PAGENO="0133" COMPETITIVE PROBLEMS IN THE DRUG INDUSTRY 1683 I have not come in here to attempt to prove to you that the polio~ vaccine problem of Cutter did indeed reduce its rate of return from plus 16 to minus 16 percent in the course of 1 year. What I did do was do what all economists standing on the outside have to do. You~ see the temporal coincidence of the two events and you at least advance it as a logical and plausible connection~ between the two sets of facts. And that is as far as I would care to take it. In fact, as far as I can take it until I have access to all of the intimate financial records ot Cutter. But I was responding to the comment you made that I was at-~ tempting to prove the general theory with one case. I am not trying to do that at all. Mr. PLOTKIN. Sir, in this morning's testimony we purposely, did not specify the causes of risks in individual industries. We specified a measure which could go over many industries and evaluate the risk. Those results are on the chart. As your example very well points up, there are sources of risk in the automobile industry different but whose causes may be the same in a sense as the causes of risk in the drug industry. This morning an analysis, which was not particularized to the automobile or to the drug industry, shows your intimate feel- ing to be quite true, that the two industries are operating at about the same level of risk and profitability. This afternoon, we are saying, well, there is the drug industry, fairly high in both risk and return. We now ask ourselves are there possible reasons, without saying these are the only reasons, are there possible explanations for why the drug industry has risks as high as the automobile industry's? I think Professor Markham, taking the reasonable man's point of view of what could cause risks, is now enumerating these possibilities. So I submit these are two legs of the same problem. Both of them are being~ addressed in our testimony. Senator NELSON. Well, all I would say in response to that is, one, I named the auto industry off the top of my head. This afternoon, I mentioned the highway construction industry, although I have no statistical basis at all, just the empirical evidence that I have looked at for many, many years, in my State. I think you will find that there are more people in and out of the highway construction business than even are in this business. Nobody goes into the auto production busi- ness. They are all going out. But you also find that those who have the capital and know how are very stable. They have a very long record. We have them in our State for 50 years, very distinguished. The reason I am raising these questions is that you tell me you do not have available any specific statistics to prove the assertions made here. The assertion on page 1 of Dr. Markham's statement is, "We have with reasonable certainty identified the following as major risks, some- what peculiar in this regard to the pharmaceutical industry." Then when I ask for the specific examples to prove the assertions which I question, I am told that you do not have them compiled yet. As just an ordinary country lawyer, when somebody makes an as- sertion, I say OK, you have drawn a conclusion. Then I ask, on what basis did you reach that specific conclusion? That is all I have been saying here. PAGENO="0134" 1684 COMPETITIVE PROBLEMS IN THE DRUG INDUSTRY Mr. CONRAD. The assertions we make are that we have analyzed the statements of the executives in the drug industry and we have analyzed the drug industry ourselves in terms of the types of risks that are inherent and operating in the industry. We are not claiming and we did not come in to say that we are proving through examples of past evidence that these risks have taken place. We are saying that the combination of these risks in prospect and these risks that we have enumerated, these five risks are present, I did not think anyone who has studied the drug industry in depth at all would say that they were not there in the drug industry. Those are reasonable examples of risks that drug firms face. The second point is that we are trying to link that group of risks with a high concentration of profitability in the drug firms. If there is a high concentration of profitability in drug firms among a few product lines, then anyone of those risks could very significantly dis- rupt that company's earnings for any length of time from 1 year up to as long as it takes to introduce a new product into a line to replace the revenues and profits represented by the product that was lost. We in no way and nowhere in the statement, I think, can be inter- preted as coming in here to prove these risks, on the basis of their past existence. We have not attempted to do that, we have not done it in our study to date. We further do not think it necessarily germane. Senator NELSON. Well, OK. I guess we have covered that. Dr. MARKHAM. Shall I proceed, sir ? Senator NELSON. Yes. Dr. MARKHAM. These risks, referring again to five or possibly others, are not of the cyclical type where demand for a product dips only to recover again when conditions affecting its position in the economy improve. Most manufacturers whose profits are reduced during a recession can in general look forward to a restoration of profitability when the economy subsequently recovers. However, the pharmaceutical firm generally must find a replacement for its lost revenue and profits in the form of a successful new product. To the extent that the research, development, approval, and introduction of a new major product requires a prolonged period of years, the firm may suffer long periods of significantly reduced profits. In the study we are making at Arthur P. Little, Inc., we have sought through requests to a representative sample of major pharmaceutical firms to obtain the percent of total net profits and total gross profits accounted for by each of their five most important product lines in the 3 years 1961, 1964, and 1966. To facilitate this transfer of infor- mation the returns are being coded in such a way that neither of us can identify the responding companies. We are prepared to discuss with the committee our interpretation of what we believe the returns received to date show in terms of the mag- nitude of the risks encountered by the firms who have submitted the returns. I would like to just make an addition to the written statement and say, as you can pretty well appreciate, that firms who do not keep their books, or at least do not keep profits, gross profits, or net profits, in terms of specific products or product lines, are very reluctant to step out to make meaningful statements of just the extent to which PAGENO="0135" COMPETITIVE PROBLEMS IN THE DRUG INDUSTRY 1685 profits might be concentrated in a certain product line. However, on the basis of the returns that we have, and bearing in mind that these are estimates, pretty largely made by adjusting the impact of, say, the withdrawal of a product on gross revenues or on gross margins, it does turn out that in the limited number of returns we have already received, it is not unusual to find that 50 percent of the estimated profitability of a firm is accounted for by its leading product in any one of these 3 years. And it is equally not unusual to find that the top two leading products often account for as much as, say, 65 to 70 per- cent of the estimated profit of the firm. This is only an arithmetic way of saying that sales are not distributed Over a wide range of products so that any one of them happening to encounter any of these risks, the revenue impact would be slight. The fact is that they are concen- trated and when a leading product, if and when it does encounter one of these risks, the firm is at least vulnerable to a very significant shrink- age in its profit or revenue, and this, therefore, seems to us to add up to some reinforcement of the possible quantitative impact of these potential risks. Senator NELSON. Looking at your comments on page 2, we discussed at some length the five items listed as peculiar, or somewhat peculiar, as you put it, to the drug industry. Then, in the middle of the page, you state: The study just presented to you provides abundant evidence that American manufacturing industries confront varying degrees of risks and that these risks are strongly correlated with profitability. Our study of individual pharmaceutical manufacturer's experience reveals that these risks are not simply potential but that most have in fact materialized in recent years. We can provide illustrations in sufficient numbers to support this point should the Committee wish us to cite them. That is what I was asking you a while back. Can you provide these instances or are these the ones that have not materialized? Dr. MARKHAM. Yes, sir; we are compiling lists of some of these. I gave you some examples-or at least one example-of each of the five as we went through. I am certain that we can provide you with, in case you are interested in them, a much more complete list than this. Senator NELSON. Well, that is entirely up to you. I would say quite frankly, with what you have presented, the record is not convincing. The record does not, in my judgment, as an ordinary lawyer, contain specific items sufficient to support the five assertions you make. It just does not. Now, you do not even have to be a lawyer to conclude that. If you want to prove your case, I think you have to come in with a pattern such that you use to prove any situation. You would have to come in with evidence that proves the case that these five characteristics are, as you put it, somewhat peculiar to the industry and that they have had adverse effects that makes this industry a risk industry in order to develop a substantial case. Otherwise, quite frankly, I do not believe it and I do not think anybody else reading the record will. You say you can provide illustrations in a sufficient number to prove the point should the committee wish you to cite them. Then you tell us, if I understand you correctly, that you really do not have them yet. The drug industry has argued those five points for years, so anybody could have discovered them. But the drug industry PAGENO="0136" 1686 COMPETITIVE PROBLEMS IN THE DRIJG INDUSTRY has not yet supplied this committee with evidence and these witnesses have not supplied the committee with such information either. I am not trying to be querulous about it. I am just trying to say honestly to you that the record you have made today does not prove these assertions. Dr. MARKHAM. I am not sure what you are asking me to prove, Senator. Senator NELsON. I am asking you to prove the five conclusions you have drawn. Produce the evidence to support the conclusion. I can give you examples in all kinds of industries, off the top of my head, that would prove them to be high risk. For example, just look across this country. Every single distributor of cigarette machines in America is knocked galley west by the 101-millimeter cigarette, because none of those packages fits the machines. I know a poor guy back in my State who put all of his savings into a bunch of machines. They are brand new, a.nd one change in the length of the cigarette knocks out all the machines. There is no market for them, and he is bankrupt. That is a very high risk business. The whims of the cigarette makers to change the length of cigarettes by 1 millimeter or 2 erases all the business. All I am saying is that a single instance like this does not prove that this is a risky business. But you draw five conclusions. I would like to have a number of specific examples proving each of these conclusions. Not one example to prove each conclusion. I can do that for any industry you want to name. Dr. MARKHAM. Perhaps you can, sir. I would not want to go into that. The only thing I was trying to prove is their existence, not how frequently they occur. Perhaps I can use a specific example. If I were to take you to the State of Michigan and show you an apple tree and it was indeed bearing apples, I have proven to you that apples will grow in the State of Michigan. However, I have said nothing about the importance of apples in the Michigan economy. I think what you are asking me to do is to try as best I can to tabulate and get together all of the instances that we can get from the history of the drug industry, certainly the recent history of the drug industry, to give you a better idea of the frequency with which this kind of risk has occurred. I wish to make it clear that my only point was, and I have proven the case because I have shown you an example of the materialization of everyone of the five, just as I would have proven to you that apples can grow in Michigan if I could show you a live apple tree growing apples. What you are asking me for is a different level of proof and I am saying I will try as best I can to amass that amount of information. But I do not think it is quite fair to say that I have come in and tried to prove a big case with one example. I only tried to prove one marrow issue; that is that these risks not only are indicated in the history of the industry, but that in fact, a rather striking case can be found to illustrate all fi~re of them. I have also attempted to indicate that I am not making any pretense that this is a complete list of the risks that drug firms have. But I do not think anyone can deny the fact that if Cutter's vaccine has to be PAGENO="0137" COMPETITIVE PROBLEMS IN THE DRUG INDUSTRY 1687 withdrawn, this does document that it ôan happen. It does, because it did happen. What you are asking for, I think, is something else. You would like to know, if I interpret you correctly, you would like to have me give you now some evidence that this happens with some degree of fre- quency and not just that I can prove it happens. But you want me to prove the extent to which it happens. And that is a different issue. That is the reason that we do not, or one of the reasons we do not have the whole list compiled. If you would be interested in as complete a list as we can get, and I think we were prepared to gather as complete alistaswecan. Senator NELSON. Well, I think the issue is that the industry asserts that there is something special, very spdcial about the risk in the drug industry, that does not exist for very many other industries. All I am saying is that if you are claiming a special case, then we need I would think, some substantial evidence to prove the special case. And we have not received it. Dr. MARKHAM. Let me just address myself to that too. We have not said, Senator, that the drug industry is the only industry that encoun- ters risks. What we have tried to do is to fit it in a pattern along with the manufacturing industry generally. That was the presentation there. Now, the purpose of this study is to try to specify the particular kinds of risks that drugs encounter. We say that they are somewhat unique to drugs. I believe we are right. But whether they are entirely unique, each one of them, that you could not find some counterpart for example if the rest of manufacturing enterprise, would not really do very much damage to either the major risk study or to this one that is attempting to show how risk comes to rest possibly on particular product lines. There are many industries up there, as you can see, that have varying degrees of risk. I am sure if you were to ask some people knowledgeable in the auto- mobile industry to catalog their particular risks, they would come up with some risks and it may very well be that one or two of them would slightly resemble or perhaps closely resemble one or two of the risks I have enumerated here. But I am arguing that this bundle of risks is fairly unique. I believe you have made the statement that you could offer me any number of industries that encounter the five. If yOu can, then my statement falls. Senator NELSON. I am sorry, I did not intend to say the same kind of risks, but industries that have risks that are just as significant to the industry involved as these risks are to the drug industry. That is what I am really getting at. I do not doubt the scholarliness of the presentation by any means. I know your credentials are distinguished. There is no doubt about that. Perhaps everything you have said could stand unchallenged. Bu.t if you can find a group of risks in other industries that are as substantial, even if they are of a different kind, the claims of special risk in the drug industry does not mean so much. This is the only industry I know of that spends as much time, with as much money, as it does propagandizing the Nation about how great a risk it faces. I can take you to any medical society meeting and without planting a word with them, just say, "What do you think about the profits of PAGENO="0138" 1688 COMPETITIVE PROBLEMS IN THE DRUG INDUSTRY the drug industry?" And they will say, "Well, it has to have high profits because the drug companies do a lot of research and it is a very risky business." They sound like automatons. I had the very delightful experience, from my standpoint, of attend- ing a reception at which there were about four doctors. They tackled me on these hearings. I said "I will tell you what I will do. Just let one of the doctors step aside and I will tell him what you are going to tell me about the drug industry and we will come back together." And I did anticipate his words in some detail; he was outraged. I said, "You have `been receiving this education, if that is what it is, for some years, you `believe it all." This is one industry that insists upon this argument of risks. Now, I know a couple of wealthy fruit growers, millionaires. Every once in a while, they know a poor season is coming; they will have a frost and be wiped out for the summer. But they are still millionaires and they have been in the business for 50 years. Well, that is a high risk industry and I can give you other examples. Agriculture-I know of people in the beef-raising business. They are millionaires. Once in a while a hoof and mouth disease or some- thing else, although not recently in this country, will cause them a substantial loss. Pulpwood-I know of all kinds of pulp people in my State, probably 3,000 of them, who do very well for a while. All of a sudden, there comes a surplus of pulp and they are in bad trouble. All I am saying is that if you claim that a sharp distinction of great significance to the drug industry is a special risk, that has an effect on the industry different from other kinds of risks on other industries, then fine. But to claim that you have some risk nobody else has because nobody else manufactures drugs does not impress me. Mr. CUTLER. No one has claimed that, Senator. Dr. MARKHAM. I thought the pattern presented this morning sup- ports, Senator, that we are taking. 59 industries, not just one. No one has said the drug industry is the only risky industry. In fact, the chart documents fairly carefully that at least three, possibly four-my vision is not too good from here-industries ra.nk higher in risk on the chart you have been given. And there are quite a number that just lean over a little bit toward the left. That is all this is saying. In fact, all it is saying is that the drug industry's rate of return seems to fall in the ballpark it should fall within, given its risk, in the way it falls with respect to the rest of the economy. And no one is arguing that it is the only industry. You see, I am not among those doctors. I have not been propagan- dized. I look at the figures and `they speak pretty eloquently to me that it falls within the pattern of the rest of the industries of the American economy. Senator NELSON. Well, OK. Is that the conclusion of these witnesses? Mr. CUTLER. That is the conclusion of this group, Senator. We now have Professor Firestone to talk to you about the downward trend of the drug prices. I would like to add one word, Senator, on this before we close the subject. PAGENO="0139" COMPETITIVE PROBLEMS IN THE DRUG INDUSTRY 1689 You have said that we have not come forward with evidence to sup- port how products get displaced in the market by other products. We have given you that evidence in the study that was handed to you that is called, "Trends in Market Shares for Ethical Pharmaceutical Products." We have also assumed that this committee is fully aware of the economic data that was presented in the Kefauver hearings by PMA, by Professor Markham and by Arthur~ D. Little, in which we gave you detailed product line by line breakdowns showing how each company moved up, down, and sideways within each of the particular pharama- ceutical classes. That appears starting at page 2526 of part 4 of the hearing before the Kefauver committee back in 1961. We have not tried to repeat any of that data. We didn't see any point in trying to prove it again. Mr. GORDON. In "Trends in Market Shares for Ethical Pharmaceu- tical Products," where you have the table on various products, prod- ucts 1, 2, 3, 4, could we have the names of those products? Could we get the data, the basic raw material, upon which these are based? Mr. CUTLER. Well, the data on which they are based comes from th~ proprietary studies made by the Gosselin line people of Pharmaceutical Sales, by product. We are not at liberty to give you the names of the products. Perhaps you can get them from Gosselin, which owns the data. Now, in the 1961 hearings, we did provide you, again in coded form, with the name of the company turning out each of those products. If that is any help to you, I suppose we could do that again. Mr. GORDON. Since you presented this, we would like to get the raw material on which it is based. Mr. CUTLER. Perhaps you would ask Gosselin. We will join you in that request to see if they can give you those names. We can give you a set of coded names to go with the product codes, if that is any help to you, as we did in 1961. That will show you which company was the leader at any given time and how the company's fortune fared. Mr. GORDON. We may want to duplicate this work to check it. Mr. CUTLER. We have told you our source, Mr. Gordon. If you can get them to give you their names, that is fine. Senator NELSON. Let's take a 5-minute recess. I want to talk to my office. (Recess.) Senator NELSON. Come to order, please. Mr. OUTLER. Mr. Chairman, our next pair of witnesses is Dr. Joim Firestone, professor of economics at the City College of New York and Mr. Howard Binkley, vice president for administration and plan- ning of the PMA. Dr. Firestone has a statement describing the pharmaceutical prod- uct price index, which he has compiled over the last many years. He will just summarize the statement very briefly. We will then place it in the record if we may, and Dr. Firestone and Mr. Binkley will an- swer any questions you may have.1 1 The complete prepared statement of Dr. Firestone begins at p. 1702, infra. PAGENO="0140" 1690 COMPETITIVE PROBLEMS IN THE DRUG INDUSTRY Senator NELSON. We are pleased to have you here, Dr. Firestone. Your biographical sketch will be printed at this point. Please proceed. (The biographical sketch follows:) BIOGRAPHICAL SKETCH OF JOHN M. FIRESTONE, PH. D. Education: B.S., College of The City of New York, 1930; M.B.A., College of The City of New York, 1933; and Ph. D., Columbia University, 1955. Present Position: Professor of Economics, College of The City of New York. Teaching experience: Teaching elementary, intermediate and advanced courses in statistics at: College of The City of New York, 1929 to date; Southern Meth- odist University, summers of 1937 and 1938; Albany Graduate Program in Public Administration (sponsored by New York University, Syracuse University and the University of New York State), 1958; and Fashion Institute of Technology, 1954-56. Government experience: U.S. Department of Labor, Bureau of Labor Statistics Economist, New York Regional Office (part-time), 1955-64; Price Economist, Chicago Regional Office, 1951-52; National Housing Agency, Chief, Cost Analysis Section, 1946-47; Patent Royalties Commission, 1945-46; War Department, Chief, Price Index Section, 1943-45; and Office of Price Administration, 1943. Business research experience: Consultant, Pharmaceutical Manufacturers Association, 1964 to date. Consultant, various Litigating Groups of Public Utilities, 1963 to date. Consultant, Anti-Trust Investigation Group of investor owned public utilities, 1961-63. Consultant, Merck, Sharp & Dohme, 1961 to date. Consultant, Prentice-Hall, 1947-54. Consultant, Controllers Institute of America, 1949-51. Research Associate, National Bureau of Economic Research, 1950-51. Consultant, McGraw-Hill Publishing Company, 1947-50. Consultant, World Calendar Association, 1949. Professional and other memberships: American Statistical Association; Ameri- can Economic Association; National Association of Business Economists; Ameri- can Association for the Advancement of Science; Social Science Research Coun- cil; U.S. Chamber of Commerce; and United Cerebral Palsy Association of New York, 1952-58. STATEMENT OF JOHN M. FIRESTONE, PH. D., PROFESSOR OF ECO- NOMICS, THE CITY COLLEGE OF NEW YORK, NEW YORK, N.Y.; ACCOMPANIED BY HOWARD BINKLEY, VICE PRESIDENT FOR ADMINISTRATION AND PLANNING, PHARMACEUTICAL MANU- FACTURERS ASSOCIATION, WASHINGTON, D.C. Dr. FIRESTONE. Mr. Chairman, gentlemen of the committee, my name is John M. Firestone. I am a professor of economics at the Col- lege of the City of New York where I have taught since 1929. Much of my work as a consultant to industry and government in- volves my principal field of interest-price statistics and indexes. In this area I have served as a consultant to the U.S. Bureau of Labor Statistics and have designed and developed price indexes for the War Department, the National Housing Agency and for nongovernment organizations. I am a fellow of the American Statistical Association and of the American Association for the Advancement of Science, and am a mem- ber of the American Economic Association and the National Associ- ation of Business Economists. My testimony here is related to my activity ias a consultant for the past 6 years to the Pharmaceutical Manufacturers Association. I have PAGENO="0141" COMPETITIVE PROBLEMS IN THE DRUG INDUSTRY 1691 been requested to address my remarks to drug prices and particularly to testimony on this subject which has been presented by earlier witnesses. First may I express my appreciation to the chairman and to mem- bers of the committee for the opportunity to appear before you and may I say that I sincerely hope that my testimony will contribute to a better understanding of the factors entering into drug price trends. For the measurement and meaning of changes in price levels is indeed not a simple matter. Price changes cannot be taken at face value, since very often the prices compared are not strictly comparable though at first view they may seem to be. I should like to discuss briefly the ways in which wholesale and retail price indexes are constructed, hOw they differ from each other, what they mean, and as important, what they do not mean in respect to drug prices. I should also like to examine the concept of an average prescription price, and point out the pitfalls in using this as a measure of change in the price of drugs, particularly at the wholesole level. Price indexes are desigiied to do only one thing. They measure the change in prices over time for the universe of commodities which the index is meant to represent. The index~ number does not, for instance, show the effect on the housewife's market basket when she substitutes cake mixes for the crude ingredients or frozen packaged vegetables for a sack of potatoes. It does not tell us whether we are using more ex- pensive or less expensive commodities than we used in the past. It does tell us whether, once introduced into the index, they and the other components of the market basket, have risen or fallen in price from one period to the next. Once the market in which prices are to be measured has been defined, a sample must be taken of the commodities and of the specifications to be priced. For prescription pharmaceuticals for example, a selection must be made of package size, dosage form and strength, of transac- tion size, of terms of sale and of points in time and place at which these prices are to be taken. In constructing a price index, logical weights (measures of the im- portance of the products selected for the index) must be employed. These weights should, as truly as possible, reflect the relative im~ portance of the commodities in the index. Changes in the prices of the most widely and frequently used commodities should have greater in- fluence on the index than changes of price in less important or in- frequently used products. These studies of relative importance Of commodities cannot be "once and for all" studies. Like most else in our society, they become obso- lescent. For, in a dynamic economy such as ours, new products are con- stantly appearing on the market, many older products diminish in importance and often disappear altogether. Also the use of existing products changes continuously as custom and needs decree. Du~ to the high cost usually entailed in collecting data, revisions of indexes are infrequent, and the accurac~ with which the index meas- ures average price changes is impaired. Until 1967, as long as ~ years after the fact, it was necessary to em~ ploy 1958 weights for the BLS wholesale price index. The 1963 weights PAGENO="0142" 1692 CO~ETITIVE PROBLEMS nc THE DRUG ThThUSTRY could not be introduced until this year. Accordingly, when new weights are introduced into the index, they are already 3 to 4 years old, and at the time they are changed, they are nearly 10 years old. The Con- sumer Price Index has undergone even more infrequent weight changes. In general terms, for there are no objective criteria available, changes should be made as frequently and as early as possible in order to more truly reflect current conditions. As a case in point, the British Government kept the commodity list in its price index unchanged from 1913 to 1947. At this latter date, with some difficulty, women's high button shoes, whalebone-stay corsets, and kerosene lamps (for light- ing) were still being priced. The BLS Wholesale Price Index is designed to measure changes in the prices of all conimodities sold in commercial transactions in pri- mary markets of the United States including Alaska and Hawaii. All these commodities, taking account of grade, size and other speci- fications number in the hundreds of thousands and the total number of price transactions for them in a month are in the untold millions. To measure what happens to prices, the Bureau collects actual price quotations on a sample of about 2,300 representative items from man- ufacturers and other producers. Ethical pharmaceutical preparations account for about 0.5 percent tf the total importance of conimodities in the WPI. A relatively larger number of commodities are used to construct these index corn- `ponents of the WPI than for most other product classes. There are ~1 drug and pharmaceutical materials, 55 ethical pharmaceuticals divided into 15 subproduct classes and 24 proprietary pharmaceuticals divided into eight subproduct classes making a total of 110 products. A proportionate share, that is 0.9 percent of the 2,300 products used for the entire WPI would be 21 which probably results in the drug category being more representative than most other categories of like weight. Even before the revision and updating of the pharmaceutical indexes in 1961, 37 products were included. Thus, pharmaceuticals are not glossed over in the WPI. But even this list of products leaves something to be desired in a definitive study of drug prices. Before the BLS had indicated that it was revising its drug index, I had been asked by the industry to construct the index now called the PMA Wholesale Price Index of Ethical Pharmaceuticals. To do this, I went back to 1949, and with a more comprehensive list of drugs con- structed an index which showed price movements since that date. The construction of this index is described in a succeeding portion of this paper. It is pertinent to compare the BLS Drug and Pharmaceutical Index with the PMA index for the periods 1949-60, the period preceding the BLS revision, and 1961-66, the period after the revision. From table I-A and chart I-A, one can see that the BLS index in the earlier period showed practically no change in drug prices while the PMA index declined 7.6 percent. The comparison for the later period (with no change in the technique employed in constructing the PMA index) shows that the movements of the two indexes were very similar, with the BLS index showing an even larger decline than shown by the PMA index. (See table I-B and chart I-B.) PAGENO="0143" COMPETITIVE PROBLEMS IN THE DRUG INDUSTRY 1693 If you look at the period from 1949 to 1960, the BLS index is practically stationary, whereas the PMA index declined almost continuously. From 1961 on, the BLS index showed a decline and the PMA index continued to move downward as it had to that date. There is a com- parison made in the last two columns of the movement since 1961, you can see from this that the BLS index declined to 94.9 whereas the PMA went down only to 95.8. Mr. GORDON. Looking at page 10-A, we see that all of the drug prices fell over the years. There are large drops in the early 1940's and 1950's. Dr. FIRESTONE. Yes. Mr. GORr~oN. Do these drops account for the full decrease which occurred between 1940 and 1949? Dr. FIRESTONE. Not entirely. There are a few years in which the prices did rise. Mr. GORDON. I am talking about the decline. Dr. FIRESTONE. Well, there have been declines in practically every year but the last four. Mr. GoiwoN. They account for most of the declines, do they not? Dr. FIRESTONE. Yes. Mr. GOIu0N. Now, it is my understanding that the late 1940 and early 1950 drops are accounted for by the price drops in penicillin in the early forms and the early forms of steroid hormones which are unpatented or at least freely available and can be supplied and sold by many firms. Is that a fair statement? Dr. FIRESTONE. No, sir, it is not, because a little further on I discuss a patented and nonpatented index. You will find the index on the non- patented products show practically no change, whereas the patented index accounts for practically all of the change in prices. Mr. GORDON. You say the drop in penicillin prices has nothing to do with these drops? Dr. FIRESTONE. I did not say that. I said the drop in penicillin prices was at a time when penicillin w~s still under patent protection. Mr. GORDON. Penicillin was never under patent protection. The Government owned the patent on that, sir. Maybe that is why that index of yours on patented products shows a decline. Dr. FIRESTONE. Yes, but some forms of penicillin did carry some form of patent protection based on developments of the companies that produced these products. Mr. GORDON. But when did they come out? They were marketed fairly recently, were they not? Dr. FIRESTONE. No, this goes back into the 1940's. Mr. GORDON. Penicillin G was not patented. It was freely available. The U.S. Government developed and patented the fermentation proc- ess used in its manufacture. The patents were put in the public domain. Dr. FIRESTONE. The information I have is that these are covered by patents. Mr. GoRDoN. We will come to that. Dr. FIRESTONE. We come to page 11. Before detailing the construction of the PMA Wholesale Price In- dex of Ethical Pharmaceuticals, I should like to state that I have PAGENO="0144" 1694 COMPETITIVE PROBLEMS IN THE DRUG INDUSTRY always had complete freedom in designing and constructing this index. This is the only way I would do it; the Pharmaceutical Manufacturers Association never asked me to do it otherwise. As a result, I alone am responsible for this index in its content and its accuracy. The index measures the changes in prices of ethical pharmaceuticals as reflected in the price quotations for those brand-name products which are prescribed frequently enough to be recorded in national surveys. The sources of the price data used are relevant editions of a trade journal annual, the Drug Topics Red Book. I am going to skip over most of this material, and at the middle of page 12, refer to the weights for this index were derived from confi- dential proprietary data. compiled by Medical Data Services, a market research firm of Darien, Conn., from its prescription panel service and provided to me on a confidential basis. The weight used in constructing the index is the average dollar vol- ume for 2 consecutive years ending with the year of reference. Thus a product is on the market at. least 2 years with substantial dollar vol- nine in prescriptiolls before. it is then included in the index for the second year. The result is that only products which become established in the market can be included in the index. Initial exploratory prices dur- ing the introductory period for the product would not generally affect the index. Indication of the. rapid changes that have taken place in drug usage may be seen from the fact that only 27 of the original 99 products are still included in the index. It should be obvious that fail- ure to recognize and take account of such changes on a current basis when constructing the price index would result in rapid obsolescence of the index. Accordingly, the list of products and their weights are reviewed each yea.r to make sure that no important item is excluded from the index and to provide for the retirement of drugs t.hat have declined to insignificantly low levels. I note that the Commissioner of t.he Bureau of Labor Statistics in recent testimony before this coin- mit.tee stated that the BLS has recently adopted this policy for its drug index. In the 17 years covered by the PMA Index o*f Wholesale Prices of Ethical Pharmaceuticals, there has been a total decline of 14.1 percent. Declines have occurred in 11 of t.he 17 yea.rs, increases in 4 and there have been 2 years with a. change of less than 0.05 percent. Of the 17 therapeutic groups, 10 have shown increases a.nd seven have had de- creases since 1949. However, the declines of more than 50 percent that have occurred in two of the most. important groups, antibiotics and hormones, more than offset increases of much smaller magnitude in less important groups. \~Then the drug produc.ts had been classified by their patent status, patented products showed a decline of 24.8 percent (through 1966) while the nonpatented products had risen by 1.1 percent in this same period. Mr. GORDON. Would you please supply us with the raw data and your calculations. Dr. FIREsTONE. What do you mean by raw data ? Mr. Go~ox. Well, whatever information you had on which you ba.sed this particular statement. PAGENO="0145" COMPETITIVE PROBLEMS IN THE DRUG INDUSTRY 1695 Dr. FIRESTONE. Well, part of it is confidential. Mr. GORDON. In other words, we cannot check it? Dr. FIRESTONE. I would certainly have to get clearance from Medi- cal Data Services and Gosselin on the weighting material. Mr. CUTLER. Do you want the list of the patented and unpatented products? Mr. GoRDoN. `We want all information, including the calculations on which this is based. We wa.nt to be able to check it. Mr. CUTLER. Well, we can provide you with all of the data we used except the weights-well, we can provide you with the weights we used, but the sources of the weights, I take it, are confidential data furnished by Prescription Audit Service, which has restrictions on its ability to disclose the data. If the committee would ask that organization to furnish the data, perhaps they would do so, but we are not at liberty to do so. 1-lowever, the method, and, I believe, the sources, are the same as those used by the BLS. Mr. Gordon, could we satisfy you by supplying all of our raw data in confidence to the BLS and have them satisfy you as to whether they consider it was accurately constructed? Mr. GoRDON. I would like to have it myself. I would like to do it myself so I could know if this was the correct data. Mr. CUTLER. Dr. Firestone will sit down with you and show you the index. If he used data which we have received in confidence from Pre- scription Audit Service, we would suggest that you ask the Prescrip- tion Audit Service if they can give you the raw data. `We have no objection to their giving it to you if they are free themselves to do so. `We are not the ones placing the confidentiality on the data. Senator NELSON. Let me ask you a question. On the last paragraph on page 16, your sentence is that "When drug products were classified by their patent status, patented products showed a decline of 24.8 per- cent-through 1966-while the nonpaténted products had risen by 1.1 percent in this same period." What is the period? Dr. FIRESTONE. 1949 through 1966. Senator NELSON. At what stage did the dramatic declines occur in any particular product? Whe~n the patent expired? Dr. FIRESTONE. No, sir; at that time they were placed in the non- patent status. Once the patent expires. Senator NELSON. So imniediately when a patent expires it is placed for purposes of your study in a nonpatent status? Dr. FIRESTONE. That is right. Senator NELSON. So you are saying that during the course of this period from 1949 to 1966, prices for drugs that were still under patent declined 24.8 percent? Dr. FIRESTONE. Yes, sir. Senator NELSON. `Were any of these drugs ones that were patented but were licensed to other companies to produce? Dr. FIRESTONE. Probably. Senator NELSON. So that might be one fa.ctor? Dr. FIRESTONE. Yes, sir. Senator NELSON. In other words, competition in the marketplace? 81-280-08-pt. 5-10 PAGENO="0146" 1696 COMPETITIVE PROBLEMS IN THE DRUG INDUSTRY Dr. FIRESTONE. That might be the factor; yes, sir. Senator NELSON. Did your statistical analysis demonstrate that there was a pattern of decline in patented drug prices as the expiration date of the patent approached? Dr. FIRESTONE. I have not examined that closely, Senator Nelson. I would be happy to go back and take another look from this aspect at tile index and see whether this is tile case or not. I do not think it is, frankly. Senator NELSON. I know of one case where that happened fairly dramatically, so I wondered if that could be the pattern. Dr. FIRESTONE. The interesting thing is when you have the most com- petition, when the products are not covered by patents, you have rising prices instead of falling prices. Senator NELSON. Where what? Dr. FIRESTONE. Where you have no patent protection and you have the greatest degree of competition, prices do not fall. Senator NELSON. That has been true recently of almost every prod- uct in the marketplace, has it not? Food has no patent and it is going up in price. Automobiles have no patent and they are going up. Clothes have no patent and they are going up. The cost of production is going up. Dr. FIRESTONE. There is some question about the rise in price of automobiles. Senator NELSON. Not in the dollar cost. Dr. FIRESTONE. Senator, one of the factors -that is often overlooked is tile change in the quality of the product. For example, you could hardly compare a 1950 automobile with a 1966 automobile. Let's talk about a Chevrolet. The price of a Chevrolet has gone up. But a car comparable to the 1950 Chevrolet, perhaps Corvair, would be at a comparable price. And it would be a better car. Senator NELSON. I think you used the wrong example. Dr. FIRESTONE. Well, perhaps. Or you might take the case of auto- mobile tires. I remember when I paid $10 for a tire back in the 1920's and I got 10,000 miles out of that tire. Now I buy a tire for $20 and get 40,000 miles. Would you say the price had gone up 100 percent or declined 50 percent? Senator NELSON. You chose another unfortunate case. I happen to have introduced the tire safety bill and we have statistics galore about the quality of low-grade tires onthe.market. - But do your statistics show that when the patent goes off and a number of other firms go into the marketplace, the price of the product drops? Dr. FIRESTONE. It does not. Senator NELSON. It does not drop? Dr. FIRESTONE. Yes, sir. Senator NELSON. Well, I would like specifically to have the list of tile patented drugs that you used, the prices for each year, and then the da.t-e the patent went off and what happened to the price then. Dr. FIRESTONE. I should be happy to give you that. Senator NELSON. If that is the case I wonder if it does not prove something that I have suspected all along and that many people have asserted. PAGENO="0147" COMPETITIVE PROBLEMS IN THE DRUG INDUSTRY 1697 There is no competition in the marketiilace; that is why the price did ~not drop. In other words, once the drug has 17 years of patent protec- tion behind it and the doctors in the country know only that particular drug, they continue to prescribe it despite the fact that in the market- place there may be a comparable drug for a half, a fifth, a 10th, or even a 20th of the price, which indicates something peculiar about the pricing structure. I will give you an example and ask you to comment on it as an economist. I have used this one before. I have here an issue of the Medical Letter, which all the industry and all the distinguished physi- cians in this country accept as a very fine publication. The Medical Letter studied the case of prednisone. The original patent on predni- sone was held by Schering, which called~ its product Meticorten, and as of the time of this publication, June 2, 1967, Meticorten was selling for $17.90 for 100 tablets. Merck was in the marketplace at $2.20 a hundred, Upjohn at $2.25, Wolins Pharmacal at 59 cents, versus $17.90, and several others were selling the drug-some at $1.84, $1.75. The Medical Letter concluded that the great price spread among the tablets pur- chased from different pharmaceutical companies suggests the desira- Ibility of prescribing by generic name, and the Medical Letter urges, at least for persons of limited means, that~ the prescription be filled with low-priced generic tablets. So, do you not have the case here where once the name gets estab- lished, the company can continue to hold its share of the marketplace despite the fact that there are competing drugs available at a fraction of the cost and the same quality? Dr. FIRESTONE. I am not technically qualified to comment on the relative merits of Meticorten versus other forms of prednisone. Senator NELSON. I just gave you that as an example. Dr. FIRESTONE. Yes; but as an economist I cannot answer that for you. Senator NELSON. As an economist, hOw would you answer the fact that here is a medical publication listing 22 drugs they say are of equivalent value. They range in price from 59 cents a hundred to $17.90 a hundred and still a substantial proportion of the market is held by the drug selling for $17.90. Dr. FIRESTONE. Apparently, for some reason, physicians did not :agree with that Medical Letter. Mr. CUTLER. Senator, I cannot help~ but suggest that as the testi- mony before your committee shows, Schering's share in this corti- costeroid market is down considerably and secondly, prednisone hap- pens to be one of the six or seven drugs that Dr. Lloyd Miller of the TJSP has informed you in writing on November 29, is a drug which has been found to have differing therapeutic effect in different product forms, even though all of them meet USP standards. Senator NELSON. Not in the case of 22 companies that are manu- facturing pTednisone according to the Medical Letter of June 2. Mr. CUTLER. That I do not know and I do not believe Dr. Miller's letter gives the source of his figures on prednisone. Senator NELSON. Well, all I am saying is that here are 22 of them and they vary in price from 59 cents to $17.90. PAGENO="0148" 1698 COMPETITIVE PROBLEMS IN THE DRUG INDUSTRY I am just asking your comment as an economist. Mr. Conzen's testi- mony, if I recall correctly, is that Schering had about 88 percent of the retail market for prednisone. Mr. CUTLER. I do not believe that is true, sir. Senator NELSON. That was his testimony. I do not. know whether he is correct or not. He may have been mistaken. Well, go ahead. Dr. FIRESTONE. We were discussing the declines in the index. On page 17, skipping down to the first paragraph, one market analysis available, a market analysis compiled by Lea Associates of Ambler, Pa., shows among other data the distribution of prescriptions for products by age group. One of these age groups is the "60 and over" category, and these data are available back to 1959. An index of wholesale drug prices for this age group was constructed, with 1959 as a base-equal to 100-and the total drug index was compared with this. The results are shown in table IV. An important reason for the difference in the decline of the two indexes is that oral contraceptives, which have shown a marked de- cline, are not included in the "over 60" index. Medical care represents nearly 6 percent of the expenditures of wage earners and clerical workers. Only 0.6 percent of the expenditures are on prescriptions. Professional services of doctors, dentists, optome- trists and nurses,hospital supplies and nursing home services, health insurance and over-the-counter drugs and sundries account for the remainder, and in all, are nearly 10 times as great as expenditures for prescriptions. It is only since 1960 that the Consumer Price Inde.x has included prescription items that by current standards can be considered truly representative. Prior to that yea.r the prescription index was based on three iteins-APC tablets, elixir terpin hydrate and codeine, and buffered penicillin-hardly a representative list of ethical pharma- ceuticals. This index rose 27.8 percent between 1949 and 1960, but I would be reluctant to accept this as an indicator of retail prescription prices. The item list was revised and expanded in 1960 to include seven thera.peutic groups, and just this year three additional groups were added. Each group is represented by only one or two products, with a total of 17 items to represent the 10 groups. Since the revision, the index has declined 12.7 percent. An index cannot tell us whether prices are too high or too low, or whether more or less expensive products are in use today than yes- terday. Senator NELSON. It does not really tell us whether the consumer is really paying more dollars or a larger percentage of his income~ for the product, either, does it? Dr. FIRESTONE. No; as a matterof fact, he is spending less. Senator NELSON. The index does not tell whether he is spending more of his dollar on drugs or on any other product, does it? Dr. FIRESTONE. No; and it does not tell us what his income is, either. Senator NELSON. So it has a very limited significance in that respect? Dr. FIRESTONE. Senator, I would not accept that as a correct state- ment, because whatever other measures we have of prices are not as good as this. PAGENO="0149" COMPETITIVE PROBLEMS IN THE DRUG INDUSTRY 1699 Senator NELSON. Well, that may be, but the fact is that you might be paying twice as much for your prescriptions and perhaps twice as much is being spent by the consumer in this country for drugs as was spent, say years ago. Dr. FIRESTONE. Senator, I will come to that very point very shortly. Senator NELSON. And still the index does not reflect an increase. Dr. FIRESTONE. I will come to that very shortly, Senator. Mr. GORDON. Dr. Firestone, just one second. You say you cannot tell whether prices are too high or too low. Prices which are lower today than they were, say, 10 years ago, may simply be only, say, 85 percent as exorbitant as they were before. If they were exorbitant in the past, they could be slightly less exorbitant as before. Dr. FIRESTONE. Conceivably. Mr. GORDON. So what the prices are, are just as important, too. Dr. FIRESTONE. Comparing prescription drug prices with all others, this is one group that has uniquely declined while practically all other groups in the consumer price and wholesale price indexes have risen. Mr. GORDON. Could that be a indication, by the way, that it is not u competitive industry? Dr. FIRESTONE. I should think not, because there is a wide range of prices for these products, as you know. Mr. CUTLER. You do not mean, Mr. Gordon, that it would please you more if prices had gone up, do you ~ Mr. GORDON. It is peculiar that if everything else goes up, prices and costs for all other commodities, and this is a thriving industry, when prices are going down, there must be quite a~ Mr. CUTLER. You have convinced me, if I needed any convincing, that no fact presented can make an impression on this committee. Senator NELSON. Just a minute. Mr. CUTLER. If the fact that consumer prices for drugs and whole- sale prices for drugs have gone down proves a lack of competition in the industry to Mr. Gordon, I wonder what the fact that they have gone up means. Senator NELSON. Well, Mr. Gordon is not the committee and not a member of the committee. Some fact may not impress him, but do not say that no fact impresses this committee. Mr. CUTLER. I am sorry, Mr. Chairman; I withdraw that remark. Mr. GORDON. May I comment on that? Senator NELSON. I do not see anything fruitful in it. We have gone through this with Mr. Ross and I will put in the record the colloquy between me and Mr. Ross. (The material referred to follows:) IExcerpts from hearing May 16, 1967, Competitive Problems In the Drug Industry, Part 1, pp. 201-2021 Senator NELSON. In your index, then, you don't include any statistics to show what consumers are paying for drugs. You don't include any statistics showing the differential in prices from year to year in prescription cost? Mr. Ross. No, we don't. Senator NELSON. Don't you think that would be of some significance? What puzzles me a little is if people are paying 50 percent more per prescription now, or 15 percent more per prescription now, than 10 years ago, that it has some significance. Mr. Ross. I don't challenge the significance of that for a minute. I merely say that is not the nature of a price index of our type. PAGENO="0150" 1700 COMPETITIVE PROBLEMS IN THE DRUG INDUSTRY Senator NELSON. But it is a price index of a type. Mr. Ross. It is a price index of a type, the problem being that what it is. measuring is open-ended. * * * * * * * Senator NELSON. I think that what is misleading is that your statistics are used to indicate there is only-I think the interpretation put on your index obviously is incorrect, that the price of drugs that the consumer is buying has. risen 2.9 percent. Mr. Ross. We have never interpreted it that way. Senator NELSON. No, no; I say I think that the interpretation put on your index, because it is not understood, whereas the fact is that the price per prescription from the Lilly statistics has gone up 50 percent in 10 years and not 2.9 percent, and I think it is very important to make a distinction between the nature of your index- Mr. Ross. Yes. Senator Nui~sox. And this other index. Mr. Ross. I fully agree that that is a very important distinction. I would argue that our type of index, which measures change in prices for specified items, is very important for many purposes, but as you say, it does not give you an average price of the other type of index. Senator NELSON. I would agree, and I am certain that from a statistical stand- point, for the purpose that you have in mind, I have great confidence in the validity of your index, but actually it was misleading to me, and I assume it is misleading to a few other people anyway who read it. Mr. Ross. Yes, sir. Senator NELSON. The fact is that if the price of my prescriptions have gone up 50 percent, I am paying out the money, my cost of living with respect to that prescription has gone up 50 and not 2.9 percent. Mr. Ross. Yes. Senator NELSON. The index does not measure some things. Dr. FmESTONE. May I comment on that? Senator NELSON. Yes. Dr. FmESTONE. At the bottom of page 19, average prescription prices, a very simple statistic that has bean used at times to measure what has happened to drug prices, is the average prescription price. This, if not properly interpreted, can be a very misleading yardstick. In an independent analysis of average prescription prices, Prof. Robert V. Eva.nson. head of the Department of Pharmacy Adminis- tration. School of Pharmacal Sciences, Purdue University, has arrived at conclusions essentially similar to those presented in my discussion. Mr. Chairman, I would like to call the committee's attention to Dr. Evanson's pa.per as published in the Septe~nber 1967 issue of Tile and Till, pp. 58-63, a.nd I respectfully request that this paper be made a~ part of the record of these hearings. Senator NELSON. Do you have it with you? You did submit that paper with your statement? Dr. FIRESTONE. Yes. Senator NELSON. That will he made a part of the record with your statement..' Dr. FmESTONE. Thank you. In contrast with a price index, an average prescription price has a number of variables other than the price of t.he drug. No a.ccount is taken of the fact that the mix of prescriptions written and filled in a given period of time may vary from the mix found at some other time. Thus, the average prescription price can rise or fall even if the price of each of the drugs included in the average has not changed by 1 cent.. 2 See p. 1714, infra. PAGENO="0151" COMPETITIVE PROBLEMS IN THE DRUG INDUSTRY 1701 Over the past few decades new drugs and medicines have made im- portant contributions to the health of this country and are in part re- sponsible for a declining death rate and a rising life expectancy. New products have been developed to provide drug therapy where none had been available previously, e.g., tranquilizers and drugs that are more certain and speedy in effecting cures, the so-called wonder drugs. Many of these drugs are higher priced than the products they re- placed on a per dose basis. But this kind of comparison is no more valid than a comparison of jet fares with railroad fares simply on the basis of miles traveled. Just as it makes more sense to measure the comparative cost of a trip from Washington to Los Angeles by plane and train taking into account not only the fare, but all other incidental costs such as meals, and time lost in travel, so is it more reasonable to measure the total drug cost over the period of an illness rather than the price of a prescription. But more than that, except for the geriatric illnesses which come with prolonged lifespans, new more sophisticated drugs shorten the duration of an illness; shorten the required period of medical, nursing, and hospital care; reduce the pain and discomfort suffered by the patient; put him back on his feet, and back to his job more quickly, reducing his earnings loss due to illness. I cite this, Mr. Chairman, not as argument, but as sound economics which goes to the question of real cost. For some specific illnesses these savings can be measures in dollars and cents. Before the development of .antibiotics, for example, lobar pneumonia meant 5 weeks of hos- pitalization, long convalescence, and several hundred dollars for doc- tors, nurses, medicines, oxygen, and hospital care. Today pneumonia means a week to 10 days of illness, usually at home, $15 to $30 for drugs, and just two or three visits from the doctor; $15 to $30 for drugs sounds expensive. What is generally overlooked is the astonishing re- duction of all other medical costs, and the reduction in loss of time from work. There is a real need therefore, I submit, for some measure which will show the true change in the total cost of treating an ailment. A some- what higher drug cost may result in a considerably lower total cost in effecting a cure. Even such an index could not evaluate the lowered mortality rate, the shorter duration of an illness with the accompany- ing relief from disability, pain and absence from one's occupation. There have been occasional attempts to develop such a measure, but to date these have not provided adequate data. A principal source of difficulty is the unavailability of suitable data concerning medical,. hospital, nursing, and medication costs in earlier years. The Depart- ment of Health, Education, and Welfare might be the appropriate agency to develop such measures. In summary, I would say that average prescription prices cannot tell us what has been happening to drug prices. There are too many factors other than changes in drug prices that tend to distort this measure. In fact, for nearly. two decades prices for ethical pharmaceuticals, as measured by the most reliable price indexes, have declined, while the price of an average prescription has been rising. Neither average prescription prices nor an index of pharmaceutical prices can tell us what is happening to our total expenditures for drugs. The increase in our lifespans and the advances of medical science inevitably must add to our drug bill. The continuing research PAGENO="0152" 1702 COMPETITIVE PROBLEMS IN THE DRUG II~DUSTR1 discoveries in the chemotherapeutic management of ailments which previously baffled the medical profession, have resulted in more ex- tensive drug usage and an increase in our total drug bill. It is impor- tant, therefore, that. while we note an increase in a patient's drug bill, we do not overlook the tremendously increased benefit to his well-being and the resulting positive contribution to our economy. I hope, Mr. Chairman, and gentlemen, that these remarks have shed additional hght on drug prices and will contribute to a better under- standing of the elements properly entering into the measuring of price changes. Senator NELSON. Thank you very much, Dr. Firestone, for your very thoughtful presentation. We appreciate your taking the time to come here. Mr. FIRESTONE. Thank you, Mr. Chairman. (The complete prepared statement and supplemental information submitted by Dr. Firestone follows:) STATEMENT OF JOHN M. FIRESTONE, Pu. D., PROFESSOR OF EcoNoMIcs, COLLEGE OF THE CITY OF NEW YORK Mr. Chairman and Members of the Subcommittee, my name is John M. Fire- stone. I am a Professor of Economics at the College of the City of New York where I have taught since 1929. Much of my work as a consultant to industry and government involves my principal field of interest-price statistics and indexes. In this area I have served as a consultant to the U.S. Bureau of Labor Statistics and have designed and developed price indexes for the War Department, the National Housing Agency and for non-governmental organizations. I am a Fellow of the American Statistical Association and of the American Association for the Advancement of Science, and am a member of the American Economic Association and the National Association of Business Economists. My testimony here is related to my activity as a consultant for the past six years to the pharmaceutical Manufacturers Association. I have been requested to address my remarks to drug prices and particularly to testimony on this subject which has been presented by earlier witnesses. First may I express my appreciation to the Chairman and to members of the committee for the opportunity to appear before you and may I say that I sincerely hope that my testimony will contribute to a better understanding of the factors entering into drug price trends. For the measurement and meaning of changes in price levels is indeed not a simple matter. Price changes cannot be taken at face value, since very often the prices com- pared are not strictly comparable though at first view they may seem to be. For example, in my opinion there has been a tendency to compare prices which are not comparable and there has been a failure in some quarters to properly differentiate between wholesale and retail prices, and wholesale and retail price indexes. All of these are appropriate measures that have their place in economic analysis. but they can be used properly only if the measures are correctly under- stood and interpreted. I should like to discuss briefly the ways in which wholesale and retail price indexes are constructed, how they differ from each other, what they mean, and as important, what they do not mean in respect to drug prices. I should also like to examine the concept of an average prescription price, and point out the pitfalls in using this as a measure of change in the price of drugs, particu- larly at the wholesale level. Finally, I should like to point out that just measur- ing the dollars and cents changes in drug prices, without taking account of the changing benefits derived from their use, may result in wholly erroneous conclusions. BL~ price indeves The United States Department of Labor's Bureau of Labor Statistics (BLS) has earned the well-deserved reputation of being the world's outstanding author- ity on the subject of price indexes. The BLS indexes of wholesale and retail PAGENO="0153" COMPETITIVE PROBLEMS IN THE DRUG INDUSTRY 1703 (consumers') prices are not equalled in quality, integrity or accuracy by any other indexes known to me. Statisticians come from almost every country to study the technique of index construction developed and executed by the Bureau's Price Division, so that they may go back to their own countries and try to develop price indexes modeled after ours. Purpose of mdccv numbers 1 Price indexes are designed to do only one thing. They measure the change in prices over time for the universe of commodities which the index is meant to represent. The index number does not, for instance, show the effect on the housewife's market basket when she substitutes cake mixes for the crude in- gredients or frozen packaged vegetables for a sack of potatoes. It does not tell us whether we are using more expensive or less expensive commodities than we used in the past. It does tell us whether, once introduced into the index, they and the other components of the market basket have risen or fallen in price from one period to the next. HOW INDEXES ARE CONSTRUCTED While technical descriptions are sometimes boring it is necessary to under- stand how indexes are put together and how they may be properly used if one is to appreciate the true significance of the changes reflected by them. Price data Index numbers are designed to show us the amount and direction of price movement in the aggregate for it is impossible to discuss the net effect of the many and diffuse price movements of all items. The total number of price quota- tions is almost infinite. Furthermore, there is a wide range of commodities, styles and sizes. Therefore, it becomes a practical necessity to select a representative sample of the many variables involved. These variables include, among others: (a) the commodity composition of the category being studied (for instance, if we are considering groceries, what items should go into the market basket) ; (b) the grades, sizes, and specifications of the commodities chosen; (c) the classes of buyers and sellers to be selected and their terms of trade; and (d) the geographic areas to be covered. Once the market in which prices are to be measured has been defined, a sample must be taken of the commodities and of the specifications to be priced. For prescription pharmaceuticals for example, a selection must be made of package size, dosage form and strength, of transaction~ size, of terms of sale and of points in time and place at which these prices are to be taken. Often these decisions are limited by the availability of data, such as when~ how often, where and for which of the items prices are published. Obviously, prices for the most important products, at the most usual terms of sale are likely to be readily available on a continuing basis. But even for these commodi- ties it is important to set rigid specifications so that variations due to changing transaction terms or grades of commodities do not affect the price comparisons. Weighing In constructing a price index, logical weights (measures of the importance of the products selected for the index) must be employed. These weights should, as truly as possible, reflect the relative importance of the commodities in the index. Changes in the prices of the most widely and frequently used commodi- ties should have greater influence on the index than changes of price in less important or infrequently used products. This information for weighting is usually harder to come by than price data and is frequently unavailable except through specially designed surveys. It is relatively easy to ascertain the going price of an item at any time. It is much harder to find out for any time period bow much of the product was produced, sold, or consumed, particularly when there are an unknown number of producers or distributers of the item. Too often this information is considered confidential by the seller or not tabulated in a way that is useful to a price statistician. 1 For a fuller discussion of Index numbers, their purposes and uses as well as methods of construction see, Mitchell, W. C., "The Making and Using of Index Numbers", Bull. #284, U.S. Bureau of Labor Statistics; Fisher, Irving, "The Making of Index Numbers"; King, W. I., "Index Numbers Elucidated"; Mills, F. C., "The Behavior of Prices"; Mudgett, R. D., "Index Numbers"; Persons, W. M., "The Construction of Index Numbers"; The National Bureau of Economic Research, "The Price Statistics of the Federal Government". PAGENO="0154" 1704 COMPETITIVE PROBLEMS fl~ THE DRUG fl~DUSTRY The BLS, in constructing its Wholesale Price Index (WPI), relies on various government and other advisory agencies to supply suitable data for weights. BLS generally has little choice as to when or how often these data are collected or as to how quickly the data are processed and released. For its Consumers' Price Index (CPI), BLS cannot fall back on any agency for data suitable for weights. It must therefore conduct its approximately decennial Survey of Consumer Expenditures. Without these special studies of consumer spending patterns, the consumer price index could not be constructed, or revised to reflect changes in these patterns. Changes in weights and commodities These studies of relative importance of commodities cannot be "once and for all" studies. Like most else in our society, they become obsolescent. For, in a dynamic economy such as ours, new products are constantly appearing on the market, many older products diminish in importance and often disappear al- together. Also the use of existing products changes continuously as custom and needs decree. Since such information can be obtained only after the fact, it is somewhat obsolete even when first used (applying information for a past period of time to a current situation) and it steadily becomes more obsolete with each passing period. The rate of change in the relative importance of commodities, and in the birth rates and death rates of commodities varies in time and with the class of products. In any event, such changes are universal. No price index in the present day can long endure without an updating of the list of items for which prices are measured and without a re-evaluation of the importance of the individual prod- ucts and the classes to which they belong. Because changes in drug usage are so rapid and steady, BLS has found it necessary to contact the American Pharma- ceutical Association annually for information on weight changes.2 Due to the high cost usually entailed in collecting data, revisions of indexes nre infrequent, and the accuracy with which the index measures average price changes in impaired. In the 1920s and 1930s, a Census of Manufactures was taken every other year. As soon as each of these censuses became available (once ~very two years), the Bureau of Labor Statistics revised its wholesale product weights and commodity lists. Since World War II, censuses were taken only in 1947, 1954, 1958, and 1963. The data from these surveys were incorporated into the Wholesale Price Index as soon as they were available. This, however, has ~ntailed a considerable time lag. Thus until 1967, as long as nine years after the fact, it was necessary to employ 1958 weights. The 1963 weights could not be introduced until this year. Accordingly, when new weights are introduced into the index, they are already three to four years old, and at the time they are changed they are nearly ten years old. The Consumer Price Index has undergone even more infrequent weight changes. The 1960-1962 weights still in use in this index were introduced in 1964. Prior to that, 1952 was the weighting base; and until 1954, the 1934-1936 Consumer Expenditure Survey was used (with some necessary interim adjust- ments). One can readily see, therefore, how unrealistic consumption patterns in the depressed mid 1930s must have seemed in the first post World War II ~1ecade. In general terms, for there are no objective criteria available, changes should be made as frequently and as early as possible in order to more truly reflect current conditions. As a case in point, the British Government kept the com- modity list in its price index unchanged from 1913 to 1947. At this latter date, with some difficulty, women's high button shoes, whalebone-stay corsets, and kerosene lamps (for lighting) were still being priced. Contrast that with the U.S. War Department's Index of Contract Price Changes, maintained during World War II, when commodity weights and lists were revised every six months to keep pace with the changing character of the war and its requirements. 2 See statement of Commissioner Ross before this Committee May 16, 1967 (transcript pp. 230-259). PAGENO="0155" COMPETITIVE PROBLEMS IN THE DRUG INDUSTRY 1705 If data are readily available, and the cost of revision is not too great, annual changes in weights and commodity lists should be made, particularly in a product -area where changes are rapid and dramatic. Such frequent changes permit in- dexes to reflect a more gradual transition to current conditions than when changes are made at infrequent intervals. In any event, changes are inevitable. They are unavoidable if the index is to have any meaning. `Quality changes The Price Statistics Review Committee, in its 1961 report to the Office of Statistical Standards of the Bureau of the Budget, stated that "if a poli were taken of professional economists and statisticians, in all probability they would designate (and by a wide majority) the failure of the price indexes to take full account of quality changes as the most important defect in these indexes. And by almost as large a majority, they would believe that this failure introduces a systematic upward bias in the price indexes-that quality changes have on average been quality improvements." This is no reflection on the ability or in- tegrity of index statisticians. It merely points up the difficulty of measuring in dollars and cents the value to the purchaser of style changes, improved efficiency of a new model of refrigerator, the more silent operation or greater durability of a room air conditioner, or the speed of action and the sustained duration of pain relief from a new drug as compared to its predecessors. WHOLESALE PRICE INDEXES The BLS Wholesale Price Indea' The BLS Wholesale Price Index is designed to measure changes in the prices of all commodities sold in commercial transactions in primary markets of the United States including Alaska and Hawaii. A primary market is defined by the BLS as the first commercial transaction for a product, so as to avoid con- iusion with subsequent wholesale transactions. Commodities produced in the United States are included, as well as those imported for sale. They include manufactured and processed goods and the output of manufacturing, agriculture, forestry, fishing, mining, gas and electricity. All these commodities, taking ac- count of grade, size and other specifications number in the hundreds of thousands -and the total number of price transactions for them in a month are in the untold millions. To measure what happens to prices, the Bureau collects actual price quotations -on a sample of about 2300 representative items from manufacturers and other producers. The actual number of price quotations collected per month numbers nearly 7000, in order to represent various markets and multiple producers of a given product. These 2300 items are classified into 15 major groups, 90 sub- groups, 314 product classes, and 37 subproduct classes. Thus, "antibiotics" is a subproduct class of the Ethical Pharmaceutical Preparations product class of the Drug and Pharmaceutical -subgroup of the Chemicals and Allied Products major group. - Ethical Pharmaceutical Preparations account for about 0.5 percent of the total importance of commodities in the WPI. If we add to this Proprietary Phar- maceutical Preparations and Drug and Pharmaceutical Materials (the other product classes of the Drugs and Pharmaceutical subgroup), the combined importance to the economy is less than 0.9 percent. I point this out to demonstrate that a relatively larger number of commodities are used to construct these index -components of the WPI than for most other product classes. There are 31 drug -and pharmaceutical materials, 55 ethical pharmaceuticals divided into 15 sub- product classes and 24 proprietary pharmaceuticals divided into 8 subproduct classes making a total of 110 products. A proportionate share, that is 0.9 percent of the 2300 products used for the entire WPI would be 21 which probably results in the drug category being more representative -than most other categories of like weight. Even before -the revision and upclating~ of the pharmaceutical indexes in 1961, 37 products were included. Thus, pharmaceuticals are not glossed over in the WPI. But even this list of products leaves something to be desired in a definitive study of drug prices, PAGENO="0156" 1706 COMPETITIVE PROBLEMS IN THE DRUG INDUSTRY At the time of the hearings on drugs conducted by the late Senator Estes Kefauver eight years ago, the PMA and the pharmaceutical industry believed, and quite correctly, that the drug and pharmaceutical component of the BLS WPI did not measure fully, nor correctly, what was happening to drug prices. This is not to say that the attention given to drug products did not represent a commensurate share of the funds available for constructing the WPI, at a time when the national spotlight had not as yet been focused on drug prices. With the vast universe of commodities that had to be accounted for in the whole- sale price index under a very limited budget, it is understandable that a group of products that accounts for less than one percent of the value of all commodities did not receive more attention, particularly with respect to the updating of the list of pharmaceuticals priced for the index. The revision of the wholesale price index of drugs by the BLS in 1961 went far toward correcting the earlier deficiencies. Prior to this revision, the index included a large number of such products as agar, citric acid, ethyl alcohol and glycerin-old standbys of the pharmacist's trade, but hardly representative of the products used even in the 1950s. The only relatively new products included as late as 1960 were penicillin, streptomycin, and four sulfa drugs. No represen- tation was accorded the hormones, diuretics, cardiovasculars, tranquilizers, corti- cal steriods or the most important of the pharmaceutical groups, broad spectrum antibiotics. Before the BLS had indicated that it was revising its drug index, I had been asked by the industry to construct the index now called the PMA Wholesale Price Index of Ethical Pharmaceuticals. To do this, I went back to 1949, and with a more comprehensive list of drugs, constructed an index which showed price movements since that date. The construction of this index is described in a succeeding portion of this paper. It is pertinent to compare the BLS Drug and Pharmaceutical Index with the PMA index for the periods 1949 to 1960, the period preceding the BLS revision. and 1961 to 1966, the period after the revision. From table IA and Chart IA, one can see that the BLS index in the earlier period showed practically no change in drug prices while the PMA index declined 7.6 percent. The comparison for the later period (with no change in the technique employed in constructing the PMA index) shows that the movements of the two indexes were very similar, with the BLS index showing an even larger decline than show-n by the PMA index. (See Table I-B and Chart I-B). TABLE 1.-COMPARISON OF THE BLS AND PMA WHOLESALE PRICE INDEXES FOR DRUGS AND PHARMACEUTICALS 1949-66 (BASE PERIOD 1949), AND 1961-66 (BASE PERIOD 1961) Year 1949-6 BLS (A) 6(194 9=100) PMA (B) 1961-66 (1961=100) BLS PMA - 1949 100.0 100.0 1950 98.8 95.9 1951 102.2 96.2 1952 98.8 93.0 1953 99.3 93.0 1954 100.3 91.8 1955 99.2 91.4 1956 98.4 91.4 1957 99.7 92.7 1958 100.4 93.0 1959 99.6 92.7 1960 100.1 92.4 1961 99.9 89.7 100.0 100.0 1962 97. 5 87. 1 97. 6 97. 1 1963 96.4 86.2 96.4 96. 1 1964 96. 1 86. 0 96. 1 95. 9 1965 95.4 86.2 95.4 96.1 1966 94. 8 85. 9 94. 9 95. 8 PAGENO="0157" COMPETITIVE PROBLEMS IN THE DRUG INDUSTRY 1707 8F 1949 1952 1955 1958 1961 1964 1966 Chart I (B) Base Period 196 1=100 ~100 ~.-BLS PMA~~ 90 _J_. L~ 1964 1966 The PITA Wholesale Price Inclew of Ethical Pharmaceuticals Before detailing the construction of the PMA Wholesale Price Index of Ethical Pharmaceuticals, I should like to state that I have always had complete freedom in designing and constructing this index. This is the only way I would do it; the Pharmaceutical Manufacturers Association never asked me to do it other- wise. As a result, I alone am responsible for, this index in its content and its accuracy. The index measures the prices of ethical pharmaceuticals as reflected in the price quotations for those brand name products which are prescribed fre- quently enough to be recorded in national surveys. The sources of the price data are revelant editions of a trade journal annual, the Drug Topics Red Book.3 While the Red Book shows prices by distributor for generic products (no brand name), to my knowledge, there are no satisfactory available data on the quan- tities of these products that are dispensed. In any event, indications point to few, if any, generic producers selling sufficient quantities or with enough market activity to prompt inclusion in the analysis. The items included in the PMA index are volume sellers and, therefore, are significant. The survey on sales volume, described here after, does not include data on generic products by distributor. For example, there is information on the sales volume of gen- erically dispensed tetracycline in the aggregate, but not by distributors. Since Published by Topics Publishing Co. of New York City. Chart I (A) bLS and-PMA WHOLESALE PRICE INDEXES for DRUGS and PHARMACEUTICALS Base Period 1949=100 Index 105 100 ~95 ~90 ~4~BLS - PMA 1949 1952 1955 1958 1961 PAGENO="0158" 1708 COMPETITIVE PROBLEMS IN THE DRUG INDUSTRY there is wide price variation in generically produced tetracycline with divergent price movements, it is impossible without quantity data to measure the effect of these price changes. The 1907 Red Book shows 5 distributors(by generic term only) increased prices for tetracycline, 2 had decreases, and 9 had no price changes. Eight distributors dropped out of the Red Book listing and nine new distributors entered. In addition, it is doubtful whether any non-specialty generic distributor has sufficient dollar volume for individual products to warrant inclusion in the index under the criteria established. It should be made clear that the prices used in the index which I have prepared, are list prices as reported by the producers or distributors. There are no adjustments for trade, quantity or cash discounts nor for free goods since such information is not available. This deficiency would affect the index only if there were trends which lead to effective increases, or decreases in dis- counts off list, over the years. Short run changes, without such changes in trends tend to counterbalance each other in the long run and probably have no lasting effect on the index. The weights for this index were derived from confidential proprietary data compiled by Medical Data Services a Market Research Firm of Darien, Con- necticut from its "Prescription Panel Service" and provided to me on a cosifiden- tial basis. Medical Data Services describes its service as providing information, from a scientifically designed sample of new prescriptions filled at pharmacies,. "on national prescription movement in order that a client may determine trends by therapeutic class by company, and by product form and strength." The cumulative total of prescriptions for each product in the sample for an entire year, multiplied by the average prescription price for that product, was. used to determine the relative importance of each of the prescription drugs. and to determine the total value of all prescriptions for a therapeutic category. This information is available within a month or so after the year-end, and provides reasonably good information on drug usage as reflected in prescrip- tions actually filled. A parallel but independent service, that of R. A. Gosselin & Company, of Dedham, Massachusetts, entitled "National Prescription Audit" was used together with the "Prescription Panel Service" in the preparation of the 1966 index to verify and reinforce the weights developed. The two services were remarkably similar in the drug usage patterns they revealed. The use of either service alone would have provided sufficiently accurate information with which to design the weighting system, but to reduce sampling errors to a minimum, the data from the two services were combined. Commodity selection From the outset, sixteen categories of drugs were selected for the index plus an "all other" group to include the remaining drug products. Drug specialties in each category were selected on the basis of their importance in their theraputic class. In each of the years since 1949 a product was included in the index if it was one of the five leading drugs in its group, so that all important specialties could be included. Items were dropped from the index only when the dollar volume fell below 0.1 percent of the group volume, to insure stability of the product list. Table II shows changes that were made in the number of drugs used in constructing the PMA index: TABLE Il-DRUGS INCLUDED IN THE PMA ETHICAL PHARMACEUTICAL INDEX, 1949-66 Number of items Total Year number Added Deleted included Number of items Total Year number Added Deleted included 1949 99 99 1950 18 1 116 1951 13 5 124 1952 46 7 163 1953 22 4 181 1954 21 7 195 1955 20 4 211 1956 14 4 221 1957 21 5 237 1958 14 7 244 1959 24 20 248 1960 18 19 247 1961 14 1 260 1967 2 0 262 1963 11 0 273 1964 8 1 280 1965 4 4 280 19661 189 106 363 1 In 1966 the basis of item selo:tion was changed somewhat to get better representation for the important products in groups like hormonas and antibiotics which were previously excluded from the index because they were not among the 5 leading products in their respective categories. Items with $1,000,000 in prescriptions in each ot 2 consecutive years. were added to the index. A product, already in the index that did not have at least $1,000,000 in prescription volume in either year, was dropped except that, no therapeutic group is ever reduced to fewer than 5 products. PAGENO="0159" COMPETITIVE PROBLEMS IN THE DRUG INDUSTRY 1709 The weight used in constructing the index is the average dollar volume for two consecutive years ending with the year of reference. Thus a product is on the market at least two years with substantial dollar volume in prescriptions before it is then included in the index for the second year. The result is that only products which become established in the market can be included in the index. Initial exploratory prices during the introductory period for the product would not generally affect the index. Indication of the rapid changes that have taken place in drug usage may `be seen from the fact that only 27 of the original 99 products are still included in the index. This is considetably less than 10 percent of the total number of products included in the 1966 index. An index `based on these 27 items at 1949 weights would stand at 105.9 in 1966. This, however, would be a most unrepresentative list of the drug items in use today. Such an index would be no more an index of drug prices today `than an index based on horse and buggy~ items would be an index of trans- portation today. It should be obvious that failure to recognize and take account of such changes on a current basis when constructing the price index would result in rapid obsolescence of the index. Accordingly, the list of products and their weigh'ts are reviewed each year to make sure that no important item is excluded from the index and to provide for the retirement of drugs that have declined to insignifl'~antly low levels of usage. I note that the Commi~s'ioner of the Bureau of Labor Statistics in recent testimony `before this Committee stated that the BLS has recently adopted this policy for its drug index.4 Uha~im indexes Whenever new items are added to the index `or old items are deleted, or when- ever the relative impottance of the included items changes, the comparability of the index with the preceding period diminishes. To avoid any distortion that might follow from such changes, and to make sure that the only movements in the index are the results of price changes, a new index series is s'tarted at the time of each change. This series has, for both the current period and the im- mediately preceding period, an identical list of commodities with identical weights. Whatever percentage change is shown by `the new series is applied to the index through the preceding period in order to measure the cumulative changes in prices for the life of the index. This procedure is called linking or chaining, and the index series is referred to as a "chain index". In this wa~ an index series reflects price changes and price changes oniy. This is a standard procedure that is followed in all indexes to preserve continuity. To illustrate in a simple fashion what is achieved, let us assume `that our hypothetical index as of 1965 stands at 90.0. In 1966, a new list of products (which normally would contain an overwhelming proportion of the prOducts in the 1965 index) is com- pared with the prices for this identical list in 1965 and it is found that the aver- age price has declined 1.0 percent. The new index for 1966 would then be equal to 90.0 (1965 index) reduced by one percent (of 90.0) or 89.1. Thus, we say in effect that prices that, up to 1965, had declined 10.0 percent (to 90.0), had declined an additional 1.0 percent in 1966 from the 1965 level, making the total decline in the index 10.9 percent (to 89.1). Sub-indeaes of the PIIIA price indev In addition to the PMA Index of Wholesale Prices for Ethical Pharmaceuticals, indexes for various reclassifications of relevent data are available. These are: 1. Therapeutic Group Indexes; 2. Indexes for the group of items covered by patents and those products without patent protection; 3. An index of drug~ used by those who, are 60 years of age and older. What the PMA inde~ves show In the seventen years covered by the PMA Index `of Wholesale Prices of Ethica'l Pharmaceuticals there has been a total decline of 14.1 percent. Declines have occurred in eleven of the seventeen years, increases in four and there `have `been two years with a change of less than 0.05 percent. Of the seventeen therapeutic groups, ten have shown increases and seven have had decreases since 1949. How- ever, the declines of more than 50 percent that have occurred in two of the most important groups, antibiotics and hormones, more than offset increases of much smaller magnitude in less important groups. Statement by Commissioner Ross, op. cit. PAGENO="0160" 1710 COMPETITIVE PROBLEMS IN THE DRUG INDUSTRY When the drug products had been classified by their patent status, patented products showed a decline of 24.8 percent (through 1966) while the non-patented products had risen by 1.1 percent in this same period. Apparently, the products with no patent protection show little price change, and such price change that occurred has been in an upward direction. At the same time, contrary to expec- tation, the existence of `a patei~t has not operated against price declines. (See Table III and Chart II, following pages.) The National Disease and Therapeutic Index, a market analysis compiled by Lea Associates of Ambler, Pennsylvania shows among other data the di~tribu- tion of prescriptions fo~ products by age group. One `of these age groujts is the "60 and over" category, and these data `are available `back to 1959. An index of whosesale drug prices for this age group w~as constructed, with 1959 as a base (equal to 100), and the total drug index was compared with this. The results are shown in the Table IV: TABLE IV.-PMA INDEX OF WHOLESALE PRICES OF ETHICAL PHARMACEUTICALS, 1959-66 [1959 =1001 Year Total index Over 60 index Year Total index Over 60 index 1959 100.0 100.0 1963 92.9 95.8 1960 99.6 99.6 1964 92.7 96.7 1961 1962 96.7 93.9 97.3 95.9 1965 1966 93.0 92.6 96.6 96.5 An important reason for the difference in the decline of the two indexes is that oral contraceptives, which have shown a marked decline, are not included in the "over 60" index. RETAIL PRICE INDEX BL~ en edical care component of the UPI Medical care represents nearly 6 percent of the expenditures of wage earners and clerical workers. Only 0.6 percent of the expenditures are on prescriptions. Professional services of doctors, dentists, optometrists and nurses, hospital sup- plies and nursing home services, health insurance and over-the-counter drugs and sundries account for the remainder, and in all, are nearly ten times as great as expenditures for prescriptions. TABLE Ill-WHOLESALE PRICE INOEXES OF ETHICAL PHARMACEUTICALS (PMA) AND CONSUMER NONDURABLE GOODS OTHER THAN FOOD (BLS) 1949-66 [1949= 1001 Year - Ethical pharmaceuticals Consumer non- durable goods other than food Patedted Nonpatented All 1949 100.0 100.0 100.0 100.0 1950 1951 1952 1953 1954 1955 1956 1957 1958 1959 1960 1961 1962 1963 1964 1965 1966 92. 1 92.2 85.4 84.9 82.7 82.3 82.6 83.8 84.0 83.6 83.4 79.8 77.1 76. 1 75. 6 75. 6 75.2 100. 1 100.7 99.8 100.0 100.6 100.8 100.9 102.7 103.6 103.5 100.8 98.7 98.2 98. 1 99. 1 101. 2 101.1 95. 9 96.2 93.0 93.0 91.8 91.4 91.4 92.7 93.0 92.7 92.4 89.7 87.1 86. 2 86. 0 86. 2 85.9 101. 6 109.4 106.7 107.7 108.0 108.6 110.8 113.3 112.6 114.3 115.1 115.1 115.2 115. 5 115. 2 116. 6 118.8 PAGENO="0161" COMPETITIVE PROBLEMS IN THE DRUG INDUSTRY 1711 Chart II WHOLESALE PRICE INDEXES of ETHICAL PHARMACEUTICALS (PMA) and NON-DURABLE CONSUMER GOODS OTHER THAN FOODS (BLS), 1949-1966 Base Period 1949100 1nde~c 120 Wholesale Prices Non-durable Consumer Goods Other T~ian Food 100 Ethical 90 ~bb., Pharmaceuticals Patented Ethical 80 ~. Pharmaceuticals 70 ~ I I t 1949 1952 1955 1958 1961 1964 1966 It is only since 1960 that the Consumer Price Index has included prescription items that by current standards can be considered truly representative. Prior to that year the prescription index was based on three items-APC tablets, elixir terpin hydrate with codeine, and buffered penicillin 5-hardly a representative list of ethical pharmaceuticals. This index rose 27.8 percent between 1949 and 1960, but I would be reluctant to accept this as an indicator of retail prescription prices. The item list was revised and expanded in 1960 to include seven thera- peutic groups, and just this year three additional groups were added. Each group is represented by only one or two products, with a total of seventeen items to represent the ten groups. While this is a rather thin basis to measure price movements of the individual therapeutic groups, it should be reasonably reliable to measure price movements for all prescriptions. The BLS retail prescription index is available on a quarterly basis while the BLS wholesale index for ethical pharmaceuticals is available on a monthly basis and thus can reflect price move- ments in the short run. Interpretation of price indexes Price indexes are designed to show what changes have taken place in average prices over time. It is therefore important, in constructing a price index, to have price as the only variable. To do this, all other factors, in so far as possible, must be kept constant. That is why in calculating price ratios between two periods of time, the list of commodities and their relative importance must be identical in the two time periods. If prices have changed, the index will reflect the net effect of all price changes that have occurred in the interim. An index Statement by Commissioner Arthur M. Ross before this Committee, May 16, 1967, p. 235. Sl-280-CS-pt. 5-l1 PAGENO="0162" 1712 COMPETITIVE PROBLEMS IN THE DRUG INDUSTRY cannot tell us whether prices are too high or too low, or whether more or less expensive products are in use today than yesterday. An index also does not reflect quality changes introduced in the new drugs. If only half as much of the new drug at twice the unit price does as much or more for the patient this cannot be determined from the price index. A price index answers only one question: for products on the market and included in the index, have prices risen, fallen, or remained stationary and by how much? An index of retail prices should not he used as a gauge of wholesale prices nor should wholesale price indexes be construed as measuring retail price move- ments. Time lag is one important variable. Also, additional elements determining retail prices may be and usually are different from the factors which affect wholesale prices. We find, for example, that consumer prices have a larger ele- ment of wage costs, often in the form of services rendered the purchaser at the retail level. These are more likely to be personal, such as waiting on a cus- tomer, filling his prescription, delivering it to his home, providing credit, and so forth. These -kinds of services are not readily amenable to automation or mech- anization, so that rising wage costs cannot be offset by higher productivity in the same degree as may be the case at the manufacturing level. This may be a principal reason for the greater rise in consumer prices than in wholesale prices during the past twenty years. Thus, since 1957-59 all wholesale prices have risen 6 percent while consumer prices have risen 15 percent. Average prescription prices A very simple statistic that has been used at times to measure what has happened -to drug prices, is the average prescription price. This, if not properly interpreted, can be a very misleading yardstick. In an independent analysis of average prescription prices, Professor Robert 17 Evanson. Head of the Department of Pharmacy Administration, School of Pharmacal Sciences, Purdue University, has arrived at conclusions essentially similar to those presented in my discussion. Mr. Chairman, I would like to call the Committee's attention to Dr. Evanson's paper as published in the September 1967 issue of Tile asd Till, pp. 58-63, and I respectfully request that this paper be made a part of the record of these hearings. In contrast with a price index, an average prescription price has a number of variables other than the price of the drug. No account is taken of the fact that the mix of prescriptions written and filled in a given period of time may vary from the mix found at some other time. Since prices for pharmaceuticals differ from item to item, we would expect the average prescription price to vary as the mix greater or lesser emphasis on higher priced drugs. Thus the average prescription price can rise or fall even if the prices of each of the drugs included in the average have not changed by one cent. This usually occurs seasonally, as the incidence of respiratory ailments changes, for example, but can also occur in an irregular fashion with the incidence of epidemics, or even as a result of new approaches in drug prescription for various ailments as the medical art progresses. More and more, the pharmaceutical industry is introducing sustained action capsules, so that one dosage now does the work of two or more of the former dosages. We can expect these new capsules to be more expensive than the former single dosage, but often the cost for a given amount of medication is less. A. single prescription of this new drug may therefore cost more than a single prescription of the older drug, but at the same time may require fewer refills. Thus the total drug cost during the course of an illness may be lower with the new drug even though the single prescription price is higher. The pharmaceutical industry has been developing combinations of drugs for physicians who wish to prescribe two or three drugs to be taken simultaneously by a patient. The combination naturally may be expected to cost more than any one of its components alone, but it will probably cost less than the total of the components purchased separately. Certainly at the retail level, there would be only one pharmacist markup for filling the prescription for the combination rather than the usual two or three for filling two or three separate prescriptions. An average prescription price would not reflect this saving. On the contrary it would show the higher price for the new single prescription. Also, with advances in medical science, there tends to be an increase in the number -of cases of chronic ailments (persons with chronic heart ailments, diabetes and arthritis, for example). Doctors in such cases are prone to prescribe PAGENO="0163" COMPETITIVE PROBLEMS IN THE DRUG INDUSTRY 1713 drugs in larger quantities to reduce the need for frequent refilling of prescrip- tions and to give the patient the benefit of whatever lower prices might accrue with volume purchases. Thus, even with lower prices for drugs (on a per dose basis), the average prescription price rises. Retail vs. wholesale prices However, what has been most serious in the misuse of average prescription prices is the use of these prices for measuring what has happened to manu- facturers' prices. Since this Committee has concerned itself primarily with manufacturers' prices, I respectfully submit that what should be examined is prices charged by manufacturers before pharmacists' margins `have been added to th'ese prices. A previous witness, for instance, testified before this committee that "In Washington, D.C., one preparation was found to retail for 75 cents at one store, and $2.45 at another".° How this difference can in any way be attributed to the misdeeds of the' manufacturer of this product I cannot under- stand. Even if this manufacturer should cut hi's price for this product in half, this reduction would be almost completely obliterated by the markup of the second pharmacist. Past studies by government agencies and Congressional committee,s have been directed at the growing spread between the prices received by producers of wheat, cotton and wool, and the prices paid by consumers for bread and clothing. Such studies have affirmed what economists are fully aware of; namely, that distribution costs have been growing steadily and that this widening spread does not of necessity reflect price gouging. Most costs have been and are still rising. Taxes, transportation costs and rentals of commercial facilities have generally been rising on an upward curve. Even sharper increase,s have taken place in wages. Automation of retail pharmacy services is practically impossible so that these increased costs cannot be absorbed by the use of more efficient equipment, as has been possible, to a large extent, in manufacturing. These cost increases must therefore be passed along to the ultimate consumer in the form of higher prices and must of necessity be reflected in the higher average prescription price. Total illness cost Over the past few decades new drugs and medicines have made important contribution,s to the health of this country and are in part responsible for a declining death rate and a rising life expectancy. New products have been de- veloped to provide drug therapy where none had been available previously, e.g. tranquilizers and drugs that are more certain and speedy in effective cures, the so-called wonder drugs. Many of these drugs are higher priced than the products `they replaced on a per dose basis. But this kind of comparison is no more valid `than a comparison of jet fares with railroad fares simply on the basis of miles traveled. Just as it makes more sense to measure the comparative c'ost of a trip from Washington to Los Angeles by plane and train taking into account not only `the fare, but all other incidental costs such as meals, and time lost in travel, so is it more rea- sonable to measure the total drug cost over the period of an illness rather than the price of a prescription. But more than that, except f'or th'e geriatric illnesses which come with prolonged life `spans, new more sophisticated drugs shorten the duration of an illness, shorten the required period of medical, nursing and l1os- pital care, reduce the pain and discomfort suffered by the patient, and put him back on his feet and back `to his job m'ore quickly, reducing hi's earnings loss due to illness. I cite this, Mr. Chairman, not as argument, but as sound economics which goes to the question of real cost. For some specific illnesses these savings can be measured in dollars and cents. Before the development of antibiotics, for example, lobar pneumonia meant five weeks of `hospitalization, long convales- cence, and several hundred dollars for doctors, ,nurses, medicines, oxygen and hospital care. Today pneumonia means a week to ten days of illness, usually at home, $15 to $30 for drugs, and just two or three visits from the doctor.7 Fifteen to thirty dollars for drugs sounds expensive. WhOt is generally overlooked is the astonishing reduction of all other medical costs, and the reduction in loss of time from work. From statement by Commissioner Arthur M. Ross before this Committee, May 16, 1967. Kramer, Lucy, "Drugs and Medicine," Public Health Reports, Vol. 73, No. 10, October 1958. PAGENO="0164" 1714 COMPETITIVE PROBLEMS IN THE DRUG INDUSTRY There is a real need; therefore, I submit, for some measure which will show the true change in the total cost of treating an ailment. A somewhat higher drug cost may result in a considerably lower total cost in effecting a cure. Even such an index could not evaluate the lowered mortality rate, the shorter duration of an illness with the accompanying relief from disability, pain and absence from one's occupation. There have been occasional attempts to develop such a measure, but to date these have not provided adequate data. A principal source of difficulty is the unavailability of suitable data concerning medical, hospital, nursing and inedi- cation costs in earlier years. The Department of Health, Education, and Welfare might be the appropriate agency to develop such measures. Several additional observations are pertinent at this point. The introduction of unit dosages in hospitals, for instance, has made available a dose in a sanitary package, properly labelled with a space for the patient's name. That patient is thus assured there is less risk of a mixup in the medication provided, and of con- tamination from handling, and simultaneously the nurse can take care of her medications in much shorter time. The higher dosage cost is more than offset by the advantages that the unit dosage system makes possible. Also an expenditure of relatively few dollars for tranqualizers now makes it possible for a person who formerly had required hospitalization in a mental institution to live at home with his family. And now chemotherapy is possible with patients, who before the advent of these drugs, were not amenable to psychiatric treatment. Our national drug bill must inevitably rise as such new products are developed to cure previously incurable illnesses whether these products implement or replace the use of other more expensive forms of medical services. As indicated previously, we cannot escape the conclusion that the reduction in the mortality rate is accompanied by a rise in the morbidity rate. Many whose lives have been saved by progress in medical care and drug therapy, must rely on continued medical care and drug usage for long periods of time, and, in many instances, for the rest of their lives. Their very existence may depend on contin- uous drug therapy. It is thus inevitable that the quantity of pharmaceuticals consumed in this country must rise, not only because of a growing population, but because of our growing per capita reliance on drugs for protracted periods. CONCLUSION In summary, I would say that average prescription prices cannot tell us what has been happening to drug prices. There are too many factors other than changes in drug prices that tend to distort this measure. I fact, for nearly two decades prices for ethical pharmaceuticals, as measured by the most reliable price indexes, have declined, while the price of an average prescription has been rising. Neither average prescription prices nor an index of pharmaceutical prices can tell us what is happening to our total expenditures for drugs. The increase in our life spans and the advances of medical science inevitably must add to our drug bill. The continuing research discoveries in the chemno-therapeutic management of ailments which previously baffled the medical profession, have resulted in more extensive drug usage and an increase in our total drug bill. It is important, therefore, that while we note an increase in a patient's drug bill, we do not over- look the tremendously increased benefit to his well-being arid the resulting Posi- tive contribution to our economy. I hope, Mr. Chairman, and Gentlemen, that these somewhat lengthy remarks have shed additional light on drug prices and will contribute to a better under- standing of the elements properly entering into the measuring of price changes. [Reprinted from Tile and Till, 53: No. 2 (June) and No. 3 (September), 1967] WHAT Is AN AVERAGE PRESCRIPTION CHARGE? (By Robert V. Evanson, Ph. D., Professor and Head, Department of Pharmacy Administration, School of Pharmacy and Pharmacal Sciences, Purdue University) An average prescription price, more properly referred to today as an average prescription charge, is a well-known item of statistical information in community and hospital pharmacy practice. Simply stated, it is an average value, expressed in dollars, of the total value of prescription revenue divided by the total number of prescription orders dispensed. PAGENO="0165" COMPETITIVE PROBLEMS IN THE DRUG INDUSTRY 1715 A given average may result from operating statistics of a single pharmacy, of several similar pharmacies in a controlled sample, of a number of pharmacies selected purely by chance, or of a sample or complete population of pharmacies in a specific geographical area. It is possible that the average of any of two or more samples may be the same, but the possibility of such a result is reduced by many factors which contribute to determining an average charge. For this reason, it is appropriate to be mindful of these factors when we view trends in average prescription charges, particularly over a period of years as related to actual changes in costs of prescription products. What is an average prescription charge? This might be better asked in the form of "What is an average?" Webster's dictionary requires several hundred words to describe the many concepts and possible applications of the "average." Statis- tically, an average represents a mean value, medial sum, or quantity, made oufof an array of unequal sums or quantities-i.e., an arithmetical mean. A group of n numbers are added, and then the sum is divided by n to arrive at a figure which should express or imply some central tendency of performance or value. In fact,. as generally used, "average" refers to a loose approximation to central tendency, but the centralness depends on the nature of the items being averaged. It is quite possible for a few high values in an otherwise low-valued distribu.tion to pull the average well above the true central tendency of the distribution. In our context, perhaps the most applicable concept of the term "average" is that of a ratio. Then it can be stated that an average is a mean value expressed as a ratio of value or performance between one set of items and another set of items. It may be expressed as a whole number, as a fraction, or as a dollar value. However, as a ratio, it is not necessarily the value of any actual item or the actual performance of any performing unit. An array of prices may produce an average which does not exist as any one of the prices in the array. An average time for a group of runners may not be the time at which any one runner ran the race. Because it is of interest for analysis and planning to review the various trends of operating data and because it is difficult to analyze masses, average prescription charges have been widely published and have been of much interest to community pharmacists. However, to have any value to a .system in which a particular item is used, the item must offer some vestige of useful information or meaning, and it must be considered in the context of its derivation. Useful information or meaning, in the case of an avearge presciption charge, would include the reflecting, meas- uring, or predicting of economic conditions, professional circumstances, inter- workings of market actions, or merely simple trends for these facts or factors. A brief review of factors involved in the determination of prescription charges will indicate the questionable value and character of a single average charge as a meaningful expression of prescription department economics or performance for a single pharmacy. If the charge cannot be meaningful in this respect, it most certainly cannot be so for any other aspect of the drug industry, including manu- facturing, distribution, and the overall practice of pharmacy. Cost factors Two major cost categories are involved in every charge: material costs and operating costs. Variability of material costs can exist in a pharmacy when the' level of operation permits added investment in quantities to realize the advan- tages of quantity-discounted cost prices. Some products are dispensed occasionally and need be purchased and stocked only in small quantities to meet the existing demand. In this case, the cost is well defined for the pharmacist. However, the rate of dispensing other products is high enough to permit buying at regularly offered quantity discounts. When this is possible, the pharmacist may use the full wholesale cost for the smallest unit, the fully discounted cost for the largest unit, or, at his discretion, any value in between necessary to `meet the needs of his customers, to enter and remain in local competition, or to allow for special considerations which are a part of his total community service. The material costs problems is not avoided when an order requires compound- ing. No one questions the fact that compounding involves time, effort, and the use of technical skills. The question arises about adding the value of these factors to the ingredient values in order to determine a cost (total value) for the final product to he dispensed. In some pharmacies, these values are added to establish such a total cost value; in others, they are assumed to be a part of the markup or fee to be added to the ingredient costs. The average charge will be lower or PAGENO="0166" 1716 COMPETITIVE PROBLEMS IN THE DRUG INDUSTRY higher, depending on the procedures followed to determine cost values for both manufactured specialties and custom-made, compounded products. Operating costs seem to present a different problem, since they can be de- termined definitely by the simple process of adding all expenditures paid out (or accrued) during the period. A small portion of the total can be added, by means of a percentage markup or a fiat fee, to each unit dispensed. However, the amount of operating costs to be charged against the prescription department is not always easily or correctly assigned. For the prescription specialty shop, little or no problem exists, since all ex- penses accrue to conducting a professional practice. Only those incurred in selling physicians' supplies or furnishing nondispensing services and supplies (for example, to nursing homes) need be deducted to get a total for the dispensing function. As soon as over-the-counter drugs, other health-needs items, and general merchandise are involved, the assignment of expenses becomes more difficult. Unless adequate records are kept to assign expenses properly, it is possible to undercharge or overcharge the department for its share of the indirect expenses which represent services rendered for the store as a whole. The point here is that, unless some valid means of defining the actual expenses incurred by the prescription department is devised, any markup or flat fee is little better than a purely arbitrary guess. Since these, too, exert a direct effect on individual charges, the resulting average charge is valueless even as an indi- cator of trends for one pharmacy. It follows that an average is even less sig- nificant for any sample of pharmacies, and significance decreases as the sample size increases. A final "cost" is that incurred by the business for the owner's investment- usually referred to as "return on investment and/or net profit." As sales volume increases, a set profit-return goal can be achieved with a small percentage return on sales. Increased turnover rates will reduce the profit needed per turnover. The economic status of customers, local competition, amount of welfare business, and internal merchandising policies all affect the potential profit or lack of profit. Each of these cost factors brings something to bear upon the potential varia- bility or level of prescription charges both individually and collectively. If so many variables exist in the seemingly simple procedure of determining a given charge for a prescription, an average of several charges will be affected by all these variables without reflecting or measuring any of them. Pre$criptio~i miclJ The strongest single control over the average charge for prescriptions is the prescription mix, because it contains all the potential variables which affect the charges represented by an average. Any given charge depends on four variables: the specific product prescribed, the unit price per dosage form, the quantity pre- scribed and/or dispensed per prescription order, and the number of prescriptions dispensed for each product. The selection of product is the function of the medical practitioner, and it depends on his course of therapy and the inherent characteristics or values of a specific product to treat specific diagnosed conditions. The incidence of certain diseases or malfunctions of the human body, specialty practices, prescribing hab- its or patterns, and many other important variables affect the choice of product. Hundreds of products are available with different components, different dosage forms, varying strengths, and different prices per dosage-form unit. The selection of any one will automatically establish the dosage-form unit price. All of the variables previously discussed as cost factors are immediately brought to bear upon the prescription mix when the product of choice is selected and re- corded on the prescription order. The quantity dispensed per order has a direct effect on the total charge and is dependent on the amount prescribed by the physician to carry out his treatment pattern or-in some instances-by the amount the patient can afford. For exam- ple, if 50 tablets at 10 cents each are ordered, the charge would be $5.00. If 100 tablets are ordered, the charge would be $10.00. However, if 100 tablets are or- dered. but the patient cannot afford to pay for more than 50 at one time, two half-orders - could be dispensed over the treatment period, each with a $5.00 charge. A similar effect on the average is exerted by the frequency and charges for prescription renewals. The number of prescriptions for each product will be affected by many different kinds of. variables, including weather, time of year, the range of age levels among a pharmacy's clientele, epidemics of specific diseases, the extent of preventive PAGENO="0167" COMPETITIVE PROBLEMS IN THE DRUG INDUSTRY 1717 health care possible among different socioeconomic groups, prescribing patterns of all the area's practitioners, and the incidence of chronic and acute illness. Any emphasis on particular products or classes of products to meet demands; occa- sioned by these variables will have a direct and immediate effect upon all those charges as well as the average charge for any representative sample of pre- scriptions. To understand better the effects of prescription mix on average charges, look at the following series of hypothetical examples that have been developed and analyzed to illustrate the mix (Table 1). Five products are structured into this example. They could be similar, competing products or different, noncompeting products. For purposes of analysis of both prices and, market fluctuations, all products are asumed to be manufactured specialties prescribed and used to treat similar or identical conditions. Each product is produced and distributed by a different company. The addition of a product indicates its introduction into the market as a new drug at the beginning of the year first shown. Table 1 details specific data for five years as they might be found in the records of a representative pharmacy. A few significant changes will be noted here without any suggestion of the reasons for which the changes occurred. During the period, four products were reduced in price by 5 to 37 percent, yet no similar direct effect was exerted on the average charge. Contrast this with the many instances in which public references to prescription charges state or imply that drug prices or prescription prices are increasing because the average prescription charge is increasing from year to year! TABLE 1.-THE EFFECT OF PRESCRIPTION MIX ON THE AVERAGE CHARGE Price/unit Year Product dosage form No. of prescriptions dispensed Average No. of units! prescriptions Total prescription revenue Average charge 1 A B 2 A B C 3 A B C 0 4 A B C D 5 A B C 0 E $05 .15 .05 .15 .30 .05 .12 .30 .20 .04 .12 .30 .20 .04 .11 .19 .19 .08 1,000 100 30 20 $1,500 300 1,100 `25 1,800 $1.64 500 500 500 30 20 10 750 1,500 1,500 1, 500 `20 3, 750 2. 50 250 600 500 350 30 20 20 10 375 1,440 3,000 700 1,700 `20 5,515 3.24 250 600 500 350 20 10 20 30 200 720 3,000 2,100 1,700 `20 6,020 3.54 200 550 600 300 150 20 10 30 30 40 160 605 3,420 1,710 480 1,800 `26 6,375 3.54 1 The averages given for average No. of units/prescriptions are raw averages not weighted for prescription frequencies. As more products in any class become available, the total number prescribed increases, but more significance is found in the effect of the shifts of each prod- net's numbers. The great number of product A in year 1 kept the average charge quite close to the $1.50 treatment `cost even though 100 patients paid $3.00 for treatment with product B. In each year, the total number dispensed had a different effect upon the average charge by affecting individual market shares and volume shares. As the reader PAGENO="0168" 1718 COMPETITIVE PROBLEMS IN THE DRUG fl~DUSTRY can see, these effects are not directly noticeable in the value of the year's average charge. The average number of units dispensed per order also moved down for some products and up for others with no apparent or predictable pattern. Although it is obvious that increases and decreases in the number dispensed will affect the individual prescription charges directly, nothing here indicates that a change in the average charge will clearly reflect such changes within the sample being studied. Total volume per product accurately describes that product's dollar share of the total revenue from all products. It is neither static nor absolutely predicta- ble for any given period of time. It depends entirely on the three factors which combine to establish individual charges, the sum of which make up the total for each product. In the data given, one product decreases in volume share each suc- cessive year, and one increases each year. Two products increase and then de- crease. However, because the average charge is determined by using the total revenue from all products, there is no way for it to represent or measure any of these changes. In fact, the average charge for year 5 is the same as that for year 4, even though significant changes occurred within each of the factors and affected its ultimate value. Average prescription charges have been determined for every imaginable sam- ple of prescription performance for decades. They have been reported religiously by every publication concerned with the economic status of the practice of phar- macy. They vary every year from report to report by as much as 10 to 20 percent. Characteristically, they have been increasing every year. They are quoted by all manner of personages in pharmacy, government, consumer groups, and the press to prove or disprove many opinions and concepts, including the biases and mis- understandings of those persons who are using an average charge out of context. The sample truth is that the "average prescription charge" is one of the most overused and least significant ratios of performance in American pharmacy. It is affected by everything, but it affects nothing. It is taken to indicate changes in pharmacy's economic facts, but in and of itself it is not capable of reflecting. measuring, or predicting any of these changes. What, then, is an average pre- scription charge? An average prescription charge is a ratio of the total revenue obtained from a specific number of pre~criptions to that number of prescriptions, reduced to a relative valve per prescription and expressed in dollars and cents. It is nothing more and nothing less. Prescription ci? arges versus price indicators The following list will serve to summarize the concepts most commonly re- ferred to in discussions of prescription economics. It summarizes also, in the order of presentation, the basic points made in the first part of this analysis. Specifically, an average prescription charge is not: L The price of any particular product dispensed or sold. 2. A charge for any professional service rendered. 3. An indicator of drug prices or trends in drug prices. 4. An average of all patient treatment costs. Product price An average prescription charge (or price) cannot be the price of commodities dispensed on prescription. Although the products, as commodities in trade, may be priced individually per bottle or per dosage-form unit, such prices are not based on the total value of any of the prescriptions dispensed. The "retail price" for each order dispensed will depend on the price per unit and the number of units involved, with consideration of breaking bulk, special quantity discounts, and certain other exceptions to basic pricing policies and procedures. It also will depend on whether the product is prefabricated or compounded. Another factor is the submersion of any product "price," other than that of the drug dispensed, in the value of the total package of goods and services received by the patient. A simple example is the added cost of a container when the manufacturers' orig- inal. ~outainer is not used. ~erv~ce charge An average charge is not a charge for any of the services rendered in connec- tion with the compounding and/or dispensing of an order. In some cases, a simple service package. properly labeled, is sufficient. In other cases, a pharmacist may have to expend much professional effort and time to satisfy the patient's needs. PAGENO="0169" COMPETITIVE PROBLEMS IN THE DRUG INDUSTRY 1719 The charge for each prescri~rtion may be varied to reflect the effort involved. It may be included in a single markup or a fee added to' the product price. Or it may be ignored entirely for certain persons who cannot afford to pay for prescrip- tion service but require it to sustain their normal health. The charge for a prescription includes the value of the total product dis- pensed-that is, the price of the commodity plus the charge for the services rendered. In some cases, these can be distinguished easily; in others, they are inseparaby intertwined. Therefore, it is impossible to show that an average charge is an average for either factor or for both together. It is clearly and irrefutably definable, but the value of any single unit or groups of units that comprise the average prescription charge actually remains indefinite until sep- arated from the total array of prescription charges being analyzed. For this rea- son, the average is not and cannot be a price or a charge for anything. This fact must be emphasized because too often, in the press and in personal statements, references are made erroneously to the "average price of drugs" in a particular year or to the "average charge for a prescription in 1966." Then, if the 1966 average is greater than that for 1965, the generalization is made that prices of drugs or charges for prescriptions increased in 1966 over 1965 charges. Such a generalization is false as well as misleading, and it cannot logically be made from the two consecutive averages. Nor ~can it be made from the data from which the averages were determined without a meticulous analysis of intragroup items that would hold size, quantity, dosage forms, and services involved con- stant, in order to measure comparable units. Drag prices An analysis of drug products sold over a number of years will bring out specific facts about the prices of pharmaceuticals at various levels in the channels of distribution, it will show that prices of many products have been reduced sig- nificantly, that other prices have necessarily been increased when costs of produc- tion and/or distribution rose as a natural consequence of economic conditions, and that still others have remained constant for long periods of time even though general conditions may have caused significant price increases for other com- modities and services. An analysis of the average prescription charge over the past twenty-three years indicates a consistent but variable rate of increase dollarwise and per- centagewise. Table 2 shows `the annual increases in cents and percentages, as reported each year in the Lilly Digest from the 1944 average of $1.10 to the 1966 average of $3.59. These are averages based on the total sample. TABLE 2.-ANNUAL INCREASES IN AVERAGE Year Average Dollar Percent of charge increase increase Year Average Dollar Percent of , charge increase. increase I 1944 $1.10 1945 1.20 $0.10 9 1946 1.33 .13 10 1947 1.41 .08 6 1948 1.51 .10 7 1949 1.60 .09 6 1950 1.77 .17 10 1951 1.90 .13 7 .1952 2.08 .18 9 1953 2.19 .11 5 1954 2.27 .08 3 1955 2.46 .19 8 1956' 2.62 .16 6 1957 2.85 .23 9 1958' 2.96 .11 4 1959 3.09 .13 4 1960 3.19 .10 3 196L' 3.25 .06 2 1962 3.32 .07 2 1963 3.39 .07 2 196L 3.41 .02 1 1965 3.48 .07 2 1966 3.59 .11 3 1 Percentages are rounded at the first whole number value. The anual increases in the average charge range from $02 to $23 and from 1 to 10 percent. No definite pattern is evident, except for some apparent stability in the last seven years at about 2 percent., Successive eleven-year comparisons provide increases as follows: 1944-1954=115 percent; 1950-1960S0 percent; 1953-1963=55 percent; 1955-1966=46 percent. The total increase from 1944 to 1966 is 226 percent. Taken at face value, these percentages offer erroneous evidence of the relative movements of drug prices. Actually, the movements are not measurable by these data. During this period of twenty-three .years, drug products have changed progressively in form. A substantial number of medications that formerly were PAGENO="0170" 1720 COMPETITIVE PROBLEMS IN THE DRUG INDUSTRY compounded extemporaneously have been replaced by an almost complete prod- uct mix of prefabricated dosage units. There has also been continual change in the kinds and classes of prescription products used. The ratio of renewals to total orders dispensed (based upon the actual numbers) has increased from 39.5 to 53.6 percent. This represents an increase from 2,505 to 10,623 units, or 324 percent. Therefore, it is practically impossible to make comparisons of aver- age charges for the purpose of demonstrating price change over time. It should be apparent at this point that the average prescription charge can in no way be an indication or evaluaton of drug prices at any level of distribu- tion. However, as recently as March 1, 1967, the average price per prescription, taken from a leading pharmacy journal, was cited in a government report for the purpose of comparing prices of drug products. Furthermore, the report aserted that the average prescription price is capable of reflecting changes in drug prices as well as in quantities of drugs dispensed and the use of new drug products. It is unfortunate that this simple ratio should have been given such official promi- nence `to support contentions which it cannot possible support or conclusions which cannot be derived from it. Treatment costs In the first part of this article, the discussion of treatment costs implied that the average charge might be interpreted as an average treatment cost for drugs. It is not possible for an average of all prescription charges to represent an aver- age treatment cost for any illness or condition requiring drugs. It may be neces- sary for the patient to have several renewals of a prescription before his health is regained. He may also require two or more prescription products, with or with- out renewals. He may need OTC drugs or other items to complete his treatment. Therefore, a patient's treatment cost may be only $1.25 (for an inexpensive, once- used prescription), or it may be $25 (for several renewals or more than one prod- uct), even though the average charge in the pharmacy which dispensed his medications was only the $3.59 reported for 1906. Price indicators Indicators in the form of carefully constructed, statistically significant price indexes that validly measure price movements over time are available to the public, the business world, and pharmacy. The Consumer Price Index (CPI) and the Wholesale Price Index (WPI) are measures of monthly price changes, prepared and published by the Bureau of Labor Statistics (BLS) of the 1J.S~ Department of Labor. The "PMA-Firestone Index of Prescription Specialties," a wholesale-level price index, is compiled by Dr. John M. Firestone, professor of economics at `the City College of New York, and published by the Pharmaceuti- cal Manufacturers Association. A brief description of these indicators will pro- vide background for understanding the inability of the average prescription' charge to indicate price trends or changes. Consumer Price Inden The CPI is computed by comparing, at different periods, costs of a fixed set of goods representative of all purchases made by urban wage and clerical workers and ranging from food to medical costs to automobiles. It is popularly called a "market-basket index" or a "cost-of-living index." Technically, it is a price index with fixed or constant weights. The weights are devised to reflect the rela- tive importance in total spending for the particular goods and services which are priced.' The concept of price in the CPI is the amount charged for a particular specified product, identifiable by style, number, or other designation, on a single sale in an outlet at specific terms of sale. Specification pricing is used for a great majority of commodities. In this case, "specification" is a detailed description of the physical characteristics which determine an item's quality and influence its price. Both producers and retailers assist in developing and making changes in the specifications. A complete tabulation includes average expenditures for 1,800 individual items in the consumer expenditure schedule. Drugs included are such products aa liquid tonics, multiple vitamin concentrates, cold tablets or capsules, cough syrups, aspirin compounds, capsules of tetracycline, chiordiazepoxide hydro- `U.S. Department of Labor, Bureau of Labor Statistics: The Consumer Price Index: History and Techniques (Bulletin No. 1517), Washington: U.S. Government Printing Office, 1066. PAGENO="0171" COMPETITIVE PROBLEMS IN THE DRUG ~DUSTRY 1721 chloride, and sodium secobarbital, and elixir of terpin hydrate with codeine. Also included in this index are tablets of sulfisoxazole, buffered penicillin G, pheno- barbital, propantheline bromide, meprobamate, reserpine, crystalline digitoxin, prednisone, and chiorothiazide. Wholesale price ~ndes~ More important to an understanding and appreciation of drug prices and price trends is the BLS Wholesale Price Index. It is structured to measure average changes in prices of commodities sold in primary markets of all fifty states of the union. Its universe comprises all commodities produced or imported for sale. Commodities selected usually include those with the largest shipment values, as shown by industrial censuses.2 The WPI is calculated as a weighted average of price changes in which the weights represent the total net selling value of the commodities. The relative im- portance of an item represents its basic weight used in the index multiplied by the index of price change between the weight date and a later date. The result is expressed as a percentage of the total for all commodities or for the major group. Several major changes have occurred in the sample of drugs and chemicals since the WPI was originated. The separation of drugs and chemicals, the re- moval of cosmetics from the drug group, and more recently, in 1961, the break- down of drug products into ethical and proprietary pharmaceuticals are exam- ples of the kind of continual revision necessary to maintain a factual and realistic analysis of prices. Because of the 1961 revision, it is possible to recognize relative trends in drug prices that were previously obscured by the grouping. A total of 111 items were included in the class of drugs and pharmaceuticals. To show the relative importance of the drug group to all commodities sampled as well as the relative importance of items within the drug group, Table 3 gives these relation- ships for 1961 and for 1964. TABLE 3.-RELATIVE IMPORTANCE OF PHARMACEUTICAL PRODUCTS IN THE WHOLESALE PRICE INDEX Subgroups Relative impo maceuticals items rtanc to e of phar- all WPI Percentage of importanceof individual items within drug group 1961 1 1964 1 1961 2 1964 2 Ethical pharmaceuticals 0. 534 Anti-infectives . 148 Vitamins .086 Ataractics . 057 Dematologicals . 036 Antiarthritics . 030 Cough and cold preparations . 027 Hormones . 023 Sedatives and hypnotics . 023 Cardiovasculars and antihypertensives . 020 Diuretics . 017 Hematinics . 017 Diabetics . 013 Antiobesity preparations . 012 Analgesics . 012 Antispasmodics and anticholinergics . 013 Proprietary pharmaceuticals . 236 Internal analgesics .069 Cough and cold preparations . 041 Antiseptics . 032 Laxatives and elimination aids . 028 Vitamins . 022 Antiacids . 021 External analgesics . 015 Tonics and alteratives . 008 0. 519 100. 00 100. 00 . 131 .087 . 057 . 038 . 030 . 026 . 023 . 022 . 021 . 017 . 017 . 013 . 013 . 012 . 012 27. 72 25. 20 16.11 16.76 10. 67 10. 98 6. 74 7. 32 5. 61 5. 78 5. 06 5. 01 4. 31 4. 43 4. 30 4. 24 3. 74 4. 05 3. 18 3. 28 3. 18 3. 28 2. 43 2. 50 2. 25 2. 50 2. 25 2. 31 2. 43 2. 31 . 241 100. 00 100. 00 .070 . 041 . 033 . 030 . 022 . 022 . 015 . 008 29.23 29.04 17. 37 17. 01 13. 56 13. 69 11. 86 12. 45 9. 32 9. 13 8. 89 9. 13 6. 36 6. 33 3. 38 3. 31 1 U.S. Department of Labor, Bureau of Labor Statistics Wholesale Prices and Price Indexes, 1963 (Bulletin No. 1513). Washington: U.S. Government Printing Office, 1966. 2 Calculated. 2 U.S. Department of Labor, Bureau of Labor Statistics: Wholesale Prices and Price Indexes, 1963 (Bulletin No. 1513). Washington: U.S. Government Printing Office, 1966. PAGENO="0172" 1.722 COMPETITIVE PROBLEMS IN THE DRUG INDUSTRY Pharmaceuticals accounted for 0.77 percent of all commodities listed in the WPI in 1901 and for 0.76 percent of those to be included in the 1964 index. Anti- infectives, vitamins, and ataractics accounted for 54.50 percent of the ethical pharmaceuticals listed in 1961 and for 52.94 percent in 1964. Among proprietary pharmaceuticals, internal analgesics, cough and cold preparations, and antiseptics accounted for 60.16 and 59.74 percent of the products listed in these same years. PMA-Firestone fndea During the 1950's circumstances evolved a need for a better, more precise evalu- ation of prescription drug prices and price changes. Because these were not re- ported individually but as part of a larger group in the BLS Wholesale Price Index, no factual data were immediately available to support the industry's con- tentions that drug prices were affected by competition and economies in produc- tion so as to be reduced over time and that they were decreasing while prices of other nondrug products increased. Therefore, the PMA engaged the services of Dr. Firestone to construct a realistic index of prescription drug prices.3 So that the index of prices would be current and meaningful, it was based on a structured sample of prices for the year 1949. The index measures the changes in prices paid by pharmacies for specialty drugs used in dispensing prescriptions. It is computed by a method called a "chain index of weighted price relatives." ~Uhis provides flexibility for changing weights as often as may be necessary. New drugs can be introduced into the index when they become sufficiently important; drugs declining in importance can be dropped without invalidating the compara- bility of the index over time, while continually reflecting the true current market conditions. The index was set up by listing the five specialties that led in dollar volume in reach of the eleven years from 1949 to 1959. A. drug was and is maintained on the list as long as it appears on the market. All dosage forms of a product are repre- vented by the price of the most widely used dosage form. Prices and sales volume data are derived from Drug Topics Red Book at its press time each year. A total of 308 drugs appeared at one time or another during the initial period. From 99 products, the base for the 1949 index, the sample was expanded, in order to reflect market conditions more accurately, to include 242 products in 1959. Table 4 indi- cates the relative importance of product classes as it changed over the initial period. TABLE 4-PERCENTAGE SHARES OF PRESCRIPTION SPECIALTIES BY PRODUCT CLASS IN THE PMA-FIRESTONE INDEX Product class 1959 1949 Product class 1959 1949 Percent Percent Antibiotics, oral 23 11 Hormones 11 10 Tranquilizers 9 0 Analgesics, narcotics 7 4 Antispasmodics 5 4 Cough and cold preparations 5 6 Vitamins 4 13 Cardiovascular agents 4 4 Sulfonamides 4 3 Sedatives, hypnotics 3 6 Dermatol3gicals 3 4 Percent Percent Diuretics 3 0 Antihistaminics, oral 2 5 Hematinics 2 4 Central-nervous-system stimulants... 2 4 Antiarthritics, antirheumatics 1 1 -- Total 100 100 - - Total number of products in base sample 242 99 -__________________ Note: Data rearranged in 1959 order. Indecces Show Change The development and use of indexes has been necessary in all areas of endeavor in which changes can be measured significantly by quantitative means. The in- dexes described here have three key factors in common: (1) use of a value repre- senting the price of a given item precisely defined (physically, chemically, or otherwise) at a specified point in time; (2) use of weighting factors which recog- nize the item's impact on the market in relation to all other products; (3) flexi- bility of the sample to expand or contract, which permits inclusion of new Firestone, J. 21.: A Price Index for Drugs, Drug & Cosmetic Industry, 87: 470 (Octo- ber), 1060. PAGENO="0173" COMPETITIVE PROBLEMS IN THE DRUG INDUSTRY 1723 products and dropping older or obsolete items to maintain a sample that accu- rately reflects the products in current use. Each index is based upon a model developed to meet its specific objective. Because the model has both fixed and variable elements, any of these which re- quire subsequent adjustment can be changed to correct the index so that it reflects the actual circumstances in each time period. At the outset, the first period's data represented the index base of 100. All subsequent values have been reported in relationship to the base. When conditions change so much that the base is no longer representative of normal circumstances, complete reevaluation of the index numbers to form a new base can be accomplished by simple arithmetic without changing the relationship of the numbers to one another. The primary focus of such an index is on the prices of the products listed. An index for one item would consider only the pricing of that product and not how many units are produced or sold. Since a price index for groups of products must measure the changes for all the items to show an over-all trend, the quantities of each item affect the model only in proportion tO the importance (weights) of the amounts produced or sold. This depends on the objective of the index: to measure production, consumption, or merely changes in prices without relation to either production or consumption. In other words, an index can be structured model- wise to measure any specific kind of change or trend one wishes to measure. Table 5 summarizes the indexes described abOve, using as a base the average of values for 1957 through 1959 (the current BLS base time period) as 100. It is easy to see that the price trends for all consumer items, medical care, and all wholesale commodities were consistently upward. However, the price trends for drugs and pharmaceuticals, when measured separately by the Prescription Drug Subsection of the CPI and by the PMA-Firestone Index or when extracted from the wholesale group, are definitely downward. TABLE 5.-GENERAL AND DRUG PRICE INDEXES 11957-59=1001 PMA- Year . Consumer price i ndex ~ Wholesale price index Firestone index All items Medical care - Prescription drugs All corn- modities Drugs and pharma- ceuticals -~-- Prescription specialties 1949 83. 0 72. 0 80. 3 83. 5 100. 1 107. 8 1950 83. 8 73. 4 82. 4 86. 8 98. 9 103. 3 1951 90. 5 76. 9 86. 4 96. 7 102. 3 103. 7 1952 92. 5 81. 1 87. 5 94. 0 98. 9 100. 2 1953 93. 2 83. 9 87. 5 92. 7 99. 4 100. 2 1954 93. 6 86. 6 89. 2 92. 9 100. 4 98. 9 1955 93. 3 88. 6 90. 4 93. 2 99. 3 98. 5 1956 94. 7 91. 8 93. 2 96. 2 98. 5 98. 5 1957 98. 0 95. 5 96. 5 99. 0 99. 8 99. 9 1958 100.7 100.1 100.7 100.4 100.5 100.2. 1959 101.5 104.4 103.0 100.6 99.7 99.9 1960 103. 1 108. 1 102. 6 100. 7 100. 2 99. 6. 1961 1962 104.2 105.4 111.3 114.2 99.2 95.3 100.3 100.6 98.3 96.0 96.7 93.9 1963 106. 7 117. 0 93. 0 100. 3 95. 1 92. 9 1964 1965 108.1 109. 9 119.4 122. 3 91.8 90. 8 100.5 102. 5 95.0 94. 4 92.7 92. 9; 1966 113. 1 127. 7 90. 6 105. 9 94. 5 92. 6 Each of the indexes discussed above measures price changes annually and accurately reflects trends over time. Should the quantities of products used in- crease or decrease during a year, costs to the consumers would increase or decrease, but prices would not be affected. If more or less expensive products are substituted, total costs may change, but prices will not change. If less of a drug is needed to cure a patient, treatment cost is reduced-but not prices. If one pharmacy charges full prices while another charges discounted prices, total sales of drugs or total costs to patients will be affected, but these indexes will not be affected and will continue to measure only prices in the marketplace. Unfortunately, this is not true for the average prescription charge. It is affected directly by every charge made for every prescription. Although the price of certain tablets may be ten cents each and may remain so for several years, the PAGENO="0174" 1724 COMPETITIVE PROBLEMS IN THE DRUG INDUSTRY dispensing of increasing quantities will cause the average charge to increase. Table 6 depicts the sharp contrast between the price indexes and an index of the average prescription charge as reported in the hilly Dige$t. Although the average prescription charge is an average of prices actually paid by consumers for prescriptions, Table 6 shows beyond doubt that it is not a measure of prices, price changes, or price trends at the consumer level. If it cannot measure or reflect any kind of valid price or charge, and if its index of change moves countercurrent to true price indexes, it can be a major source of confusion and result in a liability rather than an asset to pharmacy when it is taken out of context. TABLE 6.-THE AVERAGE PRESCRIPTION CHARGE INDEX COMPARED WITH PRICE INDEXES [1961 = 100] Year BLS-CPI Prescription drug subsection BLS-WPI ethical pharmaceuticals subsection PMA-Firestone prescription specialties Lilly Digest average prescription charge 1961 100.0 100.0 100.0 100.0 1962 96.1 97.6 97.1 102.2 1963 93.8 96.4 96.1 104.3 1964 1965 92. 5 91. 5 96. 1 95. 4 95. 9 96. 1 104. 9 107. 1 1966 91.3 94.9 95.8 110.5 CONOLTJSION The average prescription charge is not an indicator of drug prices or of prescription charges. However, it can be a useful tool to the pharmacist in under- standing his particular departmental operation. A case in point is that of a pharmacist who decided to lower his drug prices. After a reasonable length of time had passed, he was surprised to find that his average charge continued to rise. This led him to seek the reasons. He found that lowered prices had en- couraged the prescribing of larger quantities per order. His dollar volume re- mained the same, but greater quantities per order had decreased the number of prescriptions and renewals dispensed, which caused the average charge per prescription to increase. Many useful ratios are available to indicate standards for or levels of perform- ance in pharmacies. They depend on data which may vary over time, but they neither measure nor reflect changes or trends in prices. Their usefulness lies in calling attention to an area and, when necessary, triggering further investigation, a review of policies and procedures, and, finally, some specific corrective action for improving operations. As a ratio of prescription revenue to the number of prescriptions dispensed, expressed in dollars and cents per prescription, the average charge has limited value as an over-all measure of the economics of professional performance. Aside from this, it has no other useful value in itself, in economic analysis, or in pharmacy. Senator NELSON. Dr. Whitney? Mr. CUTLER. We have just one more witness, Mr. Chairman, and that is Prof. Simon Whitney, professor of economics at New York University, who has a statement on the subject of profits and prices in this industry. Senator NELSON. Dr. Whitney, we are very pleased to have you come here today. You have been very patient to wait as long as you have to he presented. Your biographical sketch will be printed at this point. Please proceed. (The biographical sketch follows:) BIoGRAPHIcAL SKETCH OF SIMoN N. WHITNEY, P11. P. Present position: Professor of Economics, University College of Arts and Sci- ence, New York University. Education: B.A., Yale University, 1924 and Ph. P., Yale University, 1931. PAGENO="0175" COMPETITIVE PROBLEMS IN THE DRUG INDUSTRY 1725 Academic appointments: Professor of Economics, New York University, 1907 to present. Professor of Economics, Rutgers University, 1901-1967. Lecturer, New York University, 1954-55. Professor, New York University, 1949-54. Associate Professor of Economics, New York University, 1948-49. Dean, Deep Springs (California) Junior College, 1942-48. Instructor of Economics, Yale University, 1926-28. Business experience: Director, Bureau of Economics, Federal Trade Commission, 1956-61. Chief, Research Division, Twentieth Century Fund, 1954-55. Associate Economist, Twentieth Century Fund, 1949-54. Vice President, Econometric Institute, 1949. President, Telluride Association (educational endowment foundation), 1930-31. Principal Economics, Office Export Control and Board of Economic Wel- fare, 1941-42. Economist, O'Ryan Financial Commission to Japan, 1940. Associate, Lionel D. Edie and Company (investment counselors), 1936-40. Staff, Division of Research and Statistics, Board of Governors Federal Reserve System, 1936. Senior Economist, Research and Planning Division, National Recovery Administration, 1934-36. Assistant Economist, Chase National Bank of New York, 1931-34. Economist, Antitrust Division, Department of Justice, 1928-29. STATEMENT OF SIMON N. WHITNEY, PH. D., PROFESSOR, OF ECONOMICS, NEW YORK UNIVERSITY, NEW YORK, N.Y. Dr. WHITNEY. Mr. Chairman, in view of the lateness of the hour, you will want me to abbreviate my talk. I can cut it by about one-third. Senator NELSON. Thank you. Your statement will appear in full in the record. Dr. WHITNEY. Thank you. Then I will just skip here and there, re- wording slightly.1 Senator NELSON. Fine. Dr. WHITNEY. I am beginning about the middle of page 1, Mr. Chairman. More than one distinguished medical witness has charged that branded ethical drug prices and manufacturing profits are "too high". Fortunately, some industry witnesses and those today have anticipated me in part;, but I shall speak as an academic economist who applies the established principles of this science to the problems before you. Economics begins with the truths that goods are not produced with- out work and progress does not occur without saving and constructive investment. Experience points to the desire to better one's self and one's family as the most effective single motive. Reliance on self-interest also minimizes compulsion, conscript;ion and penalties. Clearly important to labor, people being eager to go wherever their wages or salaries will be highest, it is the application to capital that concerns us here. Self-interest tells people to invest where the out- look for profits is best-after taking account of risks-and this in turn is normally where consumer demand is greatest. Our standard of living is proof that, on the whole, the process has worked. Countries which condemn the profit motive are glad to draw from us some of the sur- pluses which it produces. 1 complete prepared statement of Dr. Whitney begins at p. 1740, infra. PAGENO="0176" 1726 COMPETITIVE PROBLEMS IN THE DRUG ~DUSTRY How does this apply to the questions before the subcommittee? Quite simply, drugs are too important to the public welfare for us to withdraw or dampen the incentives which have created and multiplied all our goods. Any university teacher of economics denying the con- tribution of profit in attracting capital-and this might happen, since a reputation for originality is prized in my profession-should be asked what alternative system he recommends for inducing owners of capital to put it where consumers want it and why this alternative would work better. A normal profit will not attract investment where it is most needed, nor will it achieve the second fimct.ioii of profits, stimulating cost re- duction and efficiency. These essential jobs are done by differences, not uniformity, in earnings. How has this worked in practice? The First National City Bank cal- culated the average rate of return on net worth for all manufacturing since 1950 as about 65 percent of that for drugs and medicines. As a result, $3 of additional stockholders' money per dollar of net worth iii 1950 went into drug manufacturing for every dollar into all manufacturing. Mr. Gon1~oN. Dr. Whitney, can you tell us how much of this $3 was raised through the selling of new stock issues? Dr. WHITNEY. Very little. Most of it was reinvested profits. If profits had not been made, they would not have been there to invest. Mr. GORDON. They would have had to go to the capital market, would they not? Dr. WHITNEY. If they had had poorer profits, they could not have raised the money in the capital market. About 25 cents additional went into mea~packing, the least profit- able major manufacturing industry reported. Comparing drugs with meatpacking, 31/3 times the rate of earnings brought 11 or 12 times the rate of investment. The free market process-consumer demand, prof- its, attraction of capital-has worked in drugs brilliantly. Mr. GROSSMAN. Dr. Whitney, may I ask you, you seem to show a great deal of concern here for the investor. At the same time, you mention on page 2 of your original statement, quite simply, that drugs are too important to the public welfare for us to withdraw or dampen the incentives which have created and multiplied all our goods. I spoke on this question earlier, this morning. What do you consider the elements of social responsibility of the drug industry? In other words I am concerned about the investor and the need for the investor to build the industry, but are we not forgetting the other side of the corn? Dr. WHITNEY. The primary responsibility of the industry is to pro- duce the drugs that are vital to our health. 1 believe they are doing it, are they not? That is why we are investigating their practices. As for the investor, I am not one myself. I have refused to invest in this industry. I have considered it too risky. But the investor is im- portant, because only through him can we get the industry going and producing the necessary drugs. Mr. GRossMAN. You say on page 2, "But such a profit will not attract investment where it is most needed." Dr. WHITNEY. That is an average profit. PAGENO="0177" COMPETITIVE PROBLEMS IN THE DRUG INDUSTRY 1727 Mr. GROSSMAN. In other words, if we were to regulate. the drug industry, you would be able to say assuredly that we would have a serious diminution of investment, is that correct? Dr. WHITNEY. Yes, if you reduce their profits, you certainly will. They are the No. 1 industry in profits, in my opinion, over the last 17 or 18 years. The FTC just reported that the other day. Although they have been criticized this morning or this afternoon, still they have issued that, and they have been the No. 1 industry in attracting capitaL Mr. GRoss~1AN. I do not pose to be an economist, but I wonder, on page 2 also, whether you really believe-you said "Adam Smith spoke of `an invisible hand' by which the search for profit leads to the gen- eral good, and also wisely said that we expect our dinners from the self-interest, not the benevolence of butchers and bakers." Do you still believe that? Dr. WHITNEY. Did you understand how I explained that? I will give it to you again in different words. A higher profit, where consumer demand is greater, attracts capital into making that product and sell- ing to consumers. Thus the consumer gets what he wants Now, is it clear? Mr. GROSSMAN. He gets what he wants. He gets it at a better price, you would say? Dr. WHITNEY. The only alternative is to have perhaps the Govern- ment make the thing. Under a free system-we have found that a free system works well. Under our free market system, he gets it by paying enough to induce the capital to go in. He pays more where he wants the product most. Mr. GROSSMAN. I take it you would agree with, I think, Mr. Plotkin, who said earlier that the FTC reports give an improper image to the profits of the drug industry? Dr. WHITNEY. Oh, no. I am not saying that. I did not understand all of `the argument to and fro, because some of it was put forth in a low voice. I do not want to be in the position of criticizing the FTC report. Mr. GROSSMAN. In other words, they, say the drug industry is the most profitable industry. Dr. WHITNEY. Their figures show th~ year 1966, the latest figures, that this is the most profitable. I ran that back a few years and found that compressors and pumps and the ,` auto industry, were running along on that level, too. Mr. GROSSMAN. You say on page 3 `that new capital will avoid an area with the continuing risks of this industry. Can you tell me where you see the continuing risks? I know we have heard from other people where they see continuing risks, but I wanted to get your opinion. In other words, an industry that has made money over the last 16 years, where are the continuing risks? When does a risk become a con- tingency and when does a contingency become so infinitesimal that we are talking about a straw in the wind? Dr. WHITNEY. It is that straw, whatever it is, that has kept people like me from investing in the drug industry. `V\Te are afraid that al- though science in the last 20 years has discovered wonderful new drugs, it may not do it for the next 20 years. I can go on from that, but that will be one point.. 81-280-68-pt. 5-12 PAGENO="0178" 1728 COMPETITIVE PROBLEMS IN THE DRUG INDUSTRY Mr. GROSSMAN. Do you feel this is much more so than in comparable industries? Is it not the same thing? Somebody could find out that if I drink a Coca-Cola I am going to have gout, and the whole Coca- Cola Co. goes down the drain. And they have only one product. Dr. WHITNEY. That is a possibility. Do you consider that a serious one? Mr. GROSSMAN. I consider it as much of a risk as another industry that has made huge profits for 15 years. Dr. WHITNEY. The way I see it is the industry that invests the most heavily of any industry in research and development incurs thereby a tremendous risk. Their profits and so on are being risked in research and development year after year. Mr. GROSSMAN. Is there not anything that industry can do, for ex- ample, to act to prevent these things from occurring? Are there any steps that the industry could take, in terms of reserves against these risks, that they could take to lower the risk, whatever the risk is? Dr. WHITNEY. You mean, for example, if they reduced their dlvi- *dends, put their money in reserve and if 5, 10, 15 years from now, they stop being productive, they would pay the dividend out of that? Mr. GROSSMAN. I am not suggesting- Dr. WHITNEY. You are suggesting reserves. Mr. GROSSMAN. I am asking is there any way to cut down on these risks? Dr. WHITNEY. How can you cut down on the risks of research? That is something for the scientists to do. They do their best to get results. They cannot guarantee them. Mr. GROSSMAN. Are these risks not borne out by profits, then? Dr. WHITNEY. Oh, I think the profits-we do not know. In the long run, we will not know until the end of the 20th century. The profits are high, but the risks are also high. But by point is not that the risks are so high. This is my pred- ecessors' point. Mr. GROSSMAN. When does risk become contingency; when does contingency become infinitesimal? Dr. WHITNEY. In the railroad industry, I think there were no big railroad failures in the 1900's. The New Haven failed in the 1910's, the Milwaukee failed in the 1920's, and then only in the 1930's did it appear wl1at a bid situation railroads had been in for decades, and they had been considered a prune investment. Mr. GROSSMAN. In the report we have been looking at earlier, Searle reported profits of at least 20 percent since 1950. I do not see the 1966 figures. You say you are not an investor. Would you be afraid to invest in that company because of the risk? Dr. WHITNEY. I certainly would. Mr. GROSSMAN. Would you be afraid to invest in General Motors because of the risks? In other words, you are not an investor, I take it. Dr. WHITNEY. You are wrong in that. I have not considered the General Motors risk so severe as the drug company risk. It may be as *severe; I have not considered it so. If the risks were as mild as you make them out to be- Mr. GROSSMAN. I face risks every day. Don't we all? I want to know how n~uch more this risk is than the risk of walking across the :street. PAGENO="0179" COMPETITIVE PROBLEMS IN THE DRUG INDUSTRY 1729 Dr. WHITNEY. No one knows. It depends on what the investors decide. If they are willing to put in the money, then you may assume that the profit is covering the risk. If they are not willing to put it in, you may assume that the profit is not covering the risk. The money has gone in, but that is at a level of profit that has seemed to justify it. My point, in any case, is not that the risks are high. My predecessors have talked about that. But at only this profit level, whether you justify it or not by risk, it is bringing in the money, and the money must go in for the benefit of the health of the future. Mr. CUTLER. Mr. Grossman, you picked a good example in Searle, because Searle's income is concentrated very, very heavily in a partic- ular type of oral contraceptive drug. As you know, oral contraceptives and contraceptives in general are a field in which many, many drug companies around the world are operating, and the possibility of an entirely new method invented by someone else is quite strong. Mr. GROSSMAN. Do you have any idea when this product was de- veloped? Was it developed as far back as 1950? I do not know. I am asking you. Mr. CUTLER. I do not know the answer to that, either. Mr. GROSSMAN. I believe it probably had not been. But one point that I was looking for is, is there any attempt on the industry's part, perhaps, to diversify their products so they would not,- would not that cut down the risk? Dr. WHITNEY. That would cut down the risk, and the investment in lifesaving drugs. That is exactly the action we want to avoid. Do you not want to avoid this industry putting its money into some- thing else instead of into drugs? Mr. GROSSMAN. 1 want to avoid this fantastic conjured-up risk. Dr. WHITNEY. You are thinking of the investors. I am thinking of the health of the people. Do you want the money to go into drugs? Mr. GROSSMAN. Just `finally, did you or did you not answer when the risk becomes a contingency. Is there any line that you see? Dr. WHITNEY. No. Mr. GROSSMAN. You keep giving me: answers, but I would like that one just specifically. When is a risk no longer a risk? Dr. WHITNEY. The railroads prospered for 30 years and it turned out there was a risk after all underlying them. You cannot tell when it will occur. Mr. GROSSMAN. So we just have to wait and see? Dr. WHITNEY. You can sit around waiting. But I want to urge you not to throw a damper into the investment of mon:ey into research in drugs and to force companies to diversify out of a field they are doing so much good in. True, :they are making profit, but they are saving lives. Senator NELSON. Has anybody suggested forcing them? Dr. WHITNEY. Yes; Mr. Grossman has suggested that they should diversify out of drugs into some other less difficult lines. Mr. GROSSMAN. I think I suggested just diversifying t'heir products. That is all I said. I said cannot they diversify so that they `do not `have one product that they rely on. Dr. WHITNEY. Oh, into other drugs. All of them are trying to. They are `still running the risk's. Mr. GROSSMAN. Would this not lessen the risks? PAGENO="0180" 1730 COMPETITIVE PROBLEMS IN THE DRUG INDUSTRY Dr. WHITNEY. They have tried that already. They do not need our advice to do this. They are all struggling to do this. Parke, Davis just led off this week- Mr. GROSSMAN. You really think they are struggling? Dr. WHITNEY. And you think it isso easy. Mr. GROSSMAN. I certainly do not. Dr. WHITNEY. Then we agree. Mr. GROSSMAN. I know it is a struggle. Dr. WHITNEY. I shall go on. If profits are reduced to average, either new capital will avoid an area with the continuing risks of this industry but no chance of a cor- responding profit, and managements will diversify out of ethical drugs and into products of less importance where profits can be earned with- out drawing public or congressional resentment, or human nature will be found to have changed and people to be employing their sav- ings with the primary purpose of helping their fullow men. I would expect the first. What have the drug industry's reinvested profits accomplished? First let me cite two statistical series. The public demand for durgs has been met with a rapidly rising physical volume of production. Consumer purchases of ethical plus proprietary drugs and sundries- an increasing proportion, now over 60 percent, being ethical drugs- expanded by 5.5 percent annually from the first postwar year, 1946, to 1966, as compared with only 3.7 percent for all other consumer expenditures. The second series, research and development expenditures by the industry, advanced from $50 million in 1951 to $416 million in 1966- faster than either profits or net worth. Total B. & D. for 1957-66 seems to have been greater than dividends paid. In 1965, according to the National Science Foundation, 95 percent of this R. & D. was ~nanced privately, as against only 45 percent for all of industry. Many hundreds of new drugs, as documented by earlier PMA wit- nesses, resulted from this profit-motivated research. What they have done for health need not be detailed to this informed body. As an economist, I shall merely mention that the monetary value of working time saved through improvement of physical and mental health and lengthening of life is probably in the billions. We imperil further gains if we strike at so-called "excess" profits which, in fact, have furnished the driving force. For a mere 6 cents in dividends per dollar of sales of all types of products made by com- panies classified in this industry, or less than a cent and a half in dividends for each dollar of total medical care costs, the industry has done its essential job. This is not primarily one of basic pharmaceutical and biological research-though it happens to lead all industries in percentage of B. & P. going to basic research-but one of discovery and bringing into use of new medicines. Eliminate 2 of the 6 pennies in dividends-make them average for all manufacturing-and investor enthusiasm will decline to average also. Does anyone really expect enough phychological satisfaction from insuring that this one industry earns only average returns to warrant endangering future investment in the search for cures of our dread diseases? PAGENO="0181" COMPETITIVE PROBLEMS IN THE DRUG INDUSTRY 1731 It is implicit in my argument that prices are not "too high." I shall review briefly what seem to be the three most common reasons given by those who contend that they are: the size of profits, the con- trast between some branded and unbranded drug prices, and in- ability of low-income groups with chronic illnesses to pay. 1. To call higher than average profits a proof of excessive prices overlooks the economic functions of prOfits, as I have already de- tailed. It forgets also that in a free market earnings tend to vary with risks. When privately financed R. & D. is several times that of other industries in relation to sales, and when drugs become obsolete so fast that over half the 1966 prescriptions were written for drugs not known in 1956, there is plainly a high built-in risk factor. Any search for the new is more hazardous than production of the old. According to economic theory, a short to medium term higher-than- average profit is understandable; but entry of new capital will even- tually bring it to a long term level which is commensurate with risk. In this industry, capital has entered. Rising consumer demand, stimulated by the industry whenever it develops a new and effective drug, and the inherent risks of research which have prevented even more money from entering, are sustaining its earnings. 2. Prices of certain drugs of one manufacturer may be higher than those of another producer which are at least theoretically chemically equivalent. Prices could hardly be the same when one company must cover the costs, and be rewarded for taking risks, of research, testing, quality control, and original distribution, and when it makes all dos- age forms available where and when needed; whereas the maker of the other drug may only manufacture,~ and then probably only the most profitable form. Senator NELSON. On that exact point, that one company must cover the costs of research, testing, quality control, and original distribu- tion. Maybe you can answer the question I asked earlier about Thora- zine, where research was done by Rhoné-Poulenc, a French firm, and they licensed Smith Kline & French in this country, and they ii- censed Bell-Craig in Canada. The firm in Canada sells it to the Gov- ernment for $2.60 a thousand tablets. In the United States it is selling for $32.60 a thousand. As an economist, can you give me an explanation for this dramatic difference? Can you explain why 15 times as much, almost, is charged here for a licensed product in which neither the company here nor the one in Canada did the research? Dr. WHITNEY. Is that a fact, that Smith Kline & French did no research on this product? Senator NELSON. We do not have any evidence of research other than the submissions for the New Drug Application. Dr. WHITNEY. I happen to have read in the Kefauver hearings about the work they have done. Perhaps not the volume that Mr. Gor- don has, but elsewhere. Mr. CUTLER. I did have a chance at recess to check with one of the SFK people and what I said was a fact, that SFK did undertake a very large proportion of the research wOrk necesary in proving Thora- zine valuable in treating some forms of mental illness, and did under- take to obtain the right to manufacture them and in prosecuting their approval in this country. PAGENO="0182" 1732 COMPETITIVE PROBLEMS IN THE DRUG TNDUSTRY Senator NELSON. They did more of manufacturing in this respect than the Canadian company; did they not? Mr. CUTLER. I do not believe-the Canadian company did not even begin until Smith Kline & French had done all the work and the product was proven a success. Mr. GORDON. Here is testimony from Dr. Hans Lehman, on page 9024, part 16, of the Kefauver hearings. Now, on the top of page 9025, Dr. Lehman says that Rhone-Poulenc came up with such a drug and that was Thorazine, which he said was used in anesthesia. A year or two later, the same drug was used in the treatment of mentally ill people, and doctors were very excited because it had these drug-pro- ducing properties and they wanted to see what it would do in people who needed to be sedated. Then Senator Kefauver asked him if he and another physician at Verdun Hospital were the first to put the drug to this use on the Continent, to which Dr. Lehman answered, "That is right." Then Senator Kefauver asked how much later it was when son~e- thing was done about it in the United States. Dr. Lehman answered that they had been working on it about the same time here, but not as systematically, and not in this way. I have been asked, he said, why we had to jump from the United States to Canada and why the first English-speaking article came from us and not from England, the people there being so much closer to France, and so on. It seems to be pretty clear from this testimony before the Kefauver committee, that it was in Canada that the drug was used for the first time. Mr. CUTLER. We are debating a. factual issue. Let us submit to you Smith Kline & French's reports on the work it has done on this drug.' Senator NELSON. All right. Now, how much is charged by Rhone-Poulenc in Paris for that drug? Dr. WHITNEY. I do not know that fact. Senator NELSON. We have a statistic on that. We think it is about $10. If it is, would you explain the economics of that situation, where the originating company, which obviously has the patent, charges less by far-well, I guess we do not have the figures. If you will submit what the other company did, I would like to get to this later, because it is a most puzzling situation to me, in any event. You do not think it has anything to do with the lack of competition in the United States; do you? Dr. WHITNEY. You mean there is more competition in France? Senator NELSON. It is $10.80 a thousand in France, where Rhone- Poulenc got the patent., and it is $32 a thousand here, and $2.60 a thousand in Canada. Dr. WHITNEY. The average Frenchman's income is about a third or less than a third of the average American's income. Mr. GORDON. You say, then, there should be no relationship to costs? Senator NELSON. Excuse me. I have st.ated this incorrectly; I am sorry. The price to the druggist. in Paris for one hundred 25-milligram tablets is $1.08. That would be $10.80 a thousand. That is in Paris. 1 See Appendix III beginning at p. 2129, infra. PAGENO="0183" COMPETITIVE PROBLEMS IN THE DRUG INDUSTRY 1733 The price in the United States is $57.58 a thousand to the druggist. Now, as an economist, could you give me any idea why in this coun- try, where the research was not done-Rhone-Poulenc obviously did most of the work and got the patent-it sells for $57.58 a thousand, and in Paris, it is $10.80 a thousand? Dr. WHITNEY. If all goods were priced according to cost, you could throw my paper in the wastebasket. The law of attraction of capital by profits would not be allowed to apply. You have to direct capital where to go. Prices are not normally set according to costs. That would be a dif- ferent kind of economic system, one directed by the Government, perhaps. Secondly, as to prices between this country and France, most prod- ucts are different. You do not find the same price here as abroad. Their income is much lower. Their personal income, I would imagine, is one- third as much as ours, or less, and the prices are all at a much lower level there. Senator NELSON. Well, we have cases, of course, where the drug is made in this country and shipped to Europe and sold cheaper in all European countries than it is here. That is not true of the American automobile overseas. Dr. WhITNEY. Would you want drug companies to stop exporting? Senator NELSON. I think you put your finger on it precisely, and we have raised this question before: There is competition in Europe so t;hey have to meet the competition. They are licensed exclusively in this country and they do not have to meet any competition, so they can charge an exorbitant, a fantastically exorbitant price. Is that true or is it not? Dr. WHITNEY. If you reduce the price the way you are suggesting, to the French level- Senator NELSON. I did not say to the French level. I just said- Dr. WHITNEY. Something in between. There is some figure in be- tween that would put them out of business completely. They alone d'~ something like $20 million in research a year, maybe more. Then there is some higher price level that would cause them to continue their work, but gradually diversify out of drugs. Senator NELSON. When the patent runs out and everybody can make it at that price, it will come down; will it not? Dr. WHITNEY. I think so, and that is: one of the real risks that a company in. this industry has to take. Senator NELSON. Where is the risk of a company that is licensed other than their investment in distribution and advertising? They did not do the research. Dr. WHITNEY. If the product is replaced, there is a risk. The clin- ical testing they do is very great, too. Senator NELSON. Everybody does that. Mr. CUTLER. Smith Kline & French did it in this country. Rhone- Poulenc did not conduct the clinical testing in this country. Senator NELSON. You do not see anything wrong at all in charging so much in this country, $10.80 in France, and up in Canada, selling at $2.60 to the Government, not to the druggist? Mr. CUTLER. I do not draw the conclusions you draw from it. Senator NELSON. I am trying to find some explanation. PAGENO="0184" 1734 COMPETITIVE PROBLEMS IN THE DRUG INDUSTRY Dr. WHITNEY. If they are going to `sell in France at all, they have to meet the local price. Senator ~ELSON. So there is no local competition here, and what you are saying is that because there is no competition in this country, they can charge what the traffic will bear. Dr. WhITNEY. They are doing the same thing in France, selling in France for what they can get in France. You h'ave not told me Smith Kline & French sells in France? Senator NELSON. Yes; in France. They cannot do it unless they meet the competition. Dr. WHITNEY. It is the same thing in France. Rhone-Poulenc has the monopoly in France. Senator NELSON. Has it? Dr. WHITNEY. Has `their patent expired? Senator NELSON. I do not know whether it has, whether it is cross-licensed or not. I can give you a couple of examples of drugs manufactured in this country and sold at one-third the price in France and in Geneva after being shipped over there. What is your explanation for that? Dr. WHITNEY. They are meeting the local prices. Otherwise they could not export. Senator NELSON. And in this country, there is no competition. Dr. WHITNEY. No, I am not admitting that. On a patent drug, there is only one product. Senator NELSON. Right. You are correct. I just would like to have you honestly confess, as an honest professor, that the reason for it is a monopoly in this country which allows a com- pany to charge any price it pleases, and it will charge what the traffic ~will bear. And they will ship a product to a.nother country where there `is competition which they have to meet. So, the record is clear, they will manufacture drugs, `and charge an exorbitant price here and then go overseas and meet the competition. When I asked one of the manufacturers about this he said, "Well, the standard of living is lower in Geneva; that is why the price is lower." I said, "All right, how about Mexico City, where you charge three times as much as you do in Geneva, Switzerland?" He said. "I could not answer that, I would have t'o ask the comp- troller of the company." What you have here is a price structure that defies explanation unless you are going to confess that they are monopoly situations, `where the firms are going to charge whatever they can get. Dr. WHITNEY. As a lawyer and as `a Senator, you know that the ~patent law intends to give companies 17-year monopolies and they "do it in order to give them enough profit to induce them to find further discoveries. You are stating what is expected. I do not think you should call it exorbitant. Mr. CUTLER. Mr. Chairman, on your general proposition, some of the companies publish the proportion of their earnings that come from their foreign business and the proportion that comes from the domestic business. Generally speaking, their rate of return on investment on `their foreign business is more or less the same as their rate of return ~on their domestic business. If they were by your standards charging PAGENO="0185" COMPETITIVE PROBLEMS IN THE DRTJG INDUSTRY 1735 an exorbitant price at home and a competitive price abroad, that would not be so. Senator NELSON. That does not really explain how they will sell cheaper in Geneva, Paris, Bonn, and Rome, and then turn around and charge three times as much in Mexico City. Mr. CUTLER. But they have about the same rate of return on their foreign business as on their domestic business. Senator NELSON. That does not explainwhat Mr. Conzen could not explain that day. I think the fact is that there was competition there and none in Mexico City. Dr. WHITNEY. You expect competition when a drug is patented in this country? Senator NELSON. I am not talking about that at all. Dr. WHITNEY. Oh, you are not. All right. Senator NELSON. I am just saying I think the American public ought to know that companies who manufacture drugs here can sell them overseas at a third of the price they sell at to American people. Dr. WHITNEY. But if the American public then reacts by demanding that the price be cut by two-thirds here, that will almost destroy this extremely valuable industry. Senator NELSON. Well, go ahead. Dr. WHITNEY. We are talking about generic drug prices. I will skip a bit. At the bottom of page 6, I say we do not have published reports on how much generic firms are making. I say I am willing to assume that the risk-taking companies earn higher profits than the limited-service firms and will do so as long as, though no longer than, their research is productive. One reason I am willing to say it is that it is the very function of profit that my paper is about. What does it mean, after all, to say that a price is "too high"? Con-' sumers make this complaint when they are paying for necessities- milk at 30 cents a quart, meat at a dollar a pound, or prescriptions at~ $4 or $5. But millions of television sets are bought at $200 or so and~ cars at 10 or 15 times as much. Buyers must think these prices are~ "reasonable" if the product is worth more to them than some other use of their money, whether for a substitute, an unrelated article, or an addition to saving. If a $5 prescription, or 6 of them, will keep a patient from losing a couple of days' pay or spending a night in a hospital, the price is reasonable. It does not cease to be so because those whose venture capital developed the drug keep for dividends and re- investment a 20th of the retail price, which is about average, or a 10th, or whatever is necessary to bring this drug or a later and better one into existence. Nor does a theatre ticket, costing what a prescription would or sometimes more, become unreasonable if, instead of giving one a seat at a profitless "flop," it is for a "hit" whose backers are doubling their money in a couple of years. Mr. GROSSMAN. Dr. Whitney, may I stop you there? Dr. WHITNEY. Certainly. Mr. GROSSMAN. I want to read the sentence over again: If a $5 prescHption, or 6 of them, will keep a patient from losing a couple of days' pay or spending a night in a hospital, the price is reasonable. `What is the basis for your economic thinking here? PAGENO="0186" 1736 COMPETITIVE PROBLEMS IN THE DRUG Th~DUSTRY Dr. WHITNEY. I would not be here today if I were not able to get these antibiotics at $4.50 a couple of days ago, which made it possible for me to come down here. Mr. GROSSMAN. You say you think these company prices are reason- able. You did not have any choice. You go in and pay a couple of dol-. lars, a.nd that is it. I just cannot believe that any serious man would come here and say that to this committee. Dr. WHITNEY. Try to open your mind and listen to my case. I had, 2 or 3 days ago, the question of bowing out to keep from testifying. I would normally let my virus run its course. But in view of the pres- ence of antibiotics, I made a deliberate choice to spend $4.50 to come down here. What is so unreasonable about that? I made a deliberate choice. Mr. GROSSMAN. Frankly, I think your example is not the norm. I `think most people do not have the choice. They just go in and get a prescription; the doctor makes the choice for them. You said if a $5 prescription or six of them will keep a patient from losing a day's pay, the price is reasonable. You are just throwing this open, right. In other words, there is no limitation as far as you are concerned; the price is always reasonable? Dr. WHITNEY. Oh, no. Mr. GROSSMAN. Oh, no? Dr. WHITNEY. If a person were given $100 worth of antibiotics- Mr. GROSSMAN. How is he to know? Dr. WHITNEY. If he were forced to pay $100 in order to save himself a day's work and a day's work would pay him $20 a day, it would be better to lose the day's work. You measure what you get. Mr. GROSSMAN. I just do not think the buyer has that much choice. I think you are giving him a better position in the marketplace than he really has. Dr. WHITNEY. Do you think the buyer is being overcharged for his product if it keeps him out `of a one-night stay in a hospital? Mr. GROsSMAN. I think if it is reasonably produced and a reasonable profit could be elicited for $1 and the product costs $2, he is being overcharged. Dr. WHITNEY. What is your standard for a reasonable profit? Mr. GROSSMAN. I do not think you and I agree on any social respon- sibility the manufacturer has. I think you are more concerned with the investor. I underst.and your argument on- Dr. WHITNEY. I am afraid you do not understand it. Mr. GROSSMAN. I understand your argument is the consumer is bene- fited by the high prices because the industry has more money for re- search `and all this. Dr. WHITNEY. You see it. Mr. GROSSMAN. I see what you are saying, but I think we have to bear a lot to go as far as you do. I do not think we have to bear it. I think ~the statement here takes us way beyond anything I could reasonably `accept. Dr. WHITNEY. It is mainly true. Mr. G-Ross~IAN. in other words, I am just going to set here and wait for the industry to decide what they are going to charge. Dr.. WHITNEY. Shall I go ahead, or shall I go on arguing about this? PAGENO="0187" COMPETITIVE PROBLEMS IN THE DRUG INDUSTRY 1737 Mr. GROSSMAN. I would like also to ask you on this `other point. This 20-year business is another argument `of yours which I cannot, frankly, follow. Twenty years are too `few, you say. I asked that earlier this morning. I think you remember. Dr. WHITNEY. Yes; first I skipped that this afternoon but then I mentioned the railroads, which were considered a prime investment for three decades. Mr. GROSSMAN. Why do these companies not build up reserves, then? Dr. WHITNEY. What do yon mean by building up reserves? I dis- cussed that in terms of future dividends- Mr. GROSSMAN. I am just wondering, in Searle's case, if Searle has only one product, this oral contraceptive, the consumer seems to be keeping them in a profitable business by supporting this one drug. When you say they are struggling to find- Dr. WHITNEY. I th'ink that `was your word, sir. Mr. GROSSMAN. "Struggling" is `my word? I think not. Dr. WHITNEY. Is it not? Well, go ahead. Mr. GROSSMAN. All I am trying to say is I do not understa'nd why the consumer `has to pay because these companies are not diversifying. Dr. WHITNEY. Diversifying what? Mr. GROSSMAN. Their products. You are saying there is a risk be- cause Searle has an oral contraceptive, and if we find something wron'g, their whole pricing, whole profit structure is going to go tumbling down. I am saying that since 1950, we have facts to show that probably 10 to 15 years before the oral contraceptive was developed, th'ere was 30 percent profit in 1950. You can `look down your `own chart. The profits-they are never below 23 percent. Mr. CUTLER. `Mr. Grossman, you could probably show an even higher rate of return for Polaroid. Would you say that made Polaroid excessive? Mr. GROSSMAN. No~; I think here we go back to the social responsibil- ity t;hat `Mr. Squibb talked about `last week. Mr. CUTLER. Let us do th'at. Mr. GROSSMAN. I say it is relative. I am not saying it is complete; it is a relative thing. Mr. `CUTLER. I understand, but your assumption seems to be that a manufacturer should compute h'is costs,~ then he should add a reason- able profit that he gets out of the sky, perhaps from a law or something else, and add that on. That `is not the way the economic system works. T'hat is `what Profes- sor Whitney is trying to say to you. Five years ago, in the Kefauver committee, Professor Rostow, now the Under Secretary of State, used the example of a farmer who was able to produce tomatoes at half the price of `his neighbors because he was more efficient or because his land was richer, or something else. That did not mean that he would go to the market and s~ll hi's `tomatoes at the same profit margin that every'body else did; he would sell them at what they would bring in the market against the competition of other tomatoes. If you say that just shows a lack of social responsibil- it.y, we disagree with you. Mr. GROSSMAN. I would only say we are talking about a different PAGENO="0188" 1738 COMPETITIVE PROBLEMS IN THE DRUG INDUSTRY market. We are talking about somebody who goes out and is told by his doctor to buy a certain drug and he buys it. Mr. CUTLER. Every one of these drugs competes with other drugs, different patented drugs, other generic varieties of the same drug in many cases. There is no drug that has a total monopoly on the market. Everything is competing with a variety of other drugs or, as Dr. Whit- ney said, with foregoing the use of any drug at all. There are controls on prices created by competition. We are all in favor of having as much competition in this industry as we can. Mr. GROSSMAN. I want to a.sk you one last question. Does the eco- nomic theory that you have been talking about justify the price dif- ferentials between what the companies are charging to the hospitals a.nd what the companies are charging to the retail druggists? I am not talking about foreign sales. I want to know how you, as an economist, can explain that and whether perhaps we should do some- thing about the way the Robinson-Patman Act applies to these com- panies. Dr. WHITNEY. The Robinson-Patman Act, as I understand it, does not apply to them. Mr. GROSSMAN. That is right. Do you think it should? Dr. WHITNEY. I do not know whether the price you are suggesting should be higher, if that is your implication; that is what would happen. Mr. GROSSMAN. The prices would come up? Dr. WHITNEY. I assume they would have to come up. Mr. GROSSMAN. Do you think the companies are making profits on their sales to hospitals? Dr. WHITNEY. I think they must be making some. I assume they are covering just the bare incrementat costs, selling at such low figures. Mr. GROSSMAN. But you think they are making a profit, then? Dr. WHITNEY. They may be making `a sma:ll profit covering `their out-of-pocket costs, and perhaps a little contribution to overhead and profit. I would not know that. But beyond making their average profit, I would not know that. But beyond making their average `profit, if they sold to pharmacies at the prices they sell to hospitals, t'hey would all be out of business. You recognize that, do you not? Mr. GROSSMAN. Can you justify the difference in prices? Mr. CUTLER. Would you explain what you `mean by justification? Mr. GROSSMAN. Let us talk about buyer. The buyer must think these prices are reasonable-that is what you said. I go to a `hospital as an outpatient and I get something for $2. I go to a druggist `and I pay $17. Do I think `that that is reasonable? `Mr. CUTLER. The hospital charges you $1 for an aspirin; they pay a lot `less. Mr. GROSSMAN. Using the example that I gave, do y'ou think that that is reasonable? In other words, am I to interpret it-you say the buyer. You are the `buyer and I am the buyer. Would I think I have gotten a good deal if I went to the retail druggist and paid $17? Mr. CUTLER. Let us come to this again, if we can, because I think you are askking more a legal question than an economic question. If by "reasonable" or "justifiable," you mean does it meet the Robinson- Patman standard of differences in the cost of supplying one customer PAGENO="0189" COMPETITIVE PROBLEMS IN THE DRUG INDUSTRY 1739 as compared to the cost of supplying the other, assuming both are op- erating in the same market, the answer is probably no. Probably the companies would have to charge substantially higher prices to hos- pitals and to the Government, and perhaps a little bit lower prices to pharmacies in order to average out if the Robinson-Patman Act ap- plied to sales to hospitals and to the Government. If Congress thinks that is what ought to be done, let the Congress do. it and the companies will comply with that. I assume no one here is suggesting we should raise the price to hospitals and to the Government. The implication seems to be that we ought to bring down the pharmacy prices to hospital and Govern- ment level. If we did that, these figures you have before you con- clusively demonstrate that for companies earning 18 percent, if they reduce their prices by two-thirds on 70 percent of their business, they would be out of business. There is no answer for that that I know of. Senator NELSON. Go ahead. Dr. WHITNEY. I first want to make one personal statement about Mr. Grossman's first having said that I have shown all the way through this interest in the investor. I am not an investor. I am not such a quixotic person as to want to help the investor in drugs. I have never been one of them. My first interest is in the health of the public. My interest in this industry is as a tool to that end. 3. Although a case can be made for various special provisions for low-income persons with heavy outlays on drugs, no case can be made for helping them by reducing the prices to everyone. Milk is not. sold at a cut price to all so that low-income groups can afford. moie of this healthy food. Senator NELSON. Just as an aside, I might say that in the school lunch program, it is sold at a low price to everybody, for 3 cents a glass. Dr. WHITNEY. To make that relevant, to all schoolchildren. You would not want to put that to the whole population, I take it? Senator NELSON. No, just to schoolchildren. I just wanted to point that out. Dr. WHITNEY. How many dollars and cents would be saved by consumers if the industry's earnings could be brought down to average without damaging production incentives,? I have seen nothing specific on this from industry critics. Let us look at profit figures for 1966. The FTC-SEC reports earnings before income taxes on stockholders' equity for all manu- facturing as 61 percent of those for drugs. The drug profit margin- a.lso before taxes but on sales, not equity-was 19.7 cents per dollar. Sixty-one percent of this would mean a return on equity equal to that of all manufacturing, but the price would be brought down only 7.7 of the 19.7 cents. Now drug manufacturers received perhaps $1.4 billion of the $3.05 billion in prescriptions dispensed by community pharmacies. Reducing this $1.4 billion by the 7.7 percent would bring revenues down $108 million. This is 10.8 cents in $3.05, or 12.4 cents in a typical $3.50 prescription. Some critics lightly assume that the pharmacist will reduce his selling price by the same percentage that his acquisition cost falls. Certainly under the professional fee system this does not happen. This is not all. First, someone will have to make up the corporate income tax payments lost and the money now being reinvested by PAGENO="0190" 1740 CO~ETITIVE PROBLEMS IN THE DRUG INDUSTRY the industry out of its earnings, unless its expansion is to cease. At 1966 ratios, the 7.3 cents remaining drop to 4 cents-that is, a 4-cent reduction in a $3.50 prescription. This does not allow for administra- tive costs. All right, one might say, why not save the 1 cent, or the 4 cents~ or whatever it is? Certainly, if it costs nothing-but is it a good gamble to imperil industry incentives for so little? The critics seize on selected examples of drugs with the greatest price differentials. But, no economist and perhaps no accountant thinks that every product sold by a firm, whether department store or manu- facturer, should carry the sa:me profit margin. Certain trends are alarming. New chemical entities marketed per $100 million of B. & D. expenditures, for 1959 throi~gh 19G6, were 32,. 22, 21, 11, 6, 6, 7, and 3, respectively. They may rise again to 6 or 8 as the industry catches up with the new requirements of the Food and Drug Administration, but more than 6 or 8 are needed. There may be no real recovery if the profits from research are threatened. We all know how fast medical care costs are rising; but prescrip- tion drugs were only 9.8 percent c~f total medical care expenditures in 1966, down from 11.7 percent in 1957. In fact, drugs through their therapeutic effects have been alleviating the impact of other costs which are rising rapidly. The American economy leads the world. So does the American drug industry, as recognized by foreign medical auth&rities and as regis- tered very practically in an annual 12 percent increase of `foreign sales. It was not the leader until its profit opportunities became' evident after World War II. The same economic principles account for the success of both the economy and the industry. As new and better drugs are discovered through profit-seeking re- search, competitive pressures are inevitably brought against existing drugs. None of us wants to disturb this process. More important than the price of existing drugs, many of them declining or gradually dis- appearing, is the health of the future, and this will depend on the new drugs that may be, indeed must be, found. Senator NELSON. Thank you, Professor Whitney. I appreciate your taking the time to come here. Dr. WHITNEY. Thank you for staying so late, Senator and members of the staff. I appreciate that. Senator NELSON. Thank you, Mr. Cutler. We will now adjourn, subject to the call of the Chair. (The complete prepared statement of Dr. Whitney follows:) STATEMENT OF Sniox N. WHITNEY, PH. D., PROFESSOR OF EcoNoMIcs, NEW YORK UNIVERSITY Mr. Chairman and members of the subcommittee, I am a professor of economics at New York University and consultant to the U.S. Economics Corporation, a firm of business analysts retained by the Pharmaceutical Manufacturers Associ- ation to make studies in this industry. I am appearing, therefore, on behalf of the PMA. I shall discuss the allegations you have heard that `branded ethical drug prices a'nd manufacturing profits are "too high." More than one distinguished medical witness making these charges, which are plainly economic ones, admitted lack of training in economics. Industry Witnesses have anticipated me in part, hut I shall speak as an academic economist who applies the established principles of this science to the problems before you. PAGENO="0191" COMPETITIVE PROBLEMS IN THE DRUG INDUSTRY 1741 Economics begins with the truths that goods are not produced without work and progress does not occur without saving and coustructive investment. To achieve these, experience since the end o~ feudalism, and recently even in Communist countries, points to the desire to better oneself and one's family as the most effective single motive. Reliance on self-interest also minimizes corn- pulsion, conscription and penalties. Although this incentive is clearly important to labor, people being eager to go wherever their wages or salaries (after considering other aspects of the job) will be highest, it is the application to capital that concerns us here. Self-interest tells people to invest where the outlook for profits (after taking account of risks) is best; and this in turn is normally where consumer demand is greatest. Adam Smith spoke of "an invisible hand" by which the search for profit leads to the general good, and also wisely said that we expect our dinners from the self- interest, not the benevolence, of butchers and bakers.1 Our standard of living is proof that, on the whole, the process has worked. Countries which condemn the profit motive are glad to draw from us some of the surpluses which `this moti- vation produces. How does this apply to the questions before the Subcommittee? Quite simply,. drugs are too important to the public welfare for us to withdraw or dampen the incentives which have created and multiplied all our goods. Any university teacher of economics denying the contrLbution of profit in attracing capital- and `this might happen, since a reputation for originality is prized in my pro- fession-I hope you will ask two questions: First, is he indeed deliberately re- jecting the central principles of the science as it is now known? and second, what alternative system does he recommend for inducing owners of capital to put it where consumers want it, and why would this alternative work better? Several persons interested in the present inquiry concerning the pharmaceuti-. cal industry have granted that this industry should be allowed a normal or average profit. But such a profit will `not attract investment where it is most needed. Nor will it achieve the second function `of profits-stimulating cost re- duction and efficiency. These essential jobs are done by differences, not uni- formity, in earnings. How has this worked in practice? From 1950, when analyses of the drug industry were first reported separately in a leading financial publication through 1965 the rate of return on net worth for "all manufacturing" averaged 65 per- cent of that for "drugs and medicines." 2 As a~ result, $3 of additional stock- holders' money per dollar of net worth in 1950 went into drug manufacturing for every $1 into all manufacturing, and about 25 cents into meat packing, the least profitable major manufacturing industry reported. Comparing drugs with meat packing, 314~ times the rate of earnings brought 11 or 12 times the rate of investment.8 The free market process-consumer demand, profits, attraction of capital-has worked in drugs, brilliantly. 1 "By preferring the support of domestic to that of foreign industry, he intends only his own security; and by directing that industry in such a manner as its produce may be of the greatest value, he intends only his own gain, and he Is in this, as in many other cases, led by an invisible hand to promote an end which was no part of his intention." The Wealth of Nations, Modern Library ed., p. 423. "It is not from the benevolence of the. butcher, the brewer, or the baker, that we expect our dinner, but from their regard to their own interest." Ibid., p. 14. 2 These are based on arithmetic averages of the annual rates of return given in the First National City Bank of New York, Monthly Economic Letter, April issues (even- numbered years being used, since each issue covers two years). For 1966 (not included because the latest net worth figure is for January 1, 1966) the return for all manufacturing was 67.1 percent of that for drugs. The rate of return of "all manufacturing corporations, except newspapers" in 1966 was 61 percent of that of "drugs" in the Federal Trade. Commission/Securities and Exchange Commission Quarterly Financial Report for Manufac- turing Corporations. This source was not used for two reasons: it did not separate drugs prior to the 2d quarter of 1956; and it includes along with the larger drug companies many smaller ones which do little or no research. From January 1, 1950, to January 1, 1966, net worth of the drug companies included in the First National City Bank list increased by 640 percent; that of all manufacturing by 208 percent; and that of meat packing by 56 percent. Part of the increase was through companies growing large enough to publish reports and be included for the first time: drug companies increased from 19 to 39, the total of manufacturing companies from 1763 to 2279, and meat packers from 19 to 27. The additional companies are usually smaller- thus the April 1962 Monthly Economic Letter added 7 drug companies whose average net worth for January 1, 1960, was only $15.5 million as compared with $57.5 million for. the 29 companies previously included. The expansion of small firms to the point where they issue public reports, the entry of new firms, capital expansion through issue of securities, cud reinvestment of earnings are all aspects of the inflow of capital. Reinvestment is the principal source. Without the profits of the drug industry there would have been no money to reinvest in expansion; had its profit outlook not continued good, earnings would prob- ably have been used for diversification away from drugs. PAGENO="0192" ~1742 COMPETITIVE PROBLEMS IN THE DRUG INDUSTRY If legislative or regulatory action should force drug profits to the average level, one of two things will happen. Either new capital will avoid an area with the continuing risks of this industry but no chance of a corresponding profit, and managements will diversify out of ethical drugs and into products of less importance where profits can be earned without drawing public or congressional resentment; or human nature will be `found to have changed and people to be employing their savings with the primary purpose of helping their fellow men. I would expect the first. What have `the drug industry's reinvested profits accomplished? First let me cite two statistical series. The public demand for drugs has been met with a rapidly rising physical volume of production. Consumer purchases of ethical plus proprietary drugs and sundries-an increasing proportion, now over 60 percent, being ethical drugs-expanded by 5.5 percent annually from the first postwar year, 1946, to 1966, as compared with only 3.7 percent for all other consumer expenditures.4 The second series, research and development expenditures by the industry, advanced from $50 million in 1951 to $416 million in 1966-faster than either profits or net worth.5 Total R. & D. for 1957-1966 seems to have beeu greater than ~lividends paid.6 In 1965, according to `the National Science Foundation, 95 per- cent of this R. & D. was financed privately, as against only 45 percent for all of industry.7 Many hundreds of new drugs, as documented by earlier PMA witnesses, re- 5sulted from this profit-motivated research. Wrhat they have done for health need not be detailed to this informed body. As an economist, I shall merely mention ±hat the monetary value of working time saved through improvement of physical and mental health and lengthening of life is probably in the billions. We imperil further gains if we strike at so-called "exce~" profits which, in fact, have furnished the driving force. For a mere 6 cents in dividends per dollar of sales of all types of products made by companies classified in this industry,8 or less than a cent and a half in dividends for each dollar of total medical care costs,° the industry has done its essential job. This is not primarily one of basic pharmaceutical and biological research-though it happens to lead all industries in percentage of B. & D. going to basic research `°-but one of discovery and bring- ing into use of new medicines. Eliminate 2 of the 6 pennies in dividends-make them average for all manufacturing-and investor enthusiasm will decline to average also. Does anyone really expect enough psychological satisfaction from ensuring that this one industry earns only average returns to warrant endanger- ing future investment in the search for cures of our dread diseases? It is implicit in my argument that drug prices are not "too high." I shall review briefly what seem to be the three most common reasons given by those who con- tend that they are: the size of profits, the contrast `between some branded and `unbranded drug prices, and inability of low-income groups with chronic illnesses to pay. 1. To call higher than average profits a proof of excessive prices overlooks the economic functions of profits as I have already detailed. It forgets also that in a free market earnings tend to vary with risks. When privately financed B. & D. Derived from US. Department of Commerce, The National Income and Product Ac- counts of the United States, 1929-1965: A Supplement to the Survey of Current Busi- ness, pp. 48-49. The figures are dollar estimates translated into terms of quantity by adjusting for price changes, The percentage increases are, of course, compounded annually. The 5.5 percent would undoubtedly be higher If the expanding distribution of drugs through hospitals and government purchases were included. R. & D. from annual PMA surveys. The Increase was 732 percent, compared with 627 percent for net earnings and 564 percent for net worth in the First National City Bank sample. 6 The PMA R. & D. totals for these ten years amounted to $2,530 million. Drug industry dividends of $2,660 million were reported in FTC-SEC ip. cit. A substantial part of these ~ilvidends, however, may be assumed to have been earned in proprietary drugs, chemicals and other products-sales of which In 1966, according to the company reports to PMA, were about a third of the total sales of these products plus drugs for the FTC-SEC sample. PMA Report, Pharmaceutical Industry Research and Development Activity, 1966-1967, `p. 4 and National Science Foundation, Basic Research, Applied Research, and Development in Industry, 1965, p. -20. 8 Sales in 1966 -were $7,281 million; dividends, $426 million. FTC-SEC, op. cit. Private medical care expenditures (exclusive, therefore, of those paid by government) -were estimated at `$31,250 million in 1966. U.S. Department of Commerce, Survey of Current Business, July 1967, p. 23. 10 National Science ~~undatIon, op. cit., p. 78. PAGENO="0193" COMPETITIVE PROBLEMS IN THE DRUG ]NDUSTRY 1743 is several times that of other industries in relation to sales," and when drugs become obsolete so fast that over half the 1966 prescriptions were written for drugs not known in 1956, there is plainly a high built-in risk factor. Any search for the new is more hazardous than production of the old.. According to economic theory, a short to medium term higher than average profit is. understandable, but entry of new capital will eventually bring it `to a long-term level which is commensurate with risk. In this industry, capital has entered. Rising consumer demand, stimulated by the industry whenever it de- velops a new and effective drug, and the inherent risks of research which have prevented even more money from entering, are sustaining its earnings. Twenty years are too few, in many cases, to deduce absence of risk from continuing profits: look at what happened to the railroads or, for that matter, meat packing, sugar and textiles. 2. Prices of certain drugs of one manufacturer may be higher than those of another producer which are at least theoretically chemically equivalent. Prices could hardly be the same when one company must cover the costs, and be rewarded for taking the risks, of research, testing, quality control, and original distribu- tion, and when it makes all dosage forms available where and when needed; whereas the maker of the other drug may only `manufacture, and then,, probably only the most profitable form. Do `the lower prices charged by generic-name manu- facturers reflect these lower risks and costs, or are the industry critics .right in implying that they reflect willingness to operate on lower profit margins? Until a representative sample of such firms publishes its rate of return on sales or investment we `shall not know the answer.12 In the meantime, I am quite willing to assume that the risk-taking companies earn; higher profits than the limited- service firms, and will do so as long as, though `no longer than, their research is productive. What does it mean, after all, to say that a price is "too high"? Consumers make this complaint when they are paying for necessities-milk at 30 cents a quart, meat at a dollar a pound, or prescriptions at $4 or $5. But millions of television sets are bought at $200 or so and cars at 10' or 15 times as much. Buyers must think these prices are "reasonable" if the product lS; worth more to them than some other use of their money, whether for a substitute, an unrelated article, or an addition to saving. If a $5 prescription, or 6 of them, will keep a patient from losing a couple of days' pay or spending a night in a hospital, the price is reason- able. It does not cease to be so because those whose venture capital developed the drug keep for dividends and reinvestment a twentieth of the retail price which is more average, `or a tenth, or whatever is necessary to bring this drug or a later and better one into existence. Nor does a theatre ticket, costing what a prescription would or sometimes more, become unreasonable if, instead of giving one a seat at a profitless "flop," it is for a "hit" whose backers are doubling their money in a couple of years. 3. Although a case can be made for various special provi'sions for low income persons with heavy outlays on drugs, no case can be made for helping them by reducing the prices to every one. Milk is not sold at a cut price to all so that low income groups can afford more of this healthy food. How many dollars and cents would be saved by consumers if the industry's earnings could be brought down to average without damaging production in- centives? I have seen nothing specific on this from industry critics-certainly nothing in the majority repori of Senator Kefauver's Subcommittee which de- manded drug price reductions in 1961.'~ You have heard of l'arge economies made by hospitals `through purchases of drugs by generic name. Were all purchasers to do the same, many research-based companies'would be put into serious straits. 1~ In 1966 PMA surveys show 10 9 percent of the domestic pharmaceutical sales dollar of research oriented firms was allocated to the research, discovery and development of pharmaceuticals and biologicals. The National Science Foundation shows total industry R. & D. privately financed, as only 2.0 percent of sales. Ibid., p. 69. 12 39 drug companies in the First National City Bank sample earned $711 million in 1966 on $3,377 million net worth (again, these are overall company figures, applying to all products), or 21.0 percent. The FTC-SEC sample, Including these and perhaps as many small firms, earned $787 million on $3,547 million, or 22.2 percent. The difference, $76 million on $170 million, is 44.7 percent. Surely the small firms were not so profitable- but, without knowing the make-up of the samples or carefully comparing the reporting forms, we can make no affirmation. ii Administered Prices: Drugs, Report of the Committee on the Judiciary, United States Senate, Made by Its Subcommittee on Antitrust ana Monopoly. Pursuant to S. Res. 52, 87th Cong., 1st Sess., Report No. 448. 81-280-65-pt. 5-13 PAGENO="0194" 1744 COMPETITIVE PROBLEMS IN THE DRUG INDUSTRY Let us look at profit figures for 1966. The FTC-SEC reports earnings before income taxes on stockholders' equity for all manufacturing as 61 percent of those for drugs. The drug profit margin-also before taxes but on sales, not equity-was 19.7 cents per dollar. Sixty-one percent of this would mean a return on equity equal to that of all manufacturing, but the price would be brought down only 7.7 of the 19.7 cents. Now drug manufacturers received perhaps $1.4 billion of the $3.05 billion in prescriptions dispensed by community pharmacies.'4 Re- ducing this $1.4 billion by the 7.7 percent would bring revenues down $108 mil- lion. This is 10.8 cents in $3.05, or 12.4 cents in a typical $3.50 prescription. Some critics lightly assume that the pharmacist will reduce his selling price by the same percentage that his acquisition cost falls. Certainly under the professional fee system this does not happen. This is not all. First, some one will have to make up the corporate income tax payments lost as a result of declining profits of drug manufacturers. At present tax rates, the 12.4 cents the public is supposedly "saving" will be reduced by 5.1 cents. Second, someone will have to replace the money now being reinvested by the industry out of its earnings, unless its expansion is to cease. At 1966 ratios, the 7.3 cents remaining drop to 4 cents-that is, a 4 cent reduction in a $3.50 prescription. All right, one might say, why not save the 1 cent, or the 4 cents, or whatever it is? Certainly, if it costs nothing-but is it a good gamble to imperil industry incentives for so little? The critics seize on selected examples of drugs with the greatest price differen- tials. But, no economist and perhaps no accountant thinks that every product sold by a firm, whether department store or manufacturer, should carry the same profit margin. Instead of prices for each product which meet the market, such a system would create a jungle of prices reflecting the costs of each com- pany-if indeed costs could be determined. Certain trends are alarming. New chemical entities marketed per $100 million of R. & D. expenditures, for 1959 through 1966, were 32, 22, 21, 11, 6, 6, 7, and 3, respectively.'5 They may rise again to 6 or 8 as the industry catches up with the new requirements of the Food and Drug Administration, but more than 6 or 8 are needed. There may be no real recovery if the profits from research are threatened. We all know how fast medical care costs are rising; but prescription drugs were only 9.8 percent of total medical care expenditures in 1966, down from 11.7 percent in 1957.16 In fact, drugs through their therapeutic effects have been alleviating the impact of other costs which are rising rapidly. The American economy leads the world. So does the American drug industry, as recognized by foreign medical authorities and as registered very practically in au annual 12 percent increase of foreign sales.'7 It was not the leader until its profit opportunities became evident after World War II. The same economic prin ciples account for the success of both the economy and the industry. As new and bett.er drugs are discovered through profit-seeking research, com- petitive pressures are inevitably brought against existing drugs. None of us wants to disturb this process. More important than the price of existing drugs, many of them declining or gradually disappearing, is the health of the future, and this will depend on the new drugs that may be, indeed must be, found. (The charts referred to by Senator Nelson follow:) `5The $1.4 billion Is 46 percent of the $3.05 billion pharmacy ethical drug sales, as estimated by the ITS. Department of Commerce. `5Derlved from the de Haen surveys of new chemical entitles Introduced and the PMA estimates of drug R. & D. expenditures. 16 Derived from Department of Commerce estimates of consumer expenditures for medi- cal care and PMA estimates of expenditures for ethical drugs. 17 PMA data. The average annual gain from 1960 to 1966 was 12.4 percent. PAGENO="0195" DRUG PRICE COMPARISONS I .~ Item Highest city price Lowest city price ~ O lop price to Percent high- Percent trade New York Defense Veterans druggist est city over name price to ~ City Supply Administra- (Red Bunk) lowest city druggist over Agency tion 1968 price lowest insti- H tution price H L~i 341 357 367 o 816 230 ~.i 313 0) 340 Chlordiazepoxide Cl capsules (25 m~., 500's).. Des Moines, $45 Philadelphia, $18.50 $28.50 $21.75 $45.00 243.2 Meprobamato tablets (400 mg., 500 s) Des Moines, $31.20 Los Angeles, Philadelphia, 9. 45 $2. 30 2. 90 31, 20 330. 2 and San Jose, $9.50. Sulfisoxazde tablets (0.5 gm., 1000's) Des Moines, $25.30 Philadelphia, $6.90 8.35 8.50 25.30 366.7 Chleramphenicol capsules (250 mg., 100's)_ Indianapolis, $22.50 Philadelphia, $3.75 6.73 4.96 5.41 ~30.60 680 Ampicillin anhydrous or trihydrate capsules Grand Rapids (Mich.), $136.75~ Los Angeles, $59.45 59.34 136.75 230.4 (250 mg., 500's). Ampicillin oral suspension (250 mg./cc., 80/cc. Wichita and Corpus Christi, Albuquerque, $1.55 2. 70 3. 16 3. 23 4. 85 265. 8 bottle). $4.12. Nitrofurazono soluble dressing (1 lb.) Cincinnati, Corpus Christi, Nashville, $4.08 1.50 2.30 5.00 340.0 Milwaukee, Minneapolis, Cambridge, and Wichita, $5.10. Irivalent polio vaccine (10-dose package). Albuquerque, $6 Phoenix, $1.02 2.61 6.00 588.2 588 H lsoniazid tablets (100 mg., 100's) Milwaukee, $0.48 Corpus Christi, $0.07 . 1975 .26 .143 1.14 685.7 1,629 - Isoniazid tablets (50 mg 1000 s) Grand Rapids $2 60 Detroit Los Angeles and 79 1 30 5 18 329 1 666 ~j Milwaukee, $0.82. Prednisone tablets (5 mg., 1000's) Nashville, $18.00 Los Angeles, $4.35 4.58 4.52 4.10 170.00 413.8 4,146 ~ Diphenhydramine capsules (25 mg., 1000's)... Wichita and Winston-Salem, Philadelphia, $3.19 3.10 2.90 5.66 12.84 345.2 443 ~ $10.70. Diphenhydramine capsules (50 mg., 1000's) - Des Moines, $21.27 Los Angeles, $3.00 3. 25 5 5. 20 4. 39 19. 14 709. 0 709 ~ Reserpine tablets (0.1 mg., 1000's) Grand Rapids, $19.98 Newark, $1.35 .60 1.20 23.50 3,330 3,917 Reserpinetablets(0.25 mg. 5000's) Grand Rapids, $160 Chicago, $2.09 2.95 ~4.50 05.00 188.24 7,655.5 9,007 Dextroamphetamine tablets (5 mg., l000's) - Des Moines and Newark, $22.60 Los Angeles, $0.53 - 57 1. 90 1. 00 22. 60 4, 264. 2 4, 264 Phenazopyridine HCL (0.1 gm., 1000's) Erie, Indianapolis, and Winston- Phoenix, $4.13 4.80 °11.60 32.16 48.00 1,162.2 1,162 Salem, $48. 5 $1 per 1,000. 6 $0.19 per 100. I.'4 $0.58 per 100. 1 All prices are quoted from contracts let in 1967. 2 1967. 3 $0.52 per 100. $0.90 per 100. PAGENO="0196" 1746 CO~PETITWE PROBLEMS IN THE DRUG INDUSTRY PRICES OF CHLORPROMAZINE (THORAZINE) BY RHONE-POULENC'S LICENSEES 2 TO THEIR RESPECTIVE GOVERNMENTS, 1966-67 Price of U.S. licensee (Smith, Kline & French) to U.S. Defense Supply Agency Price of Canadian licensee (Bell-Craig) to Canada's Department of Veterans Affairs Smith, Kline & French's price to U.S. community pharmacist :25 mg. tablets $32.62 per thousand. 100 mg. tablets $48.22 per thousand. $2.60 per thousand.2 $7.25 per thousand.2 $57.58 per thousand. $92.64 per thousand. 1 Rhone-Poulenc is the discoverer and patent holder of thorazine. 2 Derived from prices for bottles of 500. (The report "Risk and Return in American Industry-An Econo- metric Analysis," submitted by Mr. Irving H. Plotkin follows:) RIsK AND RETURN IN AMERICAN INDUSTRY-AN ECONOMETRIC ANALYsIs (By Gordon B. Conrad and Irving H. Plotkin) This report is rendered upon the condition that it is not to be reproduced in whole or in part for advertising or other purposes without the special permission in writing of Arthur P. Little, Inc. I. SUMMARY A. PUBPOSE AND SCOPE This study was undertaken for the Pharmaceutical Manufacturers Association in order to examine possible broad economic expinnations for the levels of profit- ability of major pharmaceutical manufacturers. These explanations were sought because of the continuing criticism of the pharmaceutical industry in the press and by various members of Congress and Federal regulatory agencies directed at the supposedly unreasonably high prices and profits of this industry. The concept of the study was to relate levels of profitability among companies and industries to the amount of risk they have been subjected to over an extended period of time. In particular, our concept embraces science based industries, in- cluding the pharmaceutical industry, and examines their usually high risks which in part result from the high degree of technological innovativeness and obsolescence characterizing their operations. Our procedure involved first the development of a theoretical foundation for describing risk and its association with levels of profitability. Testing these theoretical constructs then reQuired the devising of quantitative measures of the risks and uncertainties experienced by companies or groups of companies PAGENO="0197" COMPETITIVE PROBLEMS IN THE DRUG INDUSTRY 1747 within specific industries. Although some of our analyses involved the stock market investor and how he might view risk in relation to profitability, our major concern and emphasis was on the individual company, its management, and the risk and return to assets invested in specific economic activities. In order to test the concept that the pharmaceutical "industry" or any other "industry" might have a characteristic and identifiable risk environment, we grouped com- panies by "industry" and conducted most of our analyses" on such industrial units. The data for these studies have been obtained from Standard and Poor's Gompustat Tape. The basic time period covered was 1950 through 1965. The Oompustate tape is essentially composed of audited and published financial infor- mation analyzed and adjusted by Standard and Poor's to insure comparability. Appendix B presents the rates of return definitions in terms of the Oompustat tape, while. Appendix C lists the industries and number of companies used in the analyses. B. CONCLUSIONS 1. The theoretical measure of risk developed for this study correlates signifi- cantly with various well recognized measures~ of profitability for the approxi- mately 780 companies grouped into 59 Industry's sectors which have been examined. Exhibits I and II illustrate in graphic form the statistical results obtained (each dot representing one industry sector). 2. The risk measure we developed appears to be a valid measure of the uncer- tainty that corporate managements ~face in deciding on the rates of return their investments must achieve, i.e. the profitability levels' necessary for the enterprise to survive and grow within the risk environment these managements perceive. 3. The general pattern of risk/return relationship shown in Exhibits 1 and 2 conforms with the expectations of broad'economic theory. 4. Individual industries, as grouped in this study, can be evaluated for their relative performance with respect to risk versus return, and can be judged on their conformity to the overall pattern of American industry. 5. The pharmaceutical industry fits welL within the overall pattern of risk! return relationships for American industry. While displaying a high level of profitability, the pharmaceutical industry also shows a high level of risk expecta- tions over the period 1950-1965. C. REPORT ORGANIZATION The report is organized along the following lines: In Section II we explain why questions of consumer price must be ultimately related to rates `of return, and further why questions of adequacy and excessive- ness of rates of return must be answered within the context of a risk environ- ment analysis. Section III together with Appendix A develops in words and mathematics a basic theory of risk and return. It also describes the types of statistical problems one is likely to' encounter when attempting to validate such a thiOTy. . PAGENO="0198" C) INDUSTRY RATE of RETURN (%) Industry Return (Net Income + Fixed Charges)/Total Capitalization Industry Risk Average Inter-com~asy Dispersion 10 20 30 00 SO EXHIBIT I MODEL I: INDUSTRY R1SK/RET~)RN PATTERN MEASURED AT BOOK VALUE NOTE: REPRESENTS IS)SITION OF PHARMACEUTICAL INDUSTRY 60 70 80 IP4bUSTRY RISK PAGENO="0199" INDUSTRY RATE ~ RETURN MtMt ~° ~ Risk ~° ~ ~u LBS 2W) 220 200 260 280 300 320 345 360 EXHIBIT II MODEL II: INDUSTRY RISK/RETURN PATTERN MEASURED AT MARKET VALUE INDUSTRY RISK NOTE: (~) REPRESENTS FOSITION OF PHARMACEUTICAL INDUSTRY 4 PAGENO="0200" 1750 COMPETITIVE PROBLEMS IN THE DRUG INDUSTRY Having defined our general measures of risk, Section IV defines the specific `book" and "market" value measures of return that we investigated. In this section we develop the concept of "Total Investable Funds." This concept facili- tates our making balance sheet item comparisons among the diverse industries encompassed in olir study. The results of testing of our basic models are sum- marized in Section V. A more detailed and technical description of the statistical analyses appears in Appendix D. II. MOTIVATION Economic theory has long maintained that a going concern must reward it~ Investors with a return commensurate with the risks inherent in the particular business undertaking. Although many elaborate and technical definitions have been offered for the term "risk," the underlying meaning is generally accepted to be "lack of predictability." While ruling in the Hope Natural Gas Case, 1944, the U.S. Supreme, Court stated this fact of economic life as follows: "From the investor or company point of view it is important: that there be enough revenue not only for operating expenses but also for capital costs of the business. These include service on the debt and dividends on the stock. By that standard the rate to the equity owner should be commensurate with the returns on investment in other enterprises having corresponding risk. That re- turn, moreover, should be sufficient to assure confidence of the financial in- tegrity of the enterprise, so as to maintain its credit to attract capital." In short, if an enterprise is to attract capital it must offer prospective investor, especially equity investors, an expected return high enough to compen- sate for any risks (lack of predictability) that might accompany this expectation. Until recently there have been two distinct forms of corporate endeavor. The first and most common type is the private corporation whose product prices are set in more or less competitive markets. The second type is the natural, regu- lated monopoly who Is given by government fiat exclusive geographical juris- diction, and whose product prices (or rates) are set by a regulatory body (agency). Examples of such organizations are telephone ~companies, light and power utilities, gas pipe lines, etc. A critical similarity between these two types of industrial organizations Is that both raise investment capital from the general public in the national capital markets. The question is, by what mechanism :does each type of organi- zation balance return and risk for its investors. The forces of supply and demand in the capital market allocate to each private enterprise an amount of capital that balances at the margin prospective risks and returns from the particular economic activity. It is the consumers' market place that determines the quantities and prices of the firm's products. In the case of regulated monopolies the regulatory agenlles seek to establish a rate schedule that over time will guarantee a return on invested . assets of some specified, say 6%, rate of return. In setting the price schedules for regu- lated industries the various agencies are required~ to consider the~ effect of these rates on the total return of the corporation. "A public utility is entitled to such rates as:will permit it to earn' a return on the value of the property which it employsfor~ the convenience of the public equal to that generally being made at the same time and in the same general part of the country on investments in other business undertakings which are attended by corresponding risks and uncertainties; but it has no constitutional right to profits such as are realized or anticipated in highly profitable enter- prises or speculative ventures." 1 However, today various government `bodies are attempting a ~tnique form of regulation. This regulation stems from the desire to control the final consumer prices of various goods and services.2 Therefore, like the regulatory agency, the government desires to set the market price of the product. However, unlike the regulatory agency, the total effect of ,a change In price on the earning abilities of the corporations is `usually not considered. Changes in price or rate schedules have immediate effects on the ~total revenue stream available to the corporation, and both short and long term effects on the corporation's ability to service its debt and pay dividends to its stockholders. It would be 1 U.S. `Supreme Court, Bluefield Waterwork8 Case (1923). 2Zome have called this type of regulation "Consumerism." PAGENO="0201" COMPETITIVE PROBLEMS IN THE DRUG INDUSTRY 1751 wrong to assume that one could materially affect the revenue streams of corpo- iations without affecting the rate of return a corporation is able or expected to earn Although final consumer price was the question at issue in the several court and regulatory cases mentioned above attention was directed toward rates of return in regulated industries. Questions of "adequacy," "excessiveness," and "equity" of final price were answered by examination of profit. Under assump- tions of profit maximizing behavior questions of adequacy, excessiveness, and even equity kre evaluated in terms of the rate of return and risk environment. In studying adequacy and excessiveness of rates of return and risks, the courts and agencies have generally found that the government is not free to change merely the rates of return of the industries whose prices it seeks to establish. Unless the government is willing to make certain guarantees of minimum returns to these new "semiregulated" industries, it will leave the risk environment unchanged, while usually lowering the rate o~f return by lower product prices. Such action would result in a marked reduction in the flow of capital to these corporations and, therefore, in a necessary curtailment of the normal flow of goods and services to the public. In summary then, any discussion of prices and exces:s'ivenes!s of prices in the pharmaceutical industry should be focused on the underlying issues of profit- ability of this industry, and even more importantly on the relation of this indus- try's profit level to the risks inherent in its operation. If prices are to be challenged or if suggestions are to be made by the Government for new price mechanism, the industry should be in a position to deal with such matters on grounds that are truly pertinent. What follows in this report is a theoretical and statistical development of risk/return relationships in American industry which can be used to place the entire issue of possible price regulation for drugs in the perspective of: (1) the Pharmaceutical Industry's position in our economy; and (2) the relationship of its profit level to its risk environment III. Paoroszn THEORY OF RISK AND RET[JRN3 Our objective in general terms has been to determine the relation~ship, if any, between -the levels of profitability and varying degrees of risk experienced in American industry. The first variable, rate -of profitability (or return) is rela- tively easy to measure conceptually. As will he explained in Section IV, we have used a number of book value and market value financial statistics to measure it. Problems of insuring inter-company accounting comparability were minimized by our use of the Compustat data, in whose preparation pains were taken to insure comparability. The concept of risk, bewever, is a more trotiblesome problem involving seman- tics. Risk is basically -a subjective phenomenon' `and not as susceptible to precise or direct measures What we have done therefore is to theorize that certain objectively measurable concepts are related-tO, and to some extent describe, risk. Our next step was, by statistical `techniques, to correlate rates of return with these objective risk measures.4 Two types of problems can arise First of all because of some logical error or assumption contrary to fact there may be no correspondence between the con cept of risk and -our objective measures. Our sole technique for dealing with this possibility has been to express as explicitiy as we can the steps in our logic and our assumptions so that they may be critiCally examined. * The second `type -of problem is that `we' can find `a relationship which turns out to be spurious i e some statistical fluke This type of problem is easier to deal with, at least conceptually. It is discussed --in `our analysis of the statistical results below. ` -- `-- ` The basic unit of concern in our risk/return analysis must be the individual corporation It is within the individual corporation that the balance between Parts `of this section are based on the* Illustrations given by Paul Cootner and Daniel Holland In' their study of Risk and Return forthe American Telephone and Telegraph Com- pany (M I T DSR Project No 9565) 4Economlsts and -financial analysts have' long proclaimed the existence of a "risk premium." By this phrase they'usually mean that! prospective Investors must be offered aboye average expectations of return (premium returns) in order to induce them to in- vest In projects having above average uncertainty (risk) Hence the higher yield on a corporate debenture than on a government bond is believed to be caused by adding the appropriate risk premium to the government's (riskiess) interest rate. - PAGENO="0202" 1752 COMPETITIVE PROBLEMS IN THE DRuG INDUSTRY expected returns and expected risks is struck. However, this is not meant to suggest that the only or best source of information about risk expectation `by entrepreneurs is historical statistics developed on a company `by company basis. Surely management forms its risk expectations on the `basis of experience; but not solely on the experience of its own firm. Likewise, entrepreneurs considering entry into an industry will assess `the general riskiness of the industry by exam- ining the range of corporations active (or previously active) in that industry. It is therefore, not unreasonable to seek a measure of e~vpected risks based on historical industrial experiences. It is our contention that returns among industry groups cannot be compared unless one has a measure of industry risks. The logic of our model of inter- industry risk/return comparisons follows from our beliefs about the inherent forces which create risky or uncertain situations within individual industries and, consequently, within companies. For the concept of an "industry" to have any validity in risk/return com- parisons it must be demonstrated that there exists sufficient similarities among various companies so that they may be meaningfully divided in industry groupings. A listing of the various phenomenon that might contribute to inter-industry difference in basic riskiness would certainly include the following: (1) Differences in the ease of entry of new firms into the industry or the ease of construction of new capacity in the industry. (2) Differences in the income elasticity of demands for the final products of the industry. (This would affect the response of the industry to general, economic activity.) (3) Differences in price flexibility in the industry. (4) Differences in the stability of major sources of raw materials. (5) Differences in storability and durability of products and raw materials. (6) Differences in exposure to foreign competition. (7) Differences in competition among existing prospective, or potential new products. Another major cause of differences in inter-industry risk character derives from differences in technological and research and development bases of industries' products. Differences in innovative processes and product obsolescence r'ates are prime examples. These considerations and others led us to stratify a number of companies into various industry groups. As will be seen below, our research has indicated that a fairly definite pattern of risk/return relationships does emerge from the indus- try grouping based on SIC stratification. The reasoning of our model is as follows: Assume an investor is seeking to enter an industry, i.e., to set up an economic concern in that industry or reinvest capital in a going concern in that industry, in such items `as plant expansion, product development, etc. The investor being reasonably experienced with the economic, financial and production problems of the industry, believes that he can expect to do as well as anyone else operating in that industry. He also knows that he is not omniscient, so that while he hopes on occasion to be more far-sighted, imaginative or effective than his competitors, he also knows that he is likely to be outwitted, or to outwit himself, or to run into a number of natural calamities as often as the reverse. His view, in that case, of the~ risks inherent in operating in the industry probably arise from observing the impact of errors and advantages upon the rates of return of the companies already engaged in the Industry. If the impact, on profitability, of some above or below average `behavior is severe the risk of entry will be lrrge. If, on the other hand, no reasonably likely error (or action of a competitor) could push him far from the average rcturn he might assume there was low risk. Thus, at least in theory, a concept of risk evolves `and as such is subject to statistical testing. (Of. Cootner and Holland p. 42). In order to perform these statstical tests it was necessary to construct `a guan- titive measure of `the industry risk (or uncertainty). We selected `measures of the dispersion of individual companies' rates of return about their industry's average rate of return for a given year. An industry which is characterized by relatively high dispersion of rates of return presents the prospective investor with much greater uncertainty as to the return he will achieve should he invest in that industry, than does an industry with low dispersion. Because our theory of return is essentially a long-run theory, we averaged over the period studies PAGENO="0203" COMPETITIVE PROBLEMS IN THE DRUG INDUSTRY 1753 (1950-1965) each industry's annual rate of return and dispersion. In this way our measures should not have been unduly iiifiuenced by abnormal years.5 (It will be readily noted by economists that in measuring risk as the average intercompany dispersion (or variance) we are, at the same time, staying with and breaking with standard economic tradition. The word risk is generally used by economists to represent certain aspects of the utility functions persons are presumed to maximize in their decision making processes. Specifically, the second moment of the expected utility function is considered the risk element 8 which one usually tries to minimize, while trying to maximize the first moment of the utility function. It is, therefore, natural to measure risk as some type of vari- ance (i.e., mathematical second moment) ; in so doing we are in keeping with tradition. However, the expected utility function concerns itself with eco ante risk. There is no strong reason to believe that temporal variance is a good measure of this quantity. In fact, when dealing with autocorrelated time series (as economic series almost always seem to be) we reject the usual reliance on individual company temporal variance.) The inter-company dispersion of returns measures one aspect of industry riskiness. We call it the interspacial component and view it as somewhat ana- logous with the uncertainty of any one company's market share in a nonregulated industry. There also exists an intertemperal component to industry riskiness. This component is analogous with the non~-predictable element of year to year changes in individual company or industry profitability. So defined it is very difficult to measure because most economic times series are highly autocor- related. We have developed some more general intertemporal dispersion measures and have used them, as far as possible, in our analyses. Our statistical tests on interspacial dispersion turned out to be significant and are discussed in full below. There are, however, a number of theoretical prob- lems which warrant discussion at this point. A few are easily disposed of. One is the problem of industry definition. The theory depends critically upon the idea of similarity between the companies assigned to an industry group. The industry groupings u-c chose are as homogeneous as the SlO based Compustat tape would allow. One may choose the precise industry composition in differing ways and thus, because the number value of our basic measure of risk is so critically dependent on tIle industry grouping, we believe that it was essential to test the sensitivity of our measure to different industry groupings. It is reassuring to report that the results were essentially unchanged when we em- ployed a small number of quite narrow and homogeneous industries,7 A similar problem arises because of the widely different sizes of the firms that are rightfully grouped in any industry and the fact that the firms used in our analyses (because of the selectivity of the Compustat tapes) tended to be the larger and more successful firms in each category. We, at this point, do not have sufficient data to investigate the effect that this may have on our results but have sufficient reason to believe that the inclusion of smaller firms would strengthen the relationship we have found. IV. DEFINITIONS OF RETURN In studying the relation between risk and return it is, of course, necessary to construct quantitative measures of both variables. The measure of risk has been defined in the previous section as the average interspacial deviation of company rates of return about the industry's rate of return. As can be seen from the mathematical deviation presented in Appendix A, the general definition of our risk measure does not depend at all on the specific definition of the rate of return. However, care must be taken in defining return, for the logic upon which the measure of risk is based maintains its economic validity if, and only if, the return calculated is a true, overall economic rate of return. Rates of return can be measured either at "book value" or at "market value." Book value returns relate the yearly income flow as reported on the company P&L statement to stock Balance Sheet items, such as total assets. Although book value figures are subject to many imperfections, in the long run they are the best indicator of real' economic return to invested resources. For a company 6Appendix A presents a detailed mathematical derivation of our measure of risk and contrasts it with other measures that have been proposed. 0 C. J. H. Markowit8, Port/oflo Beiection, New York: Wiley, 1959. May also be called cross-sectional or intercompany. 8 Throughout this paper "real" is used in its economic sense, meaning tangible or physical rather than intangible. PAGENO="0204" 1754 COMPETITIVE PROBLEMS IN THE DRTJG INDUSTRY which employed the same level of assets in an industry for an entire year and whose assets were financed solely by common equity, the ideal book rate of return would be: Net Income-Total Assets However, such a company is quite unusual, for most American corporations employ a seasonally changing level of assets, which are financed in varying degrees by both debt and equity capital. Therefore, measuring total assets at any point in time (and, consequently, at a particular season), one would find some industries with a higher than average level of assets while others would be below average due to seasonal factors. Further, net income for highly levered firms would represent a much smaller portion of the total return to the invested resources, than would it for an all equity financed company which had no fixed charges to pay. To overcome these objections we have developed the concept of total return to total permanently investable funds. Total return is defined as "Net Income plus Fixed Charges," while total permanently investable funds is "Total Assets minus Current Liabilities." It will be immediately recognized that our asset base measure-"Total Investable Funds" is the familiar Total Capitalization (Com- mon and Preferred Equity plus Long Term Debt) while our total return measure is the sum of the returns to both the equity and debt capital suppliers. We prefer to use the more general terms for they allow us to consider financial as well as industrial industries in the analyses. To summarize then, in order to adjust for industry differences in peak-sea- Sons and financing we have defined the book value rate of return as: B- Net Income + Fixed Charges - Total Asset - Current Liabilities For comparisons we have also considered other commonly used but, for our analysis, less meaningful measures of book return. They are: B' - Net Income - Common Equity and: B"- Net Income Total Assets B' is objectionable because it considers neither the totality of assets invested in the enterprise nor the totallty of return. As indicated above, B" is a mislead- ing measure when comparing industries with different degrees of leverage. Before defining our measure of market value rate of return it is important to emphasize one point. When we relate book return to enterpreneurial risks, we are asking whether resources are being efficiently allocated in the real economy. However, when we relate market returns to market risks we are concerned with the efficiency of the capital markets as allocators of financial instruments. A company may be experiencing monopoly returns (returns higher than justified by risk) on its book assets (i.e., monopoly real returns) while the holder of its equity instruments would receive a "normal" return if the monopoly profits were capitalized when the stock was issued. Although, the relationship between market return and risk does not directly bear on the question of efficient resource allocation it is of interest to us. The testing of the relationship offers an addition- PAGENO="0205" COMPETITIVE PROBLEMS IN THE DRUG INDUSTRY 1755 al test on the viability of our measure of risk and provides insight into the indus- tries deemed risky by the market. If stock values are ultimately tied to real economic phenomena, then the potential market risks should be related to real economic risks. Market rate of return is here viewed as the total income received by a pur- chaser of all the securities of a company related to his purhase price, on the assumption that he sells his holding at the end of a single year. (Other time spans can reasonably be considered.) Therefore, M Fixed Charges + Dividends + Change in Market Value Initial Market Value where Initial Market Value = Total Market Value of all debt and equity issues at beginning of year one,~ and Change in Market Value = Terminal Market Value less Initial Market Value. Fixed Charges and Dividends are those actually paid (or payable) by the corporation which the purchaser would receive (or accrue) in the course of the year. For obvious reasons we could not obtain market values for debt instruments and used Compustat book values for both debt and preferred equity. Oommon equity is evaluated as the `total market value of the common shares outstanding. This compromise with the ideal definition does not seem to be of great significance because of the relative unimportance of preferred stock and the relative stability of most corporate bonds. V. STATISTICAL ANALYSES9 The model used to test the industry risk/return relation was of the form: (I) Industry Return=a+b2 (Industry Risk) The major question we sought to answer concerned the sign and statistical sig- nificance of "b2" in the above equation. That is, if "b2" is positive and signifi- cantly greater than zero we have demonstrated our hypothesized relationship between industry return and our measure of risk-a relationship implying that high returns are associated with high risks during the period studied. Model I was fitted on both book (B) and market (M) rates of return. It is not necessary to study the detailed statistical analyses appearing in Appendix D to appreciate the relationship we have found. A simple graph can tell most of the story. Oonsidering book return and risk first, we have plotted in Figure 1 (average book return) on the vertical axis and Risk (average inter-company variance) on the horizontal axis. For each of the 59 industries studied we have plotted one point identifying its 15 year average risk/return position. The fact that the pat- tern of points form an upward sloping line demonstrates the posited relation- ship between risk and return. ° The following description of the results of our regression analyses is intended only to summarize the general findings. It is not a substitute for Appendix D which reports in detail the various models tested. PAGENO="0206" Industry Return (Net Income + Fixed Charges)/Total Capimlization Industry Risk = Average Inter-company Dispersion INDUSTRY RATE of 20 RETURN 10 20 30 30 50 60 70 330 .90 100 110 120 130 FIGURE 1 MODEL I: ~DWTRY RISK/RETURN PATTERN MEAS~ED AT 1300K VALUE ~D~TRY R~K NOTE: (~)REPRESENTS PoSITION OF PHARMACEUTICAL INDUSTRY 19 PAGENO="0207" Risk Th~DT~TRY RISK FIGURE 2 MODEL II: P'~DI~TRY RISK/RETURN PATTERN MEASURED AT MARKET VALUE NOTE: (~) REPRESENTS POSITION OF PHARMACEUTICAL JNI)USTRY PAGENO="0208" 1758 COMPETITIVE PROBLEMS IN THE DRUG INDUSTRY Exact quantification of relationship, as well as assessment of its statistical significance requires the full regression analysis presented in -Appendix E. From this analysis we have drawn the regression line fitted to the points. The equation is: (TB) Industry Return=8.6+0.038 (Industry Risk) The b2 (=0.038) coefficient is highly significant (F=50) at 57 degrees of free- dom. The correlation coefficient (R) is .68 while R2 equals .46. Figure 2 portrays the same relationship, with Return and Risk measured at market value. A glance at: the graph or the equation (1M) reveals much the same story: (IM) Industry Return=14.4+.007 (Industry Risk) In this equation b2 is again highly significant having an F test value of 82 with 57 degrees of freedom. The correlation coefficient equals .77 (out of a possible 1.00) while R2 reached 0.59-most impressive for a basically cross-sectional analysis. Now that the hypothesized relationship between risk and return has been statistically valldated, we may turn to questions concerning individual industries and their relation to the normal risk/return pattern. Given that an industry has a higher than average rate of return, it is both meaningful and possible to ask whether its risk is proportionately higher than average risk. For example, in Figure 1 the pharmaceutical industry has one of the highest book rates of return as well as one of the highest risks. Using the regression statistics we may test whether its particular risk/return point could reasonably be generated by the economic mechanism described by the regression line. Using a two standard error test-i.e., greater than 95% confidence limits-we can conclude that the point does belong to the normal pattern.15 In an attemept to examine other dimensions of risk we expanded our original model to include temporal variances as well as the spacial variance first used to measure risk. The yearly variance of the industry's rate of return about its temporal average was added first: (I') Industry Return=a+b2 (Industry Risk) +b3 (Industry Temporal Variance) In both equations I'B and I'~ there is some imphrovement in R2 from 0.46 to 0.51 and from .59 to .61., respectively. This improvement was at the cost of a reduction in degrees of freedom as well as the value of the F statistic. Including an average of the individual companies' temporal variance added little or no independent or partial explanatory power. Although we are primarily concerned with explaining industry rates of return, some time was devoted to analyses of individual company returns. The basic model used was: - (II) Company Return=a+b3 (Company Risk) +b3 (Industry Risk) Company risk is the standard temporal variance used by most researchers. While it yields the expected results (b2 positive and significant) for market rates of return, Model II yields a significafitly negative value of b2 for book rate of return. We attribute this negative coefficient to the high degree of auto-correlation found book value statistics. (In Section III we rejected the temporal variance as a measure of :risk on theoretical grounds related to auto- correlation.) In summary, we have seen that our hypothesized measure of industry risk has been statistically validated for both book and market measures of industry 10 Mathematically this finding is derived rom the limits of prediction formula (cf. Miller & Frei.inds, Probability and Statistics for Engineers, p. ~235): (a+bsro)±t~isS, .[i+1+n(xo-.~2]l/2 where t~/2=2. Setting X~ to 74 we get a y range of 19.4 to 10.2. The observed y value of 17.6 falls well within this range. PAGENO="0209" COMPETITIE PROBLEMS IN THE DRUG INDUSTRY 1759 rate'~ of return Further we have found the expected upward sloping risk/return relationship and have been able to test the éoñcurrencè of particular industries' results with the general pattern formed by American industry. The impressive magnitudes of both our t and F statistics as well as degrees of freedom, permit us to assert our results with a very high level of statistical confidence; higher in fact, than is usual in cross-sectional analysis. APPENDIx A MATHEMATICAL DERIVATION OF RISK/RETURN MEASURES Notes 1. Part I of this appendix defines in a general framework the mathematical and statistical quantities used in this study. The notation is more general than that used in Part II which defines the exact quantities used in this analysis. 2. For simplicity the letter "0" is used to represent any return quantity we are measuring. For example, C may stand for Net Income/Total Asset. PART I-GENERAL MODEL DEFINITIONS AND CALCUlATIONS OF RATES OF RETURNS AND VARIANCES FOR COMPANIES AND INDUSTRIES Each company, 1, is uniquely assigned to an industry, j. (1=1, . . ., n; j=1, . . ., N). The value of any variable for a given company In year t, may be written: C~1(t), (t=l, . . ~, T) C~1(t) is read! the value of variable C for the i" company in the j"~ industry for year t. The average value (mean) of variable C for company i is: ~ ~Ci~(t) The variance of C for company i is: T rT 12 T~J {C~,(t)]2= I ~ C~,(t) TTIf1 ~ Lt=i - T(T-i) For industry j the average (or `industry') value of variable C in year t is: ~TV11(t)C~1(t) C.1(t)= L ~ Wi1 i=i where the W~1(t)'s are the weights assigned to each company. They may be equal or, if assets are used as weights, unequal. (Note that a dot "." in place of a subscript~ means that we have summed over that subscript.) The variance of C/1(t) is given by: T T ~C.,(t)]2_I ~C.1(t) v - Lt=i T(T-l) 21-280-68-pt. 5-14 PAGENO="0210" 1760 COMPETITIVE PROBLEMS IN THE DRUG INDUSTRY This statistic, V(C.1), is the variance of the industry average over the years and must be carefully differentiated from V[C.1(t)J. The variance of the com- panies about the industry average in any one year is given by: a ~ [C~1(t)]2_[~ Cii(t)] V[C.1(t)]= ~ n(n-1) in the equal weight case, and by: ~ i=l -[C.1(t)]2 ~ W1~(t) V[C (t)}- `` - ~ r ~T~1(t) 12 ~ [~w~t~j in the case of unequal weights. The difference between V (Ci) and V[C.1 (t)] is quite significant for our analysis of interindustry riskiness. Year to year profitability for an entire industry may be quite stable. Yet in any one of those years, the individual com- panies whose profitabilities are averaged to yield the industry's figure, may be widely distributed about that mean. (That is, V[C.j (t) I may be very much larger than V(Cj).) Using the year to year industry variance tends to hide the important company to company variance. The risk of a change in market share facing individual companies is not at all reflected in the time series variance of industry sales of the detergent or automotive industries. Our pri- mary concern in this analysis is the risk as it appears to an individual company. Most scholars, on the other hand, have concentrated on the year to year industry variance. The V[O.j (t) I calculations will yield a T-element vector, V.~, each of whose elements is one year's value of V[C.1 (t) 1. The question then arises how best to summarize this vector (for graphic presentation and interindustry comparisons). One method is to indicate the range of the elements. It would be preferable if some scalar could be used rather than two numbers. Should the elements of V~ prove to be serially uncorrelated, we could construct an estimate of the variance of the underlying process by taking a (weighted) average of the elements. PART IT-SPECIFIC NOTATION DEFINITIONS AND CALCULATIONS OF RATES OF RETURNS AND VARIANCES FOR COMPANIES AND INDUSTRIES The specific quantities used in our analyses are outlined below. It should be noted that all measures are derived from the single datum point C11, which is defined in Part I, above, as Cij (t)-the value in year t of any variable for company i (in industry j). COMPANY Company value in t: Company's average value over time: 1 Company's temporal dispersion: 2 T V(C,1)=7.~ (C11_V~,)2 t 1=1 1 In Section V this statistic is called Company Return. 2 In Section V this statistic is called Company Risk. PAGENO="0211" COMPETITIVE PROBLEMS IN THE DRUG INDUSTRY 1761 Company's temporal standard deviation: = Company's temporal coefficient of variation: S/7J~ TYPICAL COMPANY (T.C.) `T.C. value in year t: C*t=N ~ i=1 Industry spacial dispersion about C*1 in t: ~ (C~t-C~)~ i i=1 Industry spacial standard deviation about ~ in t: S2(C1~) =~I[V2(C1~)] T.C. average value: -1~; U*=~ ~~~1C*t t=1 Average industry spacial dispersion about C* ~: V~(C~~) =~ ~ V2(C1~) I t=1 i~ Average industry spacial standard deviation about C~ ~: ~(C~~) ~ ~ S2(C1~) i £=1 i Industry's spacial coefficient of variation (about V*): T.C. average value: -~ t=1 T.C. temporal dispersion: V*(C~~)=~ ~ V(C~) 1=1 t T.C. temporal standard deviation: N ~ S(C~) t 1=1 t T.C. temporal coefficient of variation: PAGENO="0212" 1762 COMPETITIVE PROBLEMS IN THE DRUG INDUSTRY INDUSTRY Industry value in year t: N C.i=~ w~tC~~ i=l Industry spacial dispersion about C.~ in t: V1(C~~)=~~ (C~g_C.t)2 I t=1 Industry spacial standard deviation about C.~ in t: S1(C~1)= i/TT1(C~) Average industry value: *C.=~r ~ t=1 Average industry spacial dispersion: Vi(C~~)=~ ~V1(C~~) i t=I i Average industry spacial standard deviation: ~ S1(C~) i t=1 I Industry's spacial coefficient of variation: ~iIv. Industry's temporal dispersion about d: V(C. ~) =~ ~ (C. ~_V.)2 Industry's temporal standard deviation: S(C.~) =~/TT(C.~) t g In Section V this statistic is called Industry Return. 4 In Section V this statistic is called Industry Risk. 5 In Section V this statistic is called Industry Temporal Variance. PAGENO="0213" COMPETITIVE PROBLEMS IN THE DRUG INDUSTRY 1763 APPENDIX B DEFINITIONS OF FINANOIAL RETURNS The basic data source of our study was the Annual Industrial Compustat Tape issued by Standard Statistics, Inc., a Division of Standard & Poor's Corp. For each of the companies on the tape some or all of the quantities listed in Table B-i are given. Their definitions are those currently used in financial analysis and may be found in any of Standard & Poor's source books. The rates of return used in our study have been constructed from the S & P variables and are defined in Table B-2 in terms of the S & P number used in Table B-i. The base period for calculating the various quantities defined in Appendix A was 1950-1965. However, when a needed datum was not available in any par- ticular year, that measure and all derived measures were adjusted to permit maximum use of all available.information. TABLE B-1.-LISTOF VARIABLES APPEARING ON COMPUSTAT ANNUAL INDUSTRIAL TAPE S. & p. No. Balance sheet variables S & P. No. Balance sheet variables 1 2 Cash and equivalent. Accounts receivable. 17 18 Nonrecurring expenses. Net income. 3 Inventories. 19 Preferred dividends. 4 Current assets. 20 Available for common. 5 6 7 8 9 10 11 12 13 14 15 16 Current liabilities. Total assets. Gross plant. Net plant. Long-term debt. Preferred stock. Common equity (book value). 21 ~ 22 23 24 25 26 27 ~ 28 29 30 ~ Common dividends. Market value and miscellaneous variables ~ Stock price, high, $1 per share. Stock price, low, $1 per share. Stock price, close, $1 per share. Shares outstanding. Dividends per share. Adjustment factor (for changes in number of shares outstanding). Shares traded. Employees. Capital expenditures. Income statement variables * Net sales. Operating income. Depreciation and amortization. Fixed charges. Income taxes. TABLE B-2.-Definitions of variables used in stvdy in terms of S. & P. numbers. Symbol Accounting definition ~ S. & P. code definition t B Net income+Fixed charges 18+15 6-5tt Total assets-Current liabilities B' Netlncome Common equity 18 11 B" . Net income Total assets 18 6 M F Ixed charges±Dividends+AMarket value ftt 15+19+20+iiMVfft (25*24)+9+14 Market value ttt See table B-i. tt 6-5=9+10+11 because Total assets=Current llabilities=Total capitalization. ttf MVg(25g*24g)+9g+lOg. iiMVffiVg-MVg-i. PAGENO="0214" 1764 COMPETITIVE PROBLEMS IN THE DRUG INDUSTRY APPENDIX C-INDUSTRIES USED IN THE ANALYSES SIC No. Industry Total number of companies 0800 Forest products 7 1000 Metals, miscellaneous 12 1031 Lead and zinc 5 1042 Gold mining 1211 Coal, bituminous 7 1311 Dii, crude producers 9 1810 Motion pictures 8 Food products 43 2000 Packaged toods 9 2010 Meat packers 7 2020 Dairy products 7 2030 Canned toods 7 2046 Corn refiners 2 2070 Confectionery 2082 Beverages, brewers 17 2085 Beverages, distillers 11 2086 Beverages, soft drinks 7 Tobacco 10 2111 Cigarette manufacturers B 2121 Cigar manufacturers 4 2200 Textile products 15 2300 Textile apparel manutacturers 11 2510 Home furnishings 9 2600 Paper 17 2650 Containers, paper 10 2700 Publishing 17 2731 Publishing books 9 2800 Chemicals 43 2830 Drugs 29 2844 Cosmetics 12 2850 Paint Dii 25 2912 Integrated domestic 18 2913 Integrated international 7 2950 Building materials, roofing and wallboard 10 3000 Tire and robber goods 11 3141 Shoes 8 3221 Containers, metal and glass 7 3241 Building materials, cewent 10 3291 Abrasive products 7 3310 Steel 22 3331 Copper 6 3334 Aluminum 3400 Machinery, metal fabricating 13 3430 Building materials, heating, air conditioning, plumbing 14 3449 Miscellaneous metal work 7 Machinery 62 3511 Stoam generating 4 3522 Agricultural 6 3531 Construction and meterials handling 7 3533 Dii v~ell 6 3540 Machine tools 8 3550 Specialty 16 3560 Industrial 10 3569 General industrial 5 3570 Office and business equipment 14 3581 Vending machines 6 Electrical products 30 3600 Electrical and electronic leaders 6 3610 Electrical equipment 14 3622 Electrical industrial controls 3 3630 Electrical household appliances 7 3651 Radio-RV manufacturers 7 Electronic products 31 3670 Electronics 27 2679 Electronic components 4 3611 Automobile 5 3713 Autotrucks 3714 Auto parts and acceesories 17 3721 Aerospace 17 3740 Railroad equipment 11 3871 Watches 5 4210 Trucking 10 4400 Shipping 4511 Airtransport 13 4830 Radio, TV broadcasters 7 5311 Retail, department stores 16 5331 Retail, variety stores 10 5411 Retail, food chains 21 5600 Retail, apparel chains 6 5812 Eating places 6 Financial 14 6140 Finance 9 6145 Finance, small loans 5 PAGENO="0215" COMPETITIVE PROBLEMS IN THE DRUG INDUSTRY 1765 APPENDIX D REGRESSION ANALYSES Section V described two regression models: (I) Industry Return=a+b2 (Industry Risk) +b3 (Industry Temporal Variance) (II) Company Return=a+b2 (Company Risk) +b3 (Industry Risk). As indicated the description was quite abbreviated and did not detail the full models nor all the relevant statistic. In order to accomplish this, we shall use the notation introduced in Part II of Appendix A. Before turning to the specific models studied, let us consider a general four variable step-wise regression model: x, =a+b2x2+b3x3+b4x4 In this model x1 is the "dependent variable," while variables x2, x3, and x4 are tue "independent variables." Regressing x1 on x2 alone would yield the a and b2 coeffi- cients of the reduced model x1=a+b2x2; regressing x, on both x2 and x8 would yield a, b23, and b3.2 of the model xl=a+b2x2+b3x3. (The a's might, of course, be different.) In the following discussion and tables b2 refers to the coefficient of variable 2 in a two-variable regression, b3 refers to the coefficient of variable 3 in a two-variable regression (i.e., x1=a+bsx3), b2.3 the coefficient of variable 2 in a three-variable regression involving x3 (i.e., xi=a+b2x2+b3xs) etc. Reformulating and expanding the models present in Section V we have: (I) t~. = a+ b2[Vi(C1~)]+ b3[V(C. ~)]+ b4[V*(C~~)] i t t tY~= a+ b2[V(C~~)]+ b3[VT(C1~) J+ b4[V(C. i)]. i t As will be recalled from Appendix A, ~. measures the (15 year) average return for an industry, while C~ measures the (15 year) average return for a company. Model I was fitted with 59 observations (one per industry), while Model II was fitted with 766 observations (one per company). Each model was fitted for each of the four basic rates of return-B, B', B", and M (see Appendix B for definitions). Tables P-i and D-2 summarize almost all of the statistics for the various regression steps. In simple regressions (y=a+bx) the F statistic serves as a test of significance of both b and B2; in multiple regressions each coefficient has it own F test reported a does the R2. PAGENO="0216" 1766 COMPETITIVE PROBLEMS IN THE DRUG INDUSTRY TABLE D-l.-Regression model Ja V.= a+ b2[Vi(C~1fl+ b3[V(C. ~)1+ b4[V*(C~~)} i t Definition of rate of return b Regression statistic M B B' B' a 14.4 8.6 12.2 5.6 0. 00724 0.08314 0. 00215 0. 1300 F 82.36 49.8 12. 5 25. 6 102 0.59 0.46 0.18 0.31 a 14. 6 9.9 11. 5 6. 5 53 0.00760 0. 126 0.04463 0.213 F 67.0 16.7 6.8 8.3 0.54 0.23 0.10 0.13 a 16. 1 10.7 12.4 6.4 b4 0. 00153 0. 00053 0. 00001 0. 07432 F 35.6 0.7 0.5 7.8 102 0.38 0.01 0.01 0.12 a 14.2 8. 5 11.7 5. 5 52.3 0.00502 0.071 0.00177 0. 11496 F 9.7 32.8 7.1 17.5 53.2 0.00280 0.062 0.05197 0. 10766 102 0.61 0.51 0.21 0.34 F 43. 5 29.43 7.3 14.4 S~ 3.7632 2. 1313 3. 1105 1.7869 Mc 19.0 10.8 12.5 7.2 a 14.1 8.5 5.5 52.34 0.00311 0.07149 0. 12393 F 2.3 32.8 15.9 b3.24 0.00356 0.06065 0. 13769 F 3.8 5.0 2.7 54.32 0.00047 0.0004 (0) F 2.1 0.8 -0. 92165 0.4 0.62 0.52 0.34 F 20.4 29.4 9.6 Se 3. 7286 2.1342 1. 7966 Mc 19.0 10.8 7.2 a 59 observations. b See appendix B. `Decrease in significance if variable 4 allowed to enter regression. TABLE D-2.--Regression model ~ a t~~=a+ b2[V(C~4)I+ b3[V'i(C~1)}+ b4[V(C.~)] t i £ Definition of rate of return b Regression statistic -- M B B' . B' a 19.3 11.2 13.5 7.5 52 0.00131 -0.00084 -0.00057 -0.01611 F 647. 1 479.6 630. 5 6.9 102 0.46 0.38 0.45 0.0009 a 16.9 8.8 12.4 5.4 53 0.00351 0.04124 0.00090 0.15534 F 61.8 68.9 0.3 86.0 102 0.07 0.08 0.0004 0.09 a 17.7 10.7 13. 1 7.0 0.00341 0.06523 -0.07799 0.11425 F 41.6 5.1 0.9 5.2 102 0.0516 0.007 0.0012 0.0067 a 16.7 8.9 13.1 5.5 bnz 0.00126 -0.00084 -0.00038 -0.02786 F 596.6 551.7 650.8 22.3 53.2 0.00083 0.69116 0.00421 0.17098 F 32.7 117.7 1.2.0 102.9 .102 0.48 0.46 0.46 0.1244 F 353.3 334.9 325.6 55.3 8, 11.4 5. 1 15. 5 3.5 M'~ 21.8 11.0 12. 5 7.3 76 observations. b See app. B. PAGENO="0217" COMPETITIVE PROBLEMS IN THE DRUG INDUSTRY 1767 APPENDIX E SAMPLE OF DATA GENERATED ]~T1OM COMPUSTAT ANNUAL INDUSTRIAL TAPE' 2830 1500 ABBOTT LABORATORIES 2830 26000 AMERICAN HOME PRODUCTS CORP. 2830 26100 AMERICAN HOSPITAL SUPPLY CORP 2830 69400 BAXTER LABORATORIES 2830 72860 BECTON DICKINSON CO - 2830 91000 BRISTOL-MYERS COMPANY 2830 128251 CARTER WALLACE INC. 2830 202400 CUTTER LABORATORIES S 2830 313600 GILLETTE CO. S - 2830 397700 JOHNSON/JOHNSON S - S - 2830 -406300 KENDALL CO - : 2830 428700 LILLY ELI CO - - S S 2830 464400 MCKESSON & ROBBINS, INC. (MD.) 2830 465700 MEAD JOHNSON/CO S S 2830 471000 MERCK & COMPANY - 5 - 2830 479000 MILES LABORATORIES INC - 2830 . 543200 NORWICH PHARMACAL CO. 2830 565800 PARKE, DAVIS & CO. - S *-.-** 2830.- 579000 PFIZER UCHASCD &-COCTINCC -. . . 2830 591800 PLOUGH INC 2830 619550 R1CHARDSON-MERRELL INC. 2830 629150 RORER WM H - 2830 643300 SCHERING CORP. 2830 .648000 SEARLE G D CO . S S - - 2830 665500 SMITH KLINE/FRENCH LABORATORIES INC - .. - 2830 693600 STERLING DRUG INC. S - 2830 704920 SYNTEX CORP. . . S S 2830 755550 UPJOHN CO . S S -. - S 2830 768700 WARNER-LAMBERT PHARMACEUTICAL CO. . . . - S . - ABBOTT LABORATORIES-DRUGS - . Year - mv ret MV Ern+int B capit Earnings corn eqty Earnings assets 1950 1951 -8.287 42.327 -17.432. 1.814 11.131 . 19.575 15.076 . 12.806 13. 063 12.026 19.256 17.476 14.606 15. 051 - 13.852 14.021 10.894 - 9.445 . 9.777 .8.951 1952 1953 1954 1955 . 1956 1957 1958 1959 1960 1961 1962 1963 1964 1965 77777.000 .77777. 000 77777.000 .77777.000 .77777. 000 77777.000 .77777.000 77777.000 .77777.000 . 77777. 000 .77777.000 . 77777.000 77777. 000 . 77777.000 77777. 000 77777. 000 77777.000 77777.000 77777.000 77777.000 . 77777. 000 77777.000 15.169 ~. 16. 105 17.434 - 16. 188. 14. 936 . 13. 446. 11.477 - 13.118 . . 14.356 14. 887 15. 159. 9.837 .10. 585 11. 133 10. 707 10. 196 9.151 8.226 - 9.070 9.871 10. 778 11.024 No:years 5 5 1616 - 5.910 14. 509 15. 157 10.229 V/I . 529.167 9.285 3.516 1.745 S/I 23. 004 8. 047 1. 875 1. 321 (S/T)/(I)B 3.892 .210 .124 .129 C(.) 20. 523 17. 524 18. 942 12. 049 C( )B 25 860 17 998 19 701 12 259 1Any designation of 77777.000 IndIcates no data available. PAGENO="0218" f CC~'~'Cl)< .833 = 1.4994 or $1.50 manufacturer's selling price.) Brand product prices are verified in The Red Book or Blue Book at the net or wholesale list price level. Products specified by the generic name might be filled by a popular brand or one of a number of generic or less popular brands at a wide price range. Therefore, the calculated manufacturers' selling price does not always coincide with the manufacturer's actual prices since quantity discounts or deals to the trade are not usually reflected nor does each pharmacy follow an identical pricing policy. However, the relative positions of the products based primarily on numbers of prescriptions and also dollar volume permit a valid com- parative analysis of market shares and long term competitive trends of the products discussed in this study. The trends expressed in market shares in most cases are the same for dollar volume and numbers of prescriptions. Differences in market share between prescriptions and dollar volume for a product reflect: 1. Its price relative to others in the class; 2. Its potency and the size of the prescription; and 3. The length of time the product is taken The last point may also reflect the incidence of side effects which might se- verely limit the continuing use of any of the products. A product with fewer side effects presumably could be given in larger doses and for longer periods of time but actually resulting in fewer prescriptions written. III. ANALGESICS-NON-NARCOTIO Non-narcotic analgesic is a general classification for pain relievers, not in- cluded in the morphine class. They are used in a wide variety of medical con- ditions. The total category has increased approximately 400% in total number of prescriptions written and over 700% in dollar volume between 1956 and 1965. Non-narcotic analgesics represented 1.64% of all prescriptions written in 1956 and accounted for 4.75% in 1965. The increase in importance of this class among all products is due to the development of synthetic types of analgesics which have reduced the danger of habit or drug dependency developing in conjunction with treatment. Although 16 products represented three-fourths of the total class in 1965. there were 182 products being marketed which were audited (See `Tables 1A &1B). Over 50% of the 1965 market is represented by drugs introduced since 1956. PAGENO="0238" CC_C F\~ 0 I ~ =~ CD> CD 3g~ 02 CD~ g.:2 ~ 0 3* 0 ~c) ~ C, CI) 03 ~ N_CCDOOCC_)OC_C1CC_C * CJ000Q1 (710)0) (00 ~ 0000) -.J ~J (71 ~.J(Co~N_C~)CJ1 OV.J~0) * C ~0 C,, 0) CD C, o CD. a0 C) 30 CD -~ - C) C) ~ 4~ 0) C,.) -J 00 0) ~ 2~ p0~4P~~J~ c)3 c~:: ~ ~ H~HHHH I I I I -~ ~ 00 ~ ~ 00 -J .-J 0000 ~p5°. ~ NPN~ 0000O_C0000~~CC_CN.): ~ PAGENO="0239" COMPETITIVE PROBLEMS IN THE DRUG INDUSTRY 1789 TABLE 2-A.-ANTIARTHRITVCS, NONSTEROI DAL [Percent share of market based on numbers of prescriptions written, 1956-651 Product 1956 1957 1958 1959 1960 1961 1962 1963 1964 1965 1 47. 5 43. 8 36. 4 41. 9 43. 3 35. 9 30. 8 33. 9 28. 9 22. 9 2 15. 0 13. 6 14. 0 11. 2 8. 2 7. 9 8. 5 6. 4 3. 9 4. 3 8.3 7.2 8.2 7.4 6.5 4.7 4.4 5.7 4.5 4.2 4 6.8 5.3 3.4 1.7 .7 .8 .9 .9 1.6 1.2 5 1.6 3.4 3.6 3.5 4.5 5.7 7.0 5.9 9.0 8.4 6 3.8 8.5 5.2 4.3 3.2 5.9 1.4 1.2 .8 7 6. 1 11. 5 15. 6 13. 9 12. 6 15. 5 15. 7 13. 5 8 .1 .5 1.8 5.5 4.0 3.1 9 9.2 14.5 14.1 11.4 10.0 10 1619 Cumulative market share of above pro- share of above pro- ducts 79. 2 77. 1 80. 2 82. 4 83. 2 81. 8 86. 4 89. 3 80. 2 85. 7 All others 20. 8 20. 8 19. 8 17. 6 16. 8 18. 2 13. 6 10. 7 19. 8 14.6 Category total 100.0 100.0 100.0 100.0 100.0 100.0 100.0 100.0 100.0 100.0 TABLE 2-B.-ANTIARTHRITICS, NONSTEROI DAL [Percent share of market based on dollar volume, 1956-651 Product 1956 1957 1958 1959 1960 1961 1962 1963 1964 1965 1 52.5 46.2 36.3 41.7 41.8 34.3 28.7 32.0 25.7 19.3 2 13.0 11.7 10.9 8.9 6.4 6.0 6.6 4.8 2.9 3.0 3 7.2 6.4 7.2 6.1 5.4 4.0 3.7 4.9 4.0 3.6 4 5.3 4.2 2.8 1.3 .6 .7 .7 .7 1.4 1.4 5 3.1 6.3 6.6 6.5 8.0 9.0 11.6 10.2 14.4 13.3 6 5.4 12.2 6.8 5.9 4.3 7.9 2.0 1.3 .9 7 5.8 11.5 16.0 14.0 12. 1 15.2 14.4 11.9 8 .1 .6 2.8 8.6 6.3 5.3 9 10.3 15.4 14.9 11.8 9.9 10 19.7 Cumulative market share of above products 81. 1 80. 2 81. 8 82. 8 84. 2 83. 2 89. 5 93. 3 82. 2 88. 8 All others 18. 9 19. 8 18. 2 17. 2 15. 8 16. 8 10. 5 6. 7 17. 8 11. 2 Category total 100. 0 100. 0 100. 0 100. 0 100. 0 100. 0 100. 0 100. 0 100. 0 100. 0 V. ANTnoioTIcs-BR0AD AND MEDIUM SPECTRUM Broad and medium spectrum antibiotics represent the largest product class both in number of prescriptions written and dollar volume and are characterized by a large number of highly competitive products. The broad range of infections treated by antibiotics and their prophylactic use against reinfection contribute to the size and growth of this class. It has increased between 85% and 90% both in prescription and dollar volume over the last ten years, representing 7.3% of the total number of prescriptions written in 1956 and 7.8% in 1965. Seven major products used in 1956 accounted for nearly 70% of the class. By 1965, these same products accounted for only 35% of the prescriptions written. The introduction of other products (new agents or modifications of existing products) in the inter- vening ten years were responsible for the dediluse. Over 40% of the 1965 market was from products introduced since 1956. Thirteen products represented over 76% of the class in 1905 while 84 were identified in the audit (see Tables 3A & 3B). PAGENO="0240" 1790 COMPETITIVE PROBLEMS IN THE DRUG INDUSTRY TABLE 3-A-ANTIBIOTICS, BROAD AND MEDIUM SPECTRUM Percent share of market based on numbers of prescriptions written, 1956-651 Product 1956 1957 1958 1959 1960 1961 1962 1963 1964 1965 1 33.7 23.6 10.9 6.8 5.7 5. 1 4.5 4.1 3.7 3.7 2 9.7 10.6 12.9 14.7 14.5 10.0 5.6 4.9 4.9 4.0 3 9.3 5.4 4.5 3.6 5.2 7.4 8.3 8.1 7.8 7.9 4 7.4 5.1 6.3 3.0 1.1 .8 .9 .7 .5 .4 5 6.1 2.7 1.2 .6 .5 .3 .3 .2 .2 .3 6 1.8 1.5 1.9 2.1 1.4 1.6 3.3 4.8 7.6 7.2 7 1.6 5.5 5.0 5.0 4.3 4.9 5.8 5.2 5.1 5.5 8 .3 .2 .1 .1 .2 .3 .6 .8 2.7 5.9 9 14.9 22.7 22.3 14.7 11.6 10.5 9.0 9.4 8.8 10 . 1 2.0 3.7 4. 1 5.3 5. 1 5.0 5.2 4.13 11 .8 7.5 9.6 10.0 8.7 6.9 7.7 7.1 12 3.3 16.6 17. 5 17.6 15.8 15.4 15. 3 13 1.4 5. 1 7.6 Cumulative market 7.5 7.1 6.5 share of above products 69. 9 69. 6 68. 3 72. 7 79. 3 79. 9 78. 8 All others 30. 1 30.4 13.7 27.3 20.7 20.1 21.2 Category total 100.0 100.0 100.0 100.0 100.0 100.0 100.0 73. 0 27.0 77. 3 22.7 76. 6 23.4 100.0 100.0 100.13 TABLE 3-B--ANTIBIOTICS, BROAD AND MEDIUM SPECTRUM [Percent share of market based on dollar volume 1956-651 Product 1956 1957 1958 1959 1960 1961 1962 1963 1964 1965 1 32.8 21.3 9.4 5.5 4.7 4.0 3.8 3.5 2.9 2.9 2 10.9 11.4 13.4 15.2 15.4 11.2 6.8 6.1 6.4 5.4 3 8.8 4.7 3.8 3.0 4.8 7.3 8.0 7.7 7.4 7.3 4 7.5 4.8 5.9 2.5 .8 .6 .6 .5 .3 .3 5 6.1 2.6 1.2 .6 .5 .4 .3 .3 .2 .2 6 2.1 1.7 2.1 2.2 1.5 1.6 3.5 5.0 7.8 7.4 7 1.7 5.5 5.0 4.8 4.1 4.5 5.3 4.7 4.7 5.1 8 .3 .2 .1 .1 .3 .4 .8 1.0 2.6 5.13 9 15. 9 22. 5 21. 9 14. 3 11. 5 10. 1 8. 8 8.9 8. 0 10 .2 2.9 5.1 5.4 6.7 6.8 6.7 6.8 5.4- 11 1.0 8.2 9.3 9.6 8.2 6.4 7.3 6.7 12 3. 4 16.2 16. 8 16. 6 14. 7 14. 5 15. 2 13 1.5 5.6 8.4 8.1 7.5 7.2 Cumulative market share of above products 70. 2 68. 3 67. 3 72. 5 78. 8 80. 2 79. 2 73. 5 77. 3 76. 1 All others 29. 8 31.7 32. 7 27. 5 21. 2 19. 8 20. 8 26. 5 22. 7 23. 9 Categorytotal 100.0 100.0 100.0 100.0 100.0 100.0 100.0 100.0 100.0 100.0 VI. ANTIOBIOTICS-PENIOILLINS Penicillin is synonymous with the dramatic growth in the drug industry and frequently is regarded as the antibiotic. Penicillins represent a growth of approxi- mately 92% in the number of prescriptions written and an increase of 108% in dollar volume during the ten years. In 1965, they accounted for 3.15% of all prescriptions-slightly higher than their 2.86% share in 1956. Much of the growth can be attributed to the development of new synthetic penicillins, although some of the old established products have continued to be prescribed more frequently than a decade earlier. In 1965, eleven products accounted for almost 84% of the class while 89 products were reported in the audit (see Tables 4A & 4B). Forty- three percent of the market was represented by new products introduced since 1956. PAGENO="0241" COMPETITIVE PROBLEMS IN THE DRUG INDUSTRY 1791 TABLE 4-A-ANTIBIOTICS, PENICILLINS (Percent share of market based on numbers of prescriptions written, 1956-651 Product 1956 1957 1958 1959 1960 1961 1962 1963 1964 1965 1 19.9 20.8 12.3 5.0 1.5 2.2 1.3 1.2 1.1 0.8 2 17. 1 16. 1 16. 3 12. 5 21.8 24. 5 30. 1 27. 5 24. 2 21. 2 3 14.8 11.6 11.1 8.6 6.3 5.7 5.2 3.6 2.7 1.8 4 13.2 9.9 10.4 9.2 13.5 12.1 12.0 13.9 15.0 14.9 5 8.9 13.7 10.8 7.9 2.9 3.9 4.2 4.0 2.6 1.9 6 3. 0 15. 3 19. 3 18. 9 19. 5 20. 1 21.0 23. 2 20. 8 7 1.0 3.5 2.8 2.7 2.6 2.5 4.0 5.0 8 - .9 5.0 3.7 5.5 4.7 5.1 4.4 4.6 9 .9 8.1 6.3 4.1 4.3 4.0 2.1 10 2.8 6.3 11 4.7 Cumulative market share of above products 73. 9 75. 1 78. 1 81. 9 79.7 82. 4 84. 3 83. 1 84. 0 83. 8 All others 26.1 24.9 21.9 18.1 20.3 17.6 15.7 16.9 16.0 16.2 Category total 100. 0 100. 0 100. 0 100. 0 100.0 100. 0 100. 0 100.0 100. 0 100. 0 TABLE 4-B-ANTI BIOTICS, PENI Cl LLI NS (Percent share of market based on dollar volume, 1956-651 Product 1956 1957 1958 1959 1960 1961 1962 1963 1964 1965 1 24.4 24.9 14.1 5.5 3.0 2.1 1.3 1.2 1.0 0.7 2 13. 7 12. 4 12. 7 18. 2 18. 9 21. 2 26. 2 25. 1 22. 2 19. 3 3 16.9 12.4 12.6 8.6 6.5 6.8 5.2 4.3 3.1 2.4 4 9.0 6.7 6.9 6.2 8.9 8.4 8.7 9.8 9.6 9.2 5 10.8 15.7 12.7 8.5 6.4 4.1 4.6 3.9 2.7 1.9 6 4.2 18.9 23.1 21.4 21.4 21.4 22.4 24.7 20.9 7 1.3 4.2 3.3 3.2 3.0 4.0 4.6 5.4 8 1.0 5.9 4.3 6.0 5.0 5.4 4.4 4.4 9 1.1 10.4 8.6 5.4 5.3 4.5 2.7 10 5.2 10.3 11 5.1 Cumulative market share of above products 74. 8 76. 3 80. 2 81. 3 83. 1 81. 8 80. 9 81. 4 82. 0 81. 8 All others 25. 2 23. 7 19. 8 18. 7 16. 9 18. 2 19. 1 18. 6 18. 0 18. 2 Category total 100. 0 100. 0 100. 0 100. 0 100. 0 100. 0 100. 0 100. 0 100. 0 100. 0 \TII ANTIHISTAMINES Antihistamines represent one of the more slowly growing therapeutic cate- gories of the drug industry. Their performance depends on the severity of the hay fever and cold seasons. Both the effectiveness and the side effects of the various products vary from patient to patient. Trial of several different products may be necessary to find the one product most effective with the fewest side effects for the particular patient and disorder being treated. Dollar volume over the 1956- 1965 period increased 53.5% while the number of prescriptions increased 19.2%. In 1956 this category represented 4.25% of prescriptions written while in 1965 it represented only 2.92%. Only six products accounted for over 86% of the class although 62 products were audited in 1965 (see Table 5A & 5B). Considered a breakthrough in the late 1940's, antihistamines reached a point beyond which no significant new products have emerged. The leading antihista- mine products in 1956 continue to be the leading products ten years later. They are often switched in patients when drowsiness, the most common side effect, occurs. The leading six products which accounted for nearly 92% of the number of prescriptions written in 1956 accounted for 78.3% in 1965. The top seven products accounted for 86.2% of all antihistamine prescriptions written in 1965. 81-280-OS-pt. 5-1O PAGENO="0242" 1792 COMPETITIVE PROBLEMS IN THE DRUG INDUSTRY TABLE 5-A.--ANTIHISTAMINES [Percent share of market based on numbers of prescriptions written, 1956-65] Product 1956 1957 1958 1959 1960 1961 1962 1963 1964 1965 1 25. 2 24. 8 21. 5 18. 2 20. 3 20. 7 20. 5 19. 5 22. 8 24. 5 2 24. 4 28. 4 27. 7 19. 8 23. 0 23. 1 23. 8 20. 6 24. 0 23. 0 3 17. 0 16. 0 16. 0 12. 0 10. 1 8. 9 9. 9 8. 4 7. 5 6. 5 4 10.3 10.5 11.0 10.7 9.7 9.5 9.5 10.5 11.0 12.3 5 9.1 12.8 10.0 7.9 7.9 6.3 5.4 5.9 5.9 5.6 6 4.8 6.5 6.6 7.0 6.3 5.9 5.6 5.0 6.6 6.4 7 .3 5.6 6.3 6.1 6.5 9.0 6.2 7.9 7.9 Cumulative market share of above products 91. 8 99. 3 98. 4 81. 9 83. 4 80. 9 83. 7 76. 1 85. 7 86. 2 All others 8. 2 . 7 1. 6 18. 1 16. 6 19. 1 16. 3 23. 9 14. 3 13. 8 Category total 100.0 100.0 100.0 100.0 100.0 100.0 100.0 100.0 100.0 100.0 TABLE 5-B.-ANTIHISTAMINES [Percent share of market based on dollar volume, 1956-65] Product 1956 1957 1958 1959 1960 1961 1962 1963 1964 1965 1 20.6 19.7 16.6 14.0 15.2 15.9 15.6 14.8 17.4 18.2 2 25.7 29. 7 29. 1 20. 2 23.4 23.7 24. 1 20.9 24.6 24. 0 3 16.7 16.9 16.5 11.2 10.0 8.7 9.4 8.4 7.4 6.4 4 16. 3 15. 3 14.8 13. 6 12.3 12. 3 12. 0 13.7 13. 8 15. 1 5 7.7 10.4 7.9 6.2 6.5 5.5 4.9 5.4 5.4 5.3 6 4.8 7.0 7.4 7.7 7.0 6.5 6.1 5.2 7.0 6.7 7 .4 6.2 7.0 6.9 6.8 9.6 6.4 8.0 8.1 Cumulative market share of above products 91. 8 99. 4 98. 5 79. 9 81. 3 79. 4 81. 7 74. 8 83. 6 83. 8 All others 8. 2 . 6 1. 5 20. 1 18.7 20. 5 18. 3 25.2 16. 4 16. 2 Category total 100.0 100.0 100.0 100.0 100.0 100.0 100.0 100.0 100.0 100.0 VIII. ANTIOBESITY-AMPHETAMINES Among the products used to control weight gain and when necessary, encour- age weight loss are the amphetamines. Included in this product class are com- binations with sedatives to modify the exhilaration often produced by ampheta- mines, or tranquilizers to provide relief of tension-anxiety symptoms. The number of prescriptions written have increased approximately 36% during the ten year period while the dollar volume has increased approximately 64%. This increase has been due principally to the introduction of new products through 1960. The total class represented 3.4% of all prescriptions written in 1956 de- clining to 2.6% in 1965. Although the eleven products shown in Tables GA & GB represent over 80% of the market for amphetamine products the importance of this class is reflected by the number of products which are competing in the antiobesity market. In 1965 over 80 products were being sold by brand or generic name in the ampheta- mine class. PAGENO="0243" COMPETITIVE PROBLEMS IN THE DRUG INDUSTRY 1793 TABLE 6-A.-ANTIOBESITY-AMPFIETAMI N ES [Percent share of market based on numbers of prescriptions written, 1956-65] Product 1956 1957 1958 1959 1960 1961 1962 1963 1964 1965 1 35. 5 35. 2 31. 6 25. 0 25. 7 20. 9 17. 9 13. 0 11. 7 12. 5 2 31. 1 29. 6 33. 9 32. 9 27. 6 27. 5 25. 1 25. 8 25. 3 21. 0 3 3.6 3.1 2.2 5.3 6.8 6.3 7.5 7.8 8.4 7.4 4 2.4 2.8 3.2 2.7 2.3 1.7 1.6 1.7 2.1 2.1 5 1.7 1.0 .7 .4 .6 2.8 2.6 2.7 3.7 5.0 6 1. 4 3. 0 5. 3 8. 4 8. 1 9. 8 10. 0 7. 4 6. 7 5. 4 7 .5 2.2 1.9 1.4 1.2 1.6 3.2 3.1 8 .2 6. 5 7. 3 10. 1 13. 5 13. 8 15. 0 9 .5 1.5 2.4 2.1 2.8 4.5 10 2.9 3.0 1.5 2.8 2.2 11 .7 1.8 2.5 3.7 3.9 Cumulative market share of above products 75. 7 74.7 77. 4 77. 1 80. 0 82. 8 83. 2 79. 6 84. 2 82. 1 All others 24.3 25.3 22.6 22. 9 20.0 17.2 16. 8 20. 4 15.8 17. 9 Category total 100. 0 100.0 100. 0 100.0 - 100. 0 100. 0 100. 0 100. 0 100.0 100.0 TABLE 6-B.-ANTIOBESITY-AMPHETAMINES [Percent share of market based on dollar volume, 1956-65] Product 1956 1957 1958 1959 1960 1961 1962 1963 1964 1965 1 40. 5 38.7 33. 4 25. 9 24. 9 20. 2 17. 1 12.3 11. 0 11. 5 2 31. 8 30. 1 35.3 34.9 27.6 27. 6 24. 6 25. 0 25. 1 20.3 3 2.1 1.8 1.2 3.7 5.4 4.7 5.3 6.0 6.7 5.8 4 1.8 2.2 2.6 2.3 1.9 1.5 1.2 1.2 1.6 1.4 5 1.0 .8 .5 .3 .6 2.7 2.4 2.7 3.9 5.3 6 1.6 3.5 6.3 9.7 9.2 11.1 11.4 7.9 7.0 6.0 7 .5 2.1 1.7 1.3 1.1 1.3 2.6 2.6 8 .2 8.0 9. 1 13.0 17.0 17.2 18. 5 9 .8 2.0 3.0 2.5 3.3 5.3 10 2.9 3.1 1.4 2.8 2.1 11 1.0 2.4 3.2 4.7 4.9 Cumulative market share of above products 78. 8 77. 1 79. 8 79. 1 70. 1 84. 1 84.6 80. 5 86. 0 83.7 All others 21. 2 22. 9 20. 2 20. 9 29.9 15. 9 15. 4 19. 5 14. 0 16. 3 Category total 100. 0 100. 0 100. 0 100. 0 100. 0 100. 0 100. 0 100. 0 100. 0 100. 0 IX. ATABAXIOS The term ataraxic is generally used to designate products which are employed as calming or tranquffizing agents. However, these products may be used to treat a number of different disorders so that they vary considerably in strength and action. Our discussion is restricted to those products used to treat patients residing at home rather than those who are hospitalized in mental institutions. This classification of drugs represents one of the more important areas of the drug industry and is one which has grown significantly over the past decade in both numbers of prescriptions (93%) as well as in dollar volume (100%). One half of the tranquilizers on the market in 1965 did not exist in 1956. In 1065 these products accounted for 7.4% of prescriptions dispensed as compared to 6.6% in 1956. Twelve products represented almost 87% of the class in 1965 of the 56 which were audited (see Tables 7A & 7B). New products introduced since 1956 account for over 50% of the market. PAGENO="0244" 1794 COMPETITIVE PROBLEMS IN THE DRUG INDUSTRY TABLE 7-A.--ATARAXICS [Percent share of market based on numbers of prescriptions written 1956-651 Product 1956 1957 1958 1959 1960 1961 1962 1963 1964 1965 1 43. 0 39. 5 32. 9 32. 9 27. 4 23. 7 21. 1 17. 5 13. 6 13. 5 2 21.5 15.9 12.3 10.2 8.8 8.4 9.7 7.4 6.8 7.4 3 18.4 14.0 11.8 11.8 12.7 10.4 8.7 7.6 7.3 6.7 4 4.5 5.6 5.2 4.6 3.4 3.6 2.2 2.0 2.0 1.9 5 3.1 4.9 5.4 4.4 2.9 2.5 1.9 3.1 2.3 1.6 6 6 10.3 17.8 16.7 14.4 12.6 9.3 8.5 7.3 6.3 7 3.6 2.9 1.5 .6 .5 .2 .1 .1 .1 8 2.5 3.3 4.5 2.3 1.2 1.2 .7 .7 .7 9 5 1.6 1.1 1.4 1.4 1.3 1.7 1.9 10 3.1 7.7 6.6 7.5 7.9 7.3 5.4 11 11.4 26.0 29.4 31.4 32.1 29.7 12 2 8.0 12.9 Cumulative market share of above products 91. 1 96. 3 92. 1 91. 3 92. 7 96. 9 92. 6 87. 6 88. 3 86. 7 All other 8.9 3.7 7.9 8.7 7.3 3.1 7.4 12.4 11.7 13.3 Category total 100. 0 100. 0 100. 0 100. 0 100. 0 100. 0 100. 0 100. 0 100. 0 100. 0 TABLE 7-B.-ATARAXICS [Percent share of market based on dollar volume 1956-651 Product 1956 1957 1958 1959 1960 1961 1962 1963 1964 1965 1 46. 3 42. 4 32. 9 32. 9 27. 0 23. 1 19. 8 16. 6 12. 2 12. 3 2 18. 8 13. 5 11. 4 9. 9 8. 8 8. 2 9. 7 7. 6 7. 1 7. 8 3 19. 7 15. 2 12. 0 11. 9 12. 5 10. 4 8. 7 7. 5 7. 0 6. 4 4 4.2 4.8 4.9 4.2 2.9 3.1 1.9 1.7 1.7 1.6 5 3.1 4.8 5.6 4.4 3.0 2.8 2.1 3.4 2.7 1.8 6 .5 10.5 17.9 16.2 12.9 10.8 7.5 6.8 5.4 4.7 7 3.8 2.6 1.5 .6 .4 .2 .1 .1 .1 8 2.7 4.3 4.8 2.4 1.1 1.3 .8 .7 .7 9 .6 2.0 1.3 1.8 1.7 1.6 2.0 2.1 10 3.1 8.3 7.0 8.1 8.9 8.5 6.1 11 12.1 27.4 30.7 32.6 32.8 30.1 12 .2 8.0 12.9 Cumulative market share of above products 92.6 97.7 92.2 90.9 91.8 96.1 91.7 87.8 88.2 86.6 All others 7.4 2.3 7.8 9.1 8.2 3.9 8.3 12.2 11.8 13.4 Category total 100. 0 100. 0 100. 0 103. 0 100. 0 100. 0 100. 0 100. 0 100. 0 1 00. 0 X. RAUWOLFIA-DIURETIC COMBINATIONS Rauwolfia-diuretic combinations provide the anti-hypertensive action of ran- wolfia plus the diuretic action (increasing the secretion of urine) of the thiazide products. The latter is considered important in the management of hypertension also, producing a complementary anti-hypertensive action. The reduction of body fluid apparently is related to lessening some causal effects of bypertension~ Although rauwolfia products have been used since the early 1950's and mercurial diuretics even longer, it was not until after the introduction of thiazide diuretics that the rauwolfia-diuretic combination gained acceptance in medical prac- tice. The importance of this class can be seen by the rapid rate of growth over~ the past seven years and the many new product entries into the market (see Tables 8A & 8B). The first five rauwolfia-diuretic combinations introduced in 1959 accounted for 0.67% of all prescriptions written and by 1065 this category represented 1.71% o2 ~1l prescriptions written. As a class, the prescriptions in- creased by 344% from 1956 to 1965. The dollar volume increased by 334% during the same period. Strong interest in new combinations indicates that there is a medical need for additional compounds to try on individual patients who w-ere unresponsive to an earlier product or whose prognosis might be improved by a newer product. The demand for these products clearly indicates that a need exists and virtually PAGENO="0245" COMPETITIVE PROBLEMS IN THE DRUG INDUSTRY 1795 a new market has been created for products combining complementary thera' peutic effects. Although the fixed dosages of manufactured products of this kind limit the versatility often needed to treat individuals, they are appropriate to treat large numbers of patients. One obvious reason for the acceptance of these combinations is the greater assurance that the patient will take one tablet or capsule containing several medical ingredients than he would if several tablets or capsules had to be taken separately. If the dose of any ingredient is critical, each can be prescribed separately and patients can be controlled on an individual basis. TABLE 8-A.---RAIJWOLFIA-DIURETIC, COMBI NATION (Percent share of market based on numbers of prescriptions written, 1956-651 Product 1956 1957 1958 1959 1960 1961 1962 1963 1964 1965 1 54.2 36.0 28.9 28.5 29.1 2 20.4 19.7 11.1 8.5 7.5 3 14.1 23.4 25.6 18.0 20.5 4 5.2 10.6 6.6 5.7 2.6 5 .2 3.4 2.0 1.4 1.3 6 3.7 9.8 10.3 8.1 7 1.7 4.6 9.3 8.6 8 1.3 8.7 8.3 6.9 9 1.3 3.2 5.8 10 .7 2.3 2.6 11 .4 1.4 1.0 12 13 Cumulative markets share of above products 100.0 99.8 99.7 96.7 93.9 All others . 2 . 3 3. 3 6. 1 Categorytotal 100.0 100.0 100.0 100.0 100.0 22.9 4.6 15.7 3.4 1.0 9.4 12.9 10.2 5.8 3.8 2.0 1.5 .7 18.3 2.6 18.6 1.5 .7 7.5 17.6 8.2 3.5 6.5 1.6 2.6 3.5 93.9 6. 1 92.6 7.4 100.0 100.0 TABLE 8-B.-RAUWOLFIA-DIURETIC COMBI NATION [Percent share of market based on dollar volume, 1956-651 Product 1956 1957 1958 1959 1960 1961 1962 1963 1964 1965 1 50. 5 34. 1 27. 9 28. 2 27.2 2 22.8 22.1 11.3 9.0 8.2 3 13. 6 23. 6 26. 5 18. 2 21.2 4 6.3 8.8 5.9 5.2 2.5 5 .1 3.3 1.7 1.4 1.2 6 4.4 10.8 11.4 8.9 7 1.7 4,2 8.1 7.9 8 1.5 9.2 7.8 6.8 9 1.5 3.4 6.5 10 7 2.2 2.6 11 .5 1.6 1.3 12 13 Cumulative market share of above products 100. 0 99. 6 99. 9 96. 6 94. 3 Allothers .4 .1 3.4 5.7 Category total 100. 0 100. 0 100. 0 100. 0 100. 0 22. 8 5.7 15. 7 3.2 .9 10.5 10.7 11.6 6.0 3.8 2.2 1.4 .6 19. 1 3.2 19. 5 i.~ .8 8.9 14.9 9.1 3.5 6.6 1.6 2.4 2.8 95. 0 5.0 93. 8 6.2 100. 0 100. 0 XI. CORONARY VASODILATOR5 This therapeutic category contains the drugs which are used primarily to bring symptomatic relief of angina pectoris and essential hypertension. The prescriptions written for these drugs during 1950-1965 increased at a slower rate than that of the overall drug industry. The number of prescriptions increased 72% while the equivlilent dollar volume of th~se products increased 158%. These products accounted for 1.74% of the prescriptions written both in 1956 asid 1965. Thirteen out of 08 products represented over 90% of the class in 1965 (see Tables 9A & 9B). PAGENO="0246" 1796 COMPETITIVE PROBLEMS IN THE DRUG Th~DUSTRY Although this product class is the subject of active research efforts, may of the new products introduced have not made a significant impact. Patients responding to one kind of treatment are usually not switched when new products are introduced. Of seven products introduced between 1956 and 1965, only two have made any impact on the market. TABLE 9-A.-CORONARY VASODILATORS [Percent share of market based on numbers of prescriptions written, 1956-65J Product 1956 1957 1958 1959 1960 1961 1962 1963 1964 1965 Cumulative market share of above products 64. 0 74. 9 80. 4 76. 4 80. 5 86. 1 81. 3 84. 2 86. 3 91. 1 All othero 36. 0 25. 1 19. 6 23. 6 19. 5 13. 9 18. 7 15. 8 13. 7 8. 9 Category total 100. 0 100. 0 100. 0 100. 0 100. 0 100. 0 100. 0 100.0 100.0 100.0 TABLE 9-0.-CORONARY VASODILATORS [Percent share of market based on dollar volume, 1956-651 Product 1956 1957 1958 1959 1960 1961 1962 1963 1964 1965 Cumulative market share of above products 55. 1 70. 8 78. 4 72. 1 75. 7 79. 8 82. 4 81. 7 83. 4 90. 8 All others 44. 9 29. 2 21. 6 27. 9 24. 3 20. 2 17. 6 18. 3 16. 6 9. 2 Categorytotal 100.0 100.0 100.0 100.0 100.0 100.0 100.0 100.0 100.0 100.0 XII. DIABETIC THERAPY-OTHER Diabetic products other than insulin fall into two chemical classes-sulfony- lureas and biguanide. Each of the products shown in Tables 1OA & lOB Is a different chemical compound except one sustained release form. None of these products existed in 1956 and only one was available in 1957. They can lower the blood sugar level when given orally in the treatment of selected patients with diabetes without the need for insulin injections. The product class has grown over 1400% in total numbers of prescriptions written since 1957. By 1965 these products represented 1.26% of all prescrip- tions written and 2.3% of the total dollar market. In 1965, the total product class was represented by 8 products with the five shown accounting for over 99% of the market. 1 29. 5 39. 4 41. 8 37. 3 35. 2 36. 8 25. 6 2 19. 6 19. 4 19. 1 17. 7 22. 5 27. 5 20. 4 3 7.7 6.6 10.3 8.4 8.6 10.3 8.2 4 4.6 4.5 1.9 1.4 2.5 .9 .4 5 2.1 2.5 3.3 6.6 5.6 5.0 4.7 6 .5 2.5 2.6 2.3 1.8 1.3 1.6 7 1.0 2.1 3.0 2.5 1.7 8 .4 .6 .8 .7 2.1 9 .5 1.1 1.2 10 10.1 11 5.2 12 5.0 13 23.8 20.9 22.4 22.1 18.6 19.1 4.9 5.7 6.3 .4 1.0 .7 6.0 4.6 3.9 2.3 1.3 1.3 2.1 1.6 1.2 2.1 4.2 2.6 1.9 6.7 9.0 12.8 16.9 12.5 3.3 6.4 8.7 6.8 4.5 5.1 .5 2.1 3.5 1 32. 7 45. 0 52. 9 45. 7 44. 3 46. 1 28. 1 2 6.2 5.8 6.0 5.2 7.2 8.9 6.5 3 7.8 8.6 9.4 9.0 10.8 14.3 9.7 4 7.0 7.1 3.0 2.3 3.5 1.3 .6 5 .6 .8 1.1 2.2 1.7 1.5 1.3 6 .8 3.5 3.9 3.8 2.6 1.8 2.2 7 1.8 3.4 4.5 4.1 2.6 8 .3 .5 .6 .6 1.8 9 .5 1.2 1.3 10 18.1 11 5.2 12 5.0 13 26.4 21.8 22.1 7.0 5.4 5.3 5.1 6.0 6.2 .4 .8 .6 1.6 1.3 1.0 2.9 1.6 1.9 2.6 2.2 1.6 1.8 3.7 2.2 2.2 7.8 13.5 20.8 25.4 17.9 3.3 6.4 8.7 6.8 4.5 5.1 .8 2.9 4.7 PAGENO="0247" COMPETITIVE PROBLEMS IN THE DRUG INDUSTRY 1797 Although the products in this class have achieved a significant prescription volume, they are not a replacement for insulin injections (estimated annual sales of $22 million) where the patient's insulin requirement is high. Their principle advantage is that where they are indicated they can be given orally replacing the daily injection required to administer insulin. TABLE 10-A-DIABETIC THERAPY, 0TH ER [Percent share of market based on numbers of prescriptions written, 1956-65] Product 1956 1957 1958 1959 1960 1961 1962 1963 1964 1965 1 100. 0 99. 0 82. 2 84. 0 81. 7 77. 0 77.3 75. 0 70. 2 2 1. 0 14. 5 8. 7 12. 0 15. 3 16. 6 12. 8 13. 2 3 3.3 7.3 5.9 5.0 3.5 3.8 3.4 4 .4 2.7 2.6 5.3 8.1 5 3.1 Cumulative market 4.8 share of above products 100. 0 100. 0 100. 0 100. 0 100. 0 100. 0 100. 0 100. 0 All others 99. 7 .3 Category total 100. 0 100. 0 100. 0 100. 0 100. 0 100. 0 100. 0 100. 0 100. 0 TABLE 10-B-DIABETIC THERAPY, 0TH ER [Percent share of market based on dollar volume, 1956-65] Product 1956 1957 1958 1959 1960 1961 1962 1963 1964 1965 1 100. 0 99. 1 85. 1 85. 7 84. 9 81. 3 79. 1 77. 6 72. 8 2 .9 12.9 9.2 10.3 12.4 16.7 11.9 12.8 3 2.0 5.1 4.3 3.3 2.2 3.0 2.5 4 5 3.0 2.0 5.1 7.6 5 2.3 3.9 Cumulative market share of above products 100.0 100.0 100.0 100.0 100.0 100.0 100.0 99.9 99.6 All others .1 .4 Category total 100. 0 100. 0 100. 0 100. 0 100. 0 100. 0 100. 0 100. 0 100.0 XIII. DIURETICS Diruetic agents are used to treat edema which may be associated with a number of different disorders. The drugs may be used as `adjuncts in the manage- ment of congestive heart failure, to control the severity of cardiac decompensa- tion or to treat patients with edema associated with kidney and liver disease, pregnancy, obesity, etc. The diuretic class `had one of the most dramatic increases over the ten year span increasing 490% in th~ number of prescriptions written and 560% in dollar volume. Of the total industry volume, diuretics accounted for 1.12% of the prescriptions written in 1956 but increased to 3.68% of those written in 19&5. Eleven products accounted for 80% of the class of 43 products audited in 1965 (see Tables hA & hiB). In reviewing the 1965 data it appears that diuretics that did not exist in 1956 are responsible for nearly all of the prescriptions written. The market is fairly well fractionated with about a dozen compounds actively competing for signifi- cant positions. PAGENO="0248" 1798 COMPETITIVE PROBLEMS IN THE DRUG INDUSTRY TABLE 11-A-DIURETICS [Percent share of market based on numbers of prescriptions written, 1956-651 Product 1956 1957 1958 1959 960 1961 1962 1963 1964 1965 1 52.6 53.5 17.0 7.8 6.6 5.6 4.7 5.1 5.0 3.7 2 13.3 12.9 3.3 1.4 .9 .4 .2 .2 .4 .3 3 6.3 2.4 .2 .1 4 5.6 3.1 .6 .3 .1 .1 .1 .1 5 . 8 72. 1 61. 8 42. 3 33. 3 32. 7 23. 7 23. 1 25. 9 6 14. 3 20.6 19. 6 17. 3 18.8 14.4 20. 0 7 7.8 11.3 7.8 6.9 4.5 3.3 4.9 8 5.4 5.0 5.4 6.3 7.1 4.1 9 1.0 3.0 4.9 5.7 6.6 5.9 10 .2 2.2 3.2 5.7 8.7 3.3 11 .1 3.5 4.1 5.0 5.1 4.7 12 .4 6.1 4.9 4.4 4.9 13 .2 2.6 Cumulative market share of above products 77. 8 72.7 93.2 93. 5 88. 5 80. 9 85. 6 79. 9 78. 4 - 80.3 All others 22. 2 27. 3 6. 8 6. 5 11. 5 19. 1 14. 4 20. 1 21.6 19. 7 Category total 100. 0 100. 0 100. 0 100. 0 100. 0 100. 0 100. 0 100. 0 100. 0 100. 0 TABLE 11-B-DIURETICS [Percent share of market based on dollar volume, 1956-651 Product 1965 1957 1958 1959 1960 1961 1962 1963 1964 1965 1 49. 6 48. 7 14. 0 7. 7 7. 3 6. 4 5. 8 6. 9 6. 7 5. 4 2 18.4 15.6 3.9 1.7 1.4 .5 .2 .3 .4 .4 3 8.2 3.1 .3 .1 4 2.4 1.2 .2 .1 .1 5 1. 1 74.5 64. 1 41.5 32.8 30. 5 22. 1 21.5 24.2 6 13.4 19.4 18.9 17.1 18.0 13.9 19.3 7 7.2 10.5 7.6 6.2 4.3 3.1 4.7 8 5.2 4.7 5.6 6.6 6.8 4.4 9 .9 3.0 4.8 5.4 5.9 5.6 10 .3 2.3 3.3 6.4 9.2 3.6 11 .1 3.2 4.0 4.4 5.1 4.5 12 .4 5.9 4.7 4.3 4.7 13 .2 2.7 Cumulative market share of above products 78.6 69. 7 92.9 94.3 89.6 79. 8 83. 5 79. 1 77. 1 79. 5 All others - 21. 4 30. 3 7. 1 5. 7 10. 4 20. 2 16. 5 20. 9 22. 9 20. 5 Category total 100.0 100.0 100.0 100.0 100.0 100.0 100.0 100.0 100.0 100.0 XIV. Hon~ioxns, CORTICOIDS The term corticoid is short form of the name corticosteroid. The desired action of corticoids is to favorably influence a large number of clinical conditions by the suppression of inflammatory or immulogical processes. Unfortunately the anti- inflammatory action of the corticosteroids is inseparable from their metabolic effect which limits their usefulness for prolonged administration because of rather serious side effects. Because of these side effects a great number of deriv- atives have been prepared and are used extensively in treating patients with various types of disorders-both acute and chronic conditions (collagen diseuses, bronchial asthma, hay fever, allergies, dermatoses and certain inflammatory eye diseases). PAGENO="0249" 0 0 L~:j 0 0) 0) C) ~0 C) ~ ~ ~0 3 3 00 = 0 0 (I) C) 0 00 C) 0 00 C,) - ~ 0)00 O~ 00 N) 00 (~C)~ (00 ~ 3 2 > 00 0 00 0 z C') C) 0 - -4 C) 0 00 C,, PAGENO="0250" 1800 COMPETITIVE PROBLEMS IN THE DRIJG INDUSTRY XV. CoRTIcoIDs WITH ANTI-INFEOTIVES Corticoids with anti-infectives describes a group of products which represent the combination of the anti-inflammatory corticosteroids with various anti- infectives in topical preparations with a maximum local effect. Because of the broad range of competitive corticosteroids and anti-infective agents, there are numerous possible combinations obtaining different effects (see Tables 13A. & 13B). Consequently, the many companies which have marketed preparations make this category one of the most highly competitive in the entire field. Over the last ten years the category has grown approximately 78% in numbers of prescrip- tions written and 87% in dollar volume. The category has maintained its relative position representing 1.51% of all prescriptions written in 1956 and 1.55% of all prescriptions written in 1965. The dollar volume for this category declined from 1.63% of total in 1956 to 1.45% of total in 1965. It appears from the large number of products being offered in this class plus the interest in a variety of combinations of corticosteroids and anti-infectives that there is a need for a broad range of products to treat infections and inflam- mation from many different causes. Twenty-five products account for over 80% of the class and 76 products make up the other 20%. Many of these combination products are available in specialized forms for treatment of conditions located in the eyes, ears, and nose as well as for general topical use anywhere on the body surface. TABLE 13-A.-CORTICOIDS WITH ANTI-I NFECTIVES [Percent share of market based on numbers of prescriptions written, 1956-651 Product 1956 1957 1958 1959 1960 1961 1962 1963 1964 1965 1 41. 1 29.7 23.8 16.4 14. 9 12.7 2 18.2 14.7 12.0 7.4 5.7 4.9 3 8.8 4.2 3.3 1.9 1.7 1.7 4 4. 6 3. 8 3. 1 3. 0 3. 4 3. 9 5 2.8 4.3 2.7 3.4 3.3 2.9 6 2.8 3.7 4.4 6.2 6.2 8.4 7 2.6 1.9 1.3 .8 .7 .4 8 2.3 2.5 2.3 2.1 1.6 1.1 9 1.2 1.3 2.3 2.7 1.4 1.0 10 . 8 9. 5 12. 6 12. 1 11.7 10. 1 11 . 4 .4 2. 1 2.0 1.8 1. 5 12 2 6. 1 14.2 14.3 9.2 6.7 13 1. 0 1.9 2.2 3. 1 3.3 14 . 1 . 8 2. 4 2. 1 1.8 15 1 2.2 1.9 1.9 16 3. 5 2. 2 1. 4 17 1. ~ ~. ~ 6. 9 18 1. 8 7. 8 10. 7 19 1.4 1.4 20 8 1.2 21 1.4 22 .2 23 - 24 25 10.0 8.1 7.3 .7 3.3 2.3 2.0 2.0 .6 .6 .7 .5 2.4 1.9 2.9 2.5 2.5 2.5 2.2 1.7 7.9 8.2 7.8 8.7 .3 .2 .3 .3 1.1 .9 .6 .4 1.1 .5 .5 .5 9.6 11.0 10.0 9.5 1.4 1.3 1.1 1.4 5.4 4.2 3.5 2.6 3.0 2.4 2.4 2.5 1.3 1.5 1.9 1.8 1.0 1.3 1.3 1.2 .5 .5 .3 .1 8.2 8.3 9.3 11.3 13.4 11.1 13.2 11.5 .5 .4 .2 .1 1.8 1.6 1.6 2.4 7.1 6.8 6.4 5.4 .9 1.7 1.7 2.1 1.1 1.5 2.9 2.5 .8 1.8 2.7 3.2 .3 .8 1.1 1.2 86. 2 85. 5 85. 5 81. 4 83. 9 82. 3 13.8 14.5 14.5 18.6 16.1 17.7 Cumulative market share of above products 85. 8 83. 2 86. 9 86. 3 All others 14. 2 16. 8 13. 1 13. 7 Category total 100. 0 100. 0 100. 0 100. 0 100. 0 100. 0 100.0 100.0 100.0 100.0 PAGENO="0251" COMPETITIVE PROBLEMS IN THE DRUG INDUSTRY 1801 TABLE 13-B.-CORTICOIDS WITH ANTI-INFECTIVES (Percent share of market based on dollar volume, 1956-65] Product 1956 1957 1958 1959 1960 1961 1962 1963 1964 1965 1 41.7 29.0 23.0 16.0 14.6 12.0 9.4 2 14.2 12.7 11.2 6.3 4.7 4.0 2.8 3 6.1 2.9 2.3 1.3 1.1 1.2 .4 4 4.6 3.8 3.1 3.0 3.4 3.9 2.4 5 3.4 4.7 4.3 4.3 4.0 3.6 2.9 6 2.5 3.8 4.2 6.2 6.2 7.9 7.0 7 2.6 2.1 1.3 .8 .7 .4 .2 8 2.4 2.7 2.4 2.3 1.8 1.3 1.3 9 1.4 1.5 2.9 3.6 1.7 1.1 1.3 10 .7 10.8 14.8 14.1 13.9 12.5 11.3 11 .3 .4 1.6 1.6 1.5 1.2 1.2 12 .2 6.7 13.0 13.6 8.2 5.7 4.4 13 1.3 2.1 2.7 3.6 3.7 3.1 14 .1 .8 2.4 2.1 1.8 1.3 15 .1 2.1 2.1 2.2 1.1 16 2.4 1.5 .9 .4 17 1.8 5.6 7.0 8.6 18 1.6 7.5 10.0 12.4 19 1.1 1.1 .3 20 1.0 1.5 2.1 21 1.5 7.2 22 .2 1.1 23 1.1 24 .9 25 .5 Cumulative market share of above products 81. 7 83. 9 88. 1 87. 0 87. 1 85. 8 85. 2 All others 18. 3 16. 1 11. 9 13. 0 12. 9 14. 2 14. 8 Category total 100. 0 100. 0 100. 0 100. 0 100. 0 100. 0 100. 0 7.5 1.9 .4 1.9 2.9 7.1 .2 1.0 .6 12.6 1.0 3.5 2.6 1.5 1.6 .4 8.8 10.1 .2 1.9 7.1 2.1 1.4 1.9 1.1 6.9 1.6 .5 2.9 2.6 6.6 .3 .6 .6 11.3 .9 2.9 2.5 1.9 1.5 .2 9.7 12.2 .1 1.9 6.7 2.1 3.0 3.0 1.5 6.1 1.6 .4 2.5 1.8 7.3 .2 .5 .6 10.8 1.0 2.1 2.5 1.8 1.4 .1 12.1 10.4 .1 2.8 5.6 2.5 2.3 3.5 1.4 81. 3 18. 7 84. 2 15. 8 82. 0 18. 0 100. 0 100. 0 100. 0 XVI. ORAL MUSCLE RELAXANTS Drugs in this group act on the central nervous system and are useful in pro- moting relaxation of skeletal muscle spasm. They have been recommended as an aid in the management of almost every musculo-skeletal and neuromuscular con- dition in which painful muscle spasm occurs. Thus these agents can be used in connection with the treatment of a number of different medical conditions. Over the past decade the number of prescriptions for these products has in- creased 350%, while the equivalent dollar value has increased 405%. In 1956 these products accounted for 0.56% of the total number of the prescriptions written. In 1965 they were responsible for 1.26% indicating wider acceptance of the use of these products in various conditions. The top five products in 1956 were responsible for nearly 76% of all prescrip- tions written. By 1965, one-third of the 47 products in the class accounted for over 82% of the prescriptions (see Tables 14A & 14B). The newer products introduced the combined effect of muscle relaxants and analgesics or sedatives, or both. The effectiveness of muscle relaxants alone was not enough in many cases for satis- factory treatment. If muscle spasm caused pain, then pain relief was also indi- cated. If muscle spasm could be reduced by a sedative, working in the central nervous system, then a sedative was indicated.~ Frem these experiences, the new, convenient forms evolved, increasing competition further. PAGENO="0252" 1802 COMPETITIVE PROBLEMS IN THE DRUG INDUSTRY Product TABLE 14-A.-ORAL MUSCLE RELAXANTS [Percent share of market based on numbers of prescriptions written, 1956-65) 1956 1957 1958 1959 1960 1961 1962 1963 1964 1965 Cumulative market share of above products 75. 6 82.2 74. 4 76. 2 88. 3 74. 9 79. 8 84. 2 83. 1 82. 5 All others 24. 4 17. 8 25. 6 23. 8 12. 7 25. 1 20. 2 15. 8 16. 9 17. 5 Category total 100. 0 100. 0 100. 0 100. 0 100. 0 100. 0 100. 0 100. 0 100. 0 100. 0 TABLE 14-B-ORAL MUSCLE RELAXANTS [Percent share of market based on dollar volume, 1956-65] Product 1956 1957 1958 1959 1960 1961 1962 1963 1964 1965 1 37.0 39.2 10.6 5.5 4.8 2.0 2 13.0 10.9 4.0 1.1 .6 .7 3 14.4 10.0 4.9 1.7 1.3 1.0 4 9.1 11.9 9.4 6.0 6.8 4.4 5 4.8 2.6 1.6 1.1 2.1 2.1 6 5.6 32.9 19.3 18.4 11.4 7 10.5 5.6 1.8 1.8 8 .5 21.8 13.0 9.1 9 10.6 15.8 14.8 10 .1 4.5 5.5 11 .1 9.6 5.5 12 6.3 9.8 13 .3 4.6 14 1.6 15 16 0.1 .7 0.7 0.5 0.5 1.6 .8 .6 .8 5.3 9.5 9.1 10.5 1.8 1.6 4.3 5.2 10.5 6.0 6.6 6.0 .4 .3 .3 .2 6.4 3.9 3.5 3.7 12.5 13.3 11.6 9.1 8.7 8.6 10.4 8.6 4.0 2.6 3.7 3.6 10.3 8.1 8.3 9.2 8.7 10.4 10.0 9.8 4.9 3.8 6.5 6.9 1.0 8.6 4.1 3.0 .7 1.3 3.2 5.7 Cumulative market share of above products 78. 3 80. 2 74. 4 72. 9 85. 3 74. 3 77. 6 79. 5 82. 7 82. 8 All others 21. 7 19. 8 25. 6 27. 1 14. 7 25. 7 22. 3 20. 5 17. 3 17. 2 Category total 100.0 100.0 100.0 100.0 100.0 100.0 100.0 100.0 100.0 100.0 XVII. PSYCHOSTIMULANTS Drugs in this therapeutic category are used to treat depressed patients with widely varying causes. Thus the compounds can offer different types and different mechanisms of action. This category has shown dramatic growth although the base was small in 1950 (0.38% of total prescriptions written in 1956 to 1.3% in 1905). The number of prescriptions written has increased 510% over the period 1956-1965, while the dollar volume of these products has increased tenfold. The three principal products in 1956 accounted for 70% of the prescriptions written. Two of these were withdrawn from the market in 1962. By 1965, twelve products represented over 87% of prescriptions in a class of 36 products (see Tables 15A & 15B). New products in this field have been viewed with guarded optimism since the effects of psychostimulants might mask underlying problems. As a result, each new product is tried in different kinds of patients to find where it is most effective. The product class, therefore, has a history of product in- troduction, initial acceptance and a decline to some lower level of use. 1 28.4 33.8 11.1 5.8 4.9 2.1 2 16.0 14.1 5.8 1.6 1.0 .9 3 15.6 12.1 6.7 2.3 2.0 1.4 4 10.9 15.2 10.4 10.0 9.8 4.9 5 4.7 2.7 2.3 1.5 2.8 2.9 6 4.3 28. 3 17. 4 17. 6 10.9 7 9.4 5.2 1.7 1.7 8 4 22.1 12.8 9.2 9 10.1 15.4 14.5 10 .1 4.0 5.2 11 .1 9.7 5.4 12 6.4 9.7 13 .2 4.8 14 1.3 15 16 .8 .8 .6 .6 2.3 1.4 1.1 1.3 8.2 14.4 14.4 15.6 2.4 2.4 5.6 7.0 10.0 5.9 6.1 5.4 .4 .3 .4 .2 6.4 3.9 3.5 3.3 12.5 13.2 10.8 8.5 7.7 7.4 8.6 7.2 4.2 2.7 3.3 3.4 10.4 8.7 8.1 8.9 8.6 10.1 9.4 9.2 4.2 3.2 4.9 5.4 1.0 8.8 4.1 2.9 .6 1.0 2.2 3.6 PAGENO="0253" COMPETITIVE PROBLEMS IN THE DRUG INDUSTRY 1803 TABLE 15-A.-PSYCHOSTIMULANTS [Percent share of market based on numbers of prescriptions written, 1956-651 Product 1956 1957 1958 1959 1960 1961 1962 1963 1964 1965 I 39.2 42.8 23.6 20.1 13.6 11.0 12.5 12.5 16.0 15.9 2 19.6 10.2 1.7 1.6 .4 .1 3 11.3 7.2 2.4 2.3 1.7 .8 .4 .3 .7 1.2 4 .4 11.9 27.4 8.5 2.4 .6 5 30. 1 39. 9 21. 2 19. 0 12. 8 19. 7 18. 3 12. 5 6 8.2 17.9 18.8 16.3 15.4 23.6 15.2 7 7.0 12.7 5.3 2.0 3.2 5.0 1.8 8 6.5 13.7 12.5 8.7 5.6 3.8 2.2 9 4.8 10.9 6.3 3.6 2.8 2.4 1.7 10 10.6 16.4 20. 5 22.2 29. 5 11 6.9 12.8 21.6 4.9 3.1 12 7.3 Cumulative market share of above products 70. 5 72. 1 71. 7 80. 1 91. 0 89. 4 83. 0 96. 8 93. 7 87. 7 All others 29. 5 27. 9 28. 3 19. 9 9. 0 10. 6 17. 0 3. 2 6. 3 12. 3 Category total 100. 0 100. 0 100. 0 100. 0 100. 0 100. 0 100. 0 100. 0 100. 0 100. 0 TABLE 15-B.-PSYCFIOSTIMULANTS [Percent share of market based on dollar volume, 1956-65] Product 1956 1957 1958 1959 1960 1961 1962 1963 1964 1965 1 32.5 36.0 18.2 11.5 7.8 6.0 7.3 7.7 9.9 10.3 2 25.6 14.3 2.1 1.1 .3 .1 3 9.4 4.1 1.7 .8 .5 .3 .2 .1 .3 .5 4 .4 15.5 17.2 6.7 1.9 .6 5 30. 1 39.9 21. 2 1. 9 12. 8 19. 7 18. 3 12. 5 .6 13. 0 22. 9 23. 6 20. 8 19. 1 28. 1 18. 2 7 8.8 13.7 5.8 2.1 3.2 5.5 2.1 8 6.1 11.5 10.9 8.0 5.4 3.3 2.0 9 6.4 12.4 7.1 4.5 3.3 2.8 1.8 10 10.8 17.0 21.3 21.6 30.2 11 5.9 11.8 19.8 4.5 2.7 12 Cumulative market share of above prod- ucts 67. 9 69. 3 69. 3 94. 3 92. 2 90. 1 84. 5 98. 6 94. 3 87. 1 All others 32. 1 30. 1 30. 7 5. 7 7. 8 9. 9 15. 5 1. 4 5. 7 12. 9 Category total 100.0 100.0 100.0 100.0 100.0 100.0 100.0 100.0 100.0 100.0 XVIII. SEDATIVEB-BAIIBITURATE The barbiturates comprise an important and valuable class of central nervous system depressants. It has been reported that they are used more to produce sleep than for any other purpose. 1 Proper selection and rational use of the barbiturates depend on an intimate knowledge of their pharmacology and toxicology as well as ~an acquaintance with the important variations in action and toxicity produced by certain individual products. Barbiturates are differentiated by their duration of action ranging from longer duration barbital and phenobarbital to shorter dura- tion secobarbital, pentobarbital and hexobarbital. Combinations of barbiturates are often used to provide a more balanced effect by combining those with differing Tates of action and dissipation. The total category has increased approximately 7% in total number of prescrip- tions written and 22% in dollar volume between 1956 and 1965. Barbiturates rep- resented 7% of all prescriptions written in 1956 but only 4.3% in 1965. The lack of significant increases in demand for barbiturates may be related to the development of non-narcotic analgesics, tranquilizers and barbiturate-anal- gesic combinations. The importance of this class, however, is seen by the large number of products competing. Although the five products in Tables 1GA & 1GB account for almost three-fourths of the product class, there were 78 products on the market in 1965. 1 Tine Pharmacological Basis of Therapeutics-Second Edition Goodman and Gilman. PAGENO="0254" 1804 COMPETITIVE PROBLEMS IN THE DRUG INDUSTRY TABLE 16-A--SEDATIVES-BARBITURATE lPercent share of market based on numbers of prescriptions written, 1956-651 Product 1956 1957 1958 1959 1960 1961 1962 1963 1964 1965 1 24. 1 24. 5 26. 4 21. 5 22. 0 22. 2 22. 9 19. 7 21. 5 20. 9 2 15. 1 15. 2 13. 4 13. 8 15. 6 16. 5 15. 5 17. 3 16. 8 14. 9 3 13.2 12.8 15.2 14.7 19.3 17.7 17.4 15.1 14.2 15.4 4 10.7 12.0 11.8 14.2 14.5 13.7 10.2 12.5 14.2 13.2 5 5.8 5.5 7.3 8.6 9.4 9.6 6.8 7.8 8.0 9.7 Cumulative market share of above products 68. 9 70. 0 74. 1 72. 8 80. 8 79. 7 72. 8 72. 4 74. 7 74. 1 All others 31. 1 30. 0 25. 9 27. 2 19. 2 20. 3 27. 2 27. 6 25. 3 25. 9 Category total 100. 0 100. 0 100. 0 100. 0 100. 0 100. 0 100. 0 100. 0 100. 0 100. 0 TABLE 16-B-SEDATIVES, BARBITURATE [Percent share of market based on dollar volume, 1956-65] Product 1956 1957 1958 1959 1960 1961 1962 1963 1964 1965 1 16.7 18.1 20.1 15.9 15.7 16.9 17.3 15.1 16.5 16.2 2 14. 7 15. 5 14. 1 13. 7 15. 8 16. 4 15. 0 16. 8 16. 4 14. 1 3 13. 2 13. 1 15. 6 14. 7 19. 5 18. 0 17. 5 15. 3 14. 1 15. 4 4 10. 6 12. 9 13. 4 16. 9 16. 8 16. 3 12. 6 15. 4 17. 3 16. 1 5 5.8 6.3 7.8 9.0 10.1 10.1 7.1 8.0 8.2 9.8 Cumulative market share of above products 61. 0 65. 9 71. 0 70. 2 77. 9 77. 7 69. 5 70. 6 72. 5 71. 6 All others 39. 0 34. 1 29. 0 29. 8 22. 1 22. 3 30. 5 29.4 27. 5 28. 4 Category total 100. 0 100. 0 100. 0 100. 0 100. 0 100. 0 100. 0 100. 0 100. 0 100. 0 XIX. SULFONAMIDES Products in this group are used to treat a broad range of infections. Some prod- ucts are specific for either systemic or urinary tract infections while others treat both types. Over the ten year time span the number of prescriptions for sulfon- amides has increased 25% while the dollar volume has increased 72%. Sulfon- amides represented 2.85% of prescriptions written in 1965 as compared to 4.0% in 1956. The principal products used in 1956 continue to be major products in 1965. How- ever, approximately 76% of prescriptions written are for 14 out of 173 products audited in 1965 (see Tables 17A and 17B). PAGENO="0255" C CD CD p C CD I. ~0 3 3 -1 (0 C C 0) C) (0 N~ 0) ()~ 0)0) - C-Na Q~ - (0(0(000(0 ~ N) (0 N) N) C) ~J 3 2 0> -0~ C/) 0 0 ci ci 02 I, PAGENO="0256" PAGENO="0257" COMPETITIVE PROBLEMS IN THE DRUG INDUSTRY THURSDAY, JANUARY 18, 1968 U.S. SENATE, MONOPOLY SUBCOMMITTEE OF THE SELECT COMMITTEE ON SMALL BUSINESS, Washington, D.C. The subcommittee met, pursuant to notice, at 10 a.m., in room 318, Old Senate Office Building, Senator Gaylord P. Nelson (chairman of the subcommittee) presiding. Present: Senators Nelson, Javits, and Scott. Also present: Benjamin Gordon, staff economist; James H. Gross- man, minority counsel; Susan H. Hewman, research assistant; and William B. Cherkasky, legislative director, staff of Senator Nelson. Senator NELSON. We will open the hearing of the Subcommittee on Monopoly of the Small Business Committee. Our witness this morning is Dr. Willard F. Mueller, Chief Econ- omist and Director of the Bureau of Economics of the Federal Trade Commission. I have read your statement, Dr. Mueller, and I find it commendable and lucid compared to the economic presentation given last month by the economists sponsored by the Pharmaceutical Manufacturers Asso- ciation, which, quite frankly, I did not understand very well. Will you introduce your associates for the record, and, if, at anytime, one of them wants to make a comment, will he just identify himself so we can keep the record straight. I appreciate very much your taking the time to come here this morning and the obvious extensive work you have put into your statement. You may proceed to present it in any way you see fit. I assume you have no objection to us interrupting you with questions from time to time. Dr. MUELLER. Not at all, Mr. Chairman. Senator NELSON. And do you have a biographical statement? Dr. MUELLER. Yes; I have a sketch which I can give to the reporter or read into the record. Senator NELSON. Read it into the record for us, please. STATEMENT OF DR. WILLARD P. MUELLER, CHIEF ECONOMIST AND DIRECTOR, BUREAU OF ECONOMICS, FEDERAL TRADE COMMIS- SION, WASHINGTON, D.C.; ACCOMPANIED BY DR. RUSSELL C. PARKER, ASSISTANT DIRECTOR; AND WILLIAM H. KELLY, MEM- BER, BUREAU OP ECONOMICS Dr. MUELLER. First, I would like to introduce my two colleagues. To my immediate left is Dr. Russell C. Parker, assistant to the Direc- 1807 81-280-68-pt. 5-17 PAGENO="0258" 1808 COMPETITIVE PROBLEMS IN THE DRUG INDUSTRY tor of the Bureau of Economics, and next to him is Mr. William H. Kelly of our staff. I am Chief Economist and Director of the Bureau of Economics of the Federal Trade Commission. Prior to my current appointment, I was Chief Economist to the Select Committee on Small Business of the House of Representatives. Previous to that assignment, I was professor at the Uiiiversity of Wisconsin, from 1957 to 1961. I was assistant professor at the University of California from 1954 to 1957. I also have taught on a part-time `basis at American University; served as a visiting professor at Michigan State University, and I am cur- rently a part-time staff member of the Department of Economics, University of Maryland. It is a privilege and a pleasure to appear before this subcommittee. My appearance today is in response to the request of your Chairman that I submit testimony on the subject of profits in the drug industry, as well as present an independent analysis of the study "Risk and Return in American Industry-an Econometric Analysis," presented to this committee on December 19, 1967.' The study was sponsored by the Pharmaceutical Manufacturers Association and prepared by Gordon R. Conrad and Irving H. Plotkin of Arthur D. Little, Inc., in consultation with Prof. Jesse W. Markham and Prof. P. J. Cootner. Hereafter we shall refer to the study as the Conrad-Plot.kin study. Before turning to the study, we shall first place in perspective the profits of drug manufacturers by comparing them with those earned by business enterprises in other American industries. Figure 1 shows for 1966 the average rate of return on stockholders investments of leading firms in 22 important American manufactur- ing industries. Profit rates of leading drug manufacturers exceeded those of large firms in the 21 other industries. In fact, drug industry profits were twice as great as one-third of the remaining industries; were 44 percent or more above those of all but four other industries; and they exceeded even such traditionally high profit industries as motor vehicles and computing machines. Nor was 1966 an exceptional year. Table 1 compares over the period 1950-66, the average profits of large drug companies and large com- panies in the 22 industries shown in figure 1. Several points are of special interest. First, in the early years, 1950-55, average drug com- pany profits were about equal to or somewhat above the average of other large manufacturers. Second, beginning in 1956, however, aver- age profit rates of drug companies were well above the average of other large companies. Finally, since 1956 drug companies have con- sistently ranked either first or second among all large manufacturing industries. This indicates that during the last decade large drug companies have occupied an especially advantaged position relative to large companies in other American industries. Table 2 summarizes profit data of all drug companies and all manufacturing companies for the period of 1956-67. Over the period covered it shows essentially the same picture as table 1. Since 1956 drug manufacturers have failed to occupy first place in only one year. 1 The study, "Risk a.nd Return in American Industry-an Econometric Analysis," by G. R. Conrad and I. H. Plotkin, begins at p. 1746, supra. PAGENO="0259" COMPETITIVE PROBLEMS IN THE DRUG INDUSTRY 1809 This pattern of persistently high profits indicates that large drug companies occupy a unique position in the American economy. And they appear to have become increasingly unique since the mid-1950's. Before turning to an analysis of the Conrad-Plotkin-Markham- Cootner explanation of high profits in the drug industry, I shall briefly review what appears to be a virtual consensus of opinion among researchers in the field of industrial organization concerning the causes of high profits of drug manufacturers. In this connection I would like to emphasize the crucial role which congressional hearings have played in developing the facts necessary for scholars to study the organization and performance of the drug industry. Prior to the Kefauver drug hearings on administered prices in 1959-61, not a single article concerning the American pharmaceuti- cal industry had appeared in a professional economic journal. Senator NELSON. Are you referring to any article of any kind or just articles in economic publications? Dr. MUELLER. I am referring to articles on the structure and or- ganization of the drug industry. This observation is based on the recent study that I cited in a footnote by Professor Walker of the University of Indiana, who surveyed the literature.1 Since the first such article appeared in 1962, there has `been a growing volume of research literature on the subject, `all of which has drawn heavily on the Kefauver `and subsequent congressional hearings. The facts developed by this committee have made another enormous con- tribution to the fund of knowledge concerning the drug industry. I am confident that `scholars will be sifting and winnowing the facts for years. The preponderance of economic evidence argues that the persistently high profits of the drug industry are the result of the absence of effec- tive price competition in the sale of many products. Price competition in drugs is ineffective for several reasons. Concentration in the produc- tion of many drugs is high because of the patent privilege. And even where there are relatively many sellers, as well as many potential sellers-for example, in the case of unpatented drugs sold under generic names-effective price competition often is muted by vast ad- vertising, promotion, and other selling efforts which differentiates in the minds of consumers the products of the largest drug manufactur- ers selling under their own brand or trade names from those of other manufacturers. Hence, manufacturers selling chemically identical drugs under generic names frequently have difficulty in selling them at any price. Senator NELSON. Are you referring to' the retail market particularly in this instance? Dr. MUELLER. That is correct; the sales of druggists in the pre- scription market. Senator NELSON. We do find from the testimony, as I am sure you have noted, that generic companies or generic drugs are competing effectively in bids to nonprofit institutions and to the Federal Govern- ment, and you will also find testimony that brand name drugs drop substantially in their price when they bid in this competitive area. 1 The complete prepared statement submitted by ~" Dr. Mueller begins at p. 1824, infra. PAGENO="0260" 1810 COMPETITIVE PROBLEMS IN THE DRUG INDUSTRY Dr. MUELLER. That is correct. The evidence shows that when the buyers are well informed such as, hopefully, the purchasers for hospitals and the U.S. Government, they nsually purchase on a generic basis rather than a brand basis, or when they do buy on a brand basis they purchase the brand selling at the lowest price rather tha.n simply the most popular one or well known one. Senator NELSON. So, you are explaining, I take it, the difference between the price situation in the retail market, where you will find a brand name product sold to the pharmacist at the highest price charged in the country, and the price situation in the institutional market where the same company bidding its brand name product to the Government or some nonprofit institution will offer it at a substantially lower price; is that correct? Dr. MUELLER. That is correct. I think every person who has studied the record of congressional hearings where this sort of evidence has been developed and has been explored and subjected to extensive com- ment by informed people outside and within the drug industry, as well as medical doctors, has come essentially to this conclusion: namely, that there are very substantial differences between prices for products purchased by hospitals and other large buyers and prices paid in drug- stores for the identical product. Senator NELSON. And I understand it to be your conclusion that in general that is because you will very frequently find competition in bids to the Government or bids to nonprofit institutions while. there is much less, or in some cases not any, competition at the retail level and the price charged by the company to the retailer and the wholesaler is not a competitive one; is that correct? Dr. MUELLER. And the key factor for generating the competition is the prese.nce of an informed buyer in one case and an uninformed one in the other, and, as a result, this broadens the market. People who have a product to sell that is physically identical to that of other sellers have an opportunity to bid, and, consequently, you broaden the market, which means broadening the opportunity for competition to work. Senator NELSON. You referred to an informed buyer versus an unin- formed buyer. That factor is at work in purchases by nonprofit organi- zations or government. If the purchasing agent is informed and takes competitive bids and has the personnel to evaluate the quality of the drugs, you have an informed buyer and you have competition and a lower price. In the same institution, nonprofit or government, if the buyer is not informed the price is back up high again. Now, you are not suggesting that the question of an informed buyer makes any difference in the retail price, are you? In other words, is it not correct that the wholesaler or the retailer is charged a certain price, and although quantity buying accounts for some differential, an informed purchaser for a large drug chain is still being charged a high price, the same price as is paid by an uninformed purchaser for another drug chain; wouldn't that be correct? Dr. MUELLER. That is correct. As I interpreted the evidence that had been developed on this point, the only opportunity of getting a lower price at this level from the ultimate consumer's standpoint, namely, the patient. of a doctor, is if his doctor orders the drug on a generic basis. PAGENO="0261" COMPETITIVE PROBLEMS IN THE DRUG INDUSTRY 1811 And then, of course, we may be able to get. the product at the retail level at a lower price than for the branded product. But on your specific point as to whether the branded product sells at the same price or a higher price through drug stores in these corn- parative bids, I agree with your interpretation of the facts. Mr. GORDON. I think there may be another point, and that is that the purchasing agent for the ordinary consumer is not the one who pays~ In other words, the doctor does not pay for the medicine. He does the purchasing. He orders the drug, and the patient, the consumer, does not have much of a choice. Dr. MUELLER. I have heard many people speak to that point, includ- ing doctors, who say that some doctors are more discriminating in this respect than others. And I think increasingly, with publicity concern- ing the alternatives available, consumers often ask their doctors: "What is this generic business all about?" So this, I suspect, makes the doctor a little bit more price conscious. And I think this is one of the benefits of publicity concerning the alternatives available to con- sumers. It is sort of a counter to the advertising which promotes brand names so decisively in the minds of consumers today. Senator NELSON. Go ahead. Dr. MUELLER. The resulting wide price spread between advertised and generic drugs often applies to unpatented as well as patented drugs. These factors shelter the leading concerns, or any concern with a highly differentiated drug product, from effective price competition. This explanat ion of high drug profits is not novel. Nearly all re- searchers who have analyzed the drug industry in detail have come essentially to the same conclusion. Similarly, empirical studies which have cut across many industries have identified the elements of market structure that are primarily responsible for high noncompetitive profits. These elements are high seller concentration, high barriers to entry, and product differentiation. Mr. GORDON. Dr. Mueller, for the record, will you please explain exactly what product differentiation is? Dr. MUELLER. I had anticipated this question. I know that Senator Nelson was an undergraduate student in economics but now is primar- ily a lawyer-or, I should say, primarily a statesman. Senator NELSON. You may make that~ statement here, but no place else. Dr. MUELLER. I am thinking of back in Wisconsin. There is a definition of product differentiation which appears in the Antitrust Law and Economic Review which, presumably, is written for lawyers. I am not sure if this is an improvement on one that I could make for students. That is the way they define it-they say: Product differentiation refers to the distinguishing of substitute products from one another by advertising and the like. Whereas buyers of homogenous products regard the output of any particular seller as identical in all respects to that of all other producers of that product, the seller of a "differentiated" product enjoys a favored position over its rivals, in that the buyers consider it a superior product and are willing to pay a "premium" price for it rather than accept the substitutes offered by those rivals. Since new entrants must frequently accept a lower price than established firms are able to get for a product of equal quality and cost, this disadvantage is said to constitute a "barrier to entry," one that permits estab- lished firms to charge a supercompetitive price without attracting new entry. PAGENO="0262" 1812 COMPETITIVE PROBLEMS IN THE DRUG INDUSTRY And, as I said, I will later have some things to say about product differentiation created by advertising or promotion, which in effect creates a barrier to entry facing potential competitors. This is what this definition is supposed to explain. Senator NELSON. `When you say "product differentiation," are you referring to the same chemical compounds which are differentiated by advertising and other means? In other words, you are talking about a drug, a basic compound, which a company differentiates from the same drug put out by another firm simply by the capsule it is in and the advertising. You are not talking about two different drugs? Dr. MUELLER. No, sir, they are chemically identical. Senator NELSON. So, you would have two drugs that have the same chemical composition but are differentiated in one way or another by trade name and advertising and that sort of thing? Dr. MUELLER. That is correct. People have an impression that there is a difference. It is a psychological thing. You believe it is different, and, therefore, you behave differently, even though it is physically identical. Senator NELSON. So, a drug such as prednisone, which is the generic name, manufactured by a number of companies. One of them gives it the brand name "Paracort," and one of them "Meticorten." They are the same chemical compound, but are differentiated products in the marketplace; is that what you are saying? Dr. MUELLER. That is correct. Senator NELSON. And they have the same purpose and the same ob- jective in the treatment of the patient? Dr. MUELLER. Yes. I am assuming that they are physically or chem- ically identical, and yet, because of this differentiation in the minds of the consumer you are able to have a unique demand for them which permits you to get a higher price for one than the other. Senator NELSON. So, you are saying that the difference, when you use the word "differentiation," is in the minds of the consumers and not in the chemical compOsition of the product itself? Dr. MUELLER. That is correct. That is the way I am using it here, as I think it applies most accurately in the drug industry. In some in- stances, of course, there may be some physical difference in products, and forms differentiate them further through advertising and the like. Senator NELSON. In the method by which they are compounded or by the coating on the tablet, and that sort of thing? Dr. MUELLER. I was thinking of products other than drugs. In the case of soaps, they are very similar often, but they may be slightly different physically plus having `blue in them, and so forth, which is similar to putting a different coating on the pill. But the basic idea is that they are essentially identical products in an economic sense because of the underlying physical a.nd chemical characteristics, but that, because of this differentiation created by promotion, and so forth, the consumer thinks they are different. Senator NELSON. Soap would be one of the items that you refer to as homogeneous products; is that it? Dr. MUELLER. it is very much differentiated in the minds of con- sumers. You pay a tremendous difference in the price of Tide, say, and Safeway's detergent. And the only pdint I was making here, and I PAGENO="0263" COMPETITIVE PROBLEMS IN THE DRUG INDUSTRY 1813 probably introduced an element of confusion by doing so, is that in some cases there are slight physical differences as well as this psycho- logical difference created in the mind of the consumer; whereas, in ethical drugs most frequently the products are physically and chem- ically identical. Senator NELSON. Go ahead. Dr. MUTELLER. Perhaps the most pervasive factor blocking effective price competition in drugs is the presence Of substantial product differ- entiation of branded drug items. A recent econometric study demon- strates that advertising and promotion-created barriers to entry are the single most important explanation for differences in profit rates in American industry. The drug industry was among the industries in- cluded in that study. Professor Seymour Harris of Harvard University pretty well sum- marizes the conclusions of academic scholars concerning the organiza- tion and performance of the drug industry: Many are concerned that at industry which comes close to being a public utility achieves the highest profits in relation to sales and investment of any industry; is highly concentrated in its control of the market; reveals serious monopolistic trends; increases the ~ost to consumers by differentiating the product at a dizzy pace, with the differentiated product usually similar to or identical with exist- ing products; and greatly inflates the cost through record expenditures on selling. The competition among companies to overwhelm the doctors by repetitious and often misleading advertising, and a failure to give as much publicity to the bad side effects as to the immediate beneficial effects, are unfortunate. Thus competi- tion forces even highly moral firms to become less ethical in their behavior. In the drug industry the relation of labor to total costs is minimal; and like the soap and tobacco industries, using similar selling techniques, their relation of labor to value added is a minimum-selling expenditures and profits are the large items in gross receipts. The cost of drugs is too high. I say this, though I am aware that the research contributions of the industry are important and that the lives saved, the suffer- ing averated, `and the acceleration of recoveries `are worth more than the $4 billion spent on drugs. But the cost could be substantially less. Does this mean that risk plays no role in high drug profits? Not necessarily. Although these high profits can be explained by the struc- tural characteristics of the industry-namely, high concentration, high entry barriers, and a high degree of product differentiation-it is conceivable that risk also played some part. Conrad, Plotkin, Mark- ham and Cootner testified that they believed high drug prices and profits were due primarily to uniquely~ high risks assumed by large drug manufacturers, and that the Conrad-Plotkin study measured the magnitude of this risk. The relevant question, of course, is how much of total profits can he attributed to the risk factor. Let us therefore turn to the em~irical evidence on this subject. Mr. GROSSMAN. Dr. Mueller, before you do that, I wonder if I could take you back to Professor Harris' quote? In the first line he says: "Many are concerned that an industry which comes close to being a public utility," et cetera. * How do you feel about that? Do you think the drug industry should be treated as a public utility? Is it close to being so? Should it `be? Dr. MUELLER. I think what he is referring to is that one character- istic of a public utility is that it is providing a service to the consumer PAGENO="0264" 1814 COMPETITIVE PROBLEMS IN THE DRUG INDUSTRY which is especially unique, is required for matters of safety and public welfare. And this, I think is the definition he is using. Mr. GROSSMAN. When Mr. Squibb testified he said that this was a possible end result if the drug industry did not do something on its own, that we would eventually treat them like public utilities. Do you think the time to do this is now? Dr. MUELLER. I personally am among those who are very reluctant to see more direct regulation of American business of the kind we have in the railroads and transportation, and so on. It presents a great many difficulties. And my personal inclination is to try to make competition work wherever it is at all possible, and I would hope that there are solutions to this problem short of the public utility approach. Mr. GROsSMAN. Thank you. Senator NELsON. Is it your opinion that if there were-whatever the word means-effective competition at the retail level, that this would basically resolve the problem that we are talking about here? I realize that this involves a question of fact and all kinds of other things. Dr. MUELLER. If we could achieve effective competition, I think it would resolve the problem. This gets into the question of how you achieve it and whether it can be achieved. Mr. GRoss~rAx. I wonder if you could explain what you mean by high seller concentration? Dr. MUELLER. The fact that there are very few sellers of some drugs. Mr. GROSSMAN. Thank YOU. Senator NELSON. I believe you address yourself to that particular question a. little later on, with some comparisons with the auto industry and so forth. Dr. MUELLER. In terms of seller concentration? Senator NELSON. Yes. Dr. MUELLER. No, sir. I think earlier in my summary of what people have said of the industry I made some references to concentration. Senator NELSON. I meant lack of competition, because one or two or three companies may be the only producers of a product. Dr. MUELLER. That is correct.. I might ask, with your permission, that the fairly extensive foot- notes and so on which I ha.ve in my prepared statement be incorporated in the record. I tried to cut down on my statement by putting a good deal of the underlying data in the footnotes and references. Senator NELSON. The full statement, including the footnotes, will be included in the record, and at anytime the footnotes seem to be irn- portant to an understanding or appreciation of your ba.sic text, you may read them or explain them extemporaneously or however you see fit. Dr. MUELLER. Thank you. One way of gaining insight into the question of "risk" is to look at what investment analysts tell investors about the drug industry. This may seem to be a rather homespun approach to the problem. but after all it is what investors believe about an industry that determines investment decisions. A perusal of studies by investment analysts indicates that they generally advise investors that the drug industry is a rapid growth, PAGENO="0265" COMPETITIVE PROBLEMS IN THE DRUG INDUSTRY 1815 high profit industry where established firms hold a strong position relative to small companies and potential entrants. The industry is frequently described as "depression resistant" because, as one analyst put it, "Illness is no respector of business cycles, and Americans have shown that they will buy the medicine they need regardless of eco- nomic conditions." These `characteristics are considered to make drug stocks good "defensive" investments. As a result, drug stocks sell at relatively high price-earnings ratios, `indicating that investors are confident of a high future payout. A Standard & Poor's analysis of drugs summed up the factors affecting drug stocks as an investment as follows: Shares of drug equities have historically sold at relatively high price-earnings ratios, owing to the industry's recession-resistant characteristics, its above- average earnings growth rate, and its strong underlying position. Moreover, it is difficult to enter the drug field. Investment analysts generally emphasize that the high earnings of drug companies make drug stocks a good buy. This is not to imply, of course, that investment analysts view the `industry as completely riskiess. The staff of Moody's Investors Service, after explaining a number of reasons why drug stocks were a good investment, stated: The drug industry cannot be risk-free. The postwar years have seen periods of slowdown, and individual companies have' suffered temporary setbacks. The causes have been many. Competition has led to price-cutting in popular prod- ucts, such as penicillin, where capacity has been overexpanded. Occasionally, a profitable new drug is found to have unsuspected and unfavorable side effects. This, however, is less of a problem than product obsolescence or the expiration of patents on major drugs that have been exclusive with one company. In recent years, government regulation has been tightened at the drugmaker's expense. Finally, the ebb and flow of respiratory diseases often causes sharp fluctuations in drug sales. But after enumerating the above points, the Moody's analysts continued: The impact of such development has caused only temporary deviations in a growth curve that has pointed strongly upward. Thus, while the drug industry faces uncertainties and problems, from an investor's standpoint these "risks" apparently are no greater than those found in many other industries. On the contrary, drugs are considered a sound growth investment. Investment analysts frequently make mention of the fact that drug profits may be adversely affected by factors that will increase price competition and thereby erode high profits. The most frequent reference of this sort is the observation that anything threatening to increase the use of generic drugs as opposed to brand name drugs threatens high profits. An analysis by "Value Line" of the possible effects of medicare on drug profits is typical of investment analysts' views on the subject. After explaining that medicare very probably would increase drug sales, "Value Line" concluded that drug profits would not go up by a corresponding amount because: Hospitals and institutions usually, wherever possible, buy generic name drugs rather than brand names in order to reduce costs. The most profitable business for the drug manufacturers is that which comes through drugstores, where drugs are prescribed on a brand name basis. [Emphasis added.] PAGENO="0266" 1816 COMPETITIVE PROBLEMS IN THE DRUG INDUSTRY Senator NELSON. May I interrupt, just for clarification of a point? You state that the hospitals and institutions whenever possible buy generic drugs rather than brand names in order to reduce cost. Isn't it the actual situation that hospitals and the Government use the formulary system when they purchase drugs. They take bids on a generic basis, and then the contracts are made with companies that produce only generic drugs and companies that produce both brand names and generics. Brand name companies bid their brand, but here is where you find the example about the effect of competition on prices: They may very frequently win; they may very frequently underbid the generic company in the institutional field, that is, Govern- ment and hospitals, and therefore the trade name drug may very fre- quently be used. But here is where you see the contrast between the price that is charged by the brand named company in the competitive field versus the price they charged in the retail field. In the retail mar- ket they can hold their market price against a lower priced generic drug, hut they can't hold it against a generic drug in a Government purchase unless they substantially reduce the price when they bid. Is that a correct statement? Dr. MUELLER. That is correct. I think investment analysts view the end effect of this process. This observation, of course, is concerned with how medicare might affect competition because of the increasing use of generic drugs; it is not explaining profits associated with risks. In sum, there is no reason to conclude, on the basis of advice being given investors by investment analysts, that the drug industry is a uniquely risky industry. On the contrary, the generally glowing reports of investment analysts suggest. t.ha.t large dnig companies should have little difficulty obtaining adequate capital should they choose to go into the market for it. Actually, however, their profits are so large that drug companies seldom need go to the capital market for equity capital. And there is no reason to expect that drug companies would have difficulty getting adequate capital even if they enjoyed profit rates comparable to most other American industries. But perhaps this is a too prosaic approach to the problem. Let. us, therefore, turn to the Conrad-Plotkin-Markham- Cootner "economet- ric" explanation of high profits in the drug industry. My comments today concerning the Conrad-Plotkin analysis will be limited to an evaluation of the testimony presented to this conunittee last month. Arthur D. Little, Inc. has promised to provide us with the underlying data used in their analysis. Mr. GROSSMAN. Dr. Mueller, before you go on: At what point do you think investors would not invest? How low do we have to go in market percentages before we lose the investors? Dr. MUELLER. In the drug industry? Mr. GROSSMAN. Yes. Dr. MUELLER. I am not sure just where that level is. There are com- panies that have had fairly low earning experiences that have been able to obtain capital funds. I think you will recall Dr. Whitney's statement that in the early 1950's drug companies were investing at a greater rate than other companies, by a. 3-to-i ratio, or something like that. Actually, at that PAGENO="0267" COMPETITIVE PROBLEMS IN THE DRUG INDUSTRY 1817 time drug profits on the average were very similar to the average of all American industry. So, it is difficult to know precisely where this level is. I have little doubt in my own mind that it is considerably below the 21 percent level the large companies have earned in recent years, and there is not any persuasive evidence to argue that they would have difficulty getting capital if their profits were the same as most other American industries. Mr. GROSSMAN. In the same line of questioning, what do you think is a fair return for the drug industry? We are talking in terms of social responsibilities. What would you say would be a fair profit? We have had a lot of discussion about the percentage return. Is there any specific percentage? In other words, should we say: You can't go over 15 percent? Isn't this all rather vague? Dr. MUELLER. Yes, unless you are going to take a public utility approach which I find, personally, repugnant. Until we try other methods, I would not say that any particular percentage is appro- priate. I would like to see the industry exposed more to the breezes of price competition and see what reasonable rate of return would be generated by competitive market forces, rather than to set arbitrarily some spec~flc rate. Senator NELSON. Is there any evidence that during the period you refer to, 1950-55, the drug industry was receiving profits fairly com- parable to the average in all industries. Did the industry experience any difficulty in attracting investment capital if they needed it? Dr. MUELLER. I just have not had an opportunity to look into that point. I think it would be interesting to see what the experience was. The observation I make, though, is that there are many examples of companies that were making profits by which this 21 percent standard would be very modest~, in fact quite mediocre profits, or close to those earned by the rest of the American industry on the average. Yet they have been able to generate capital on their own or borrow it and grow effectively. So, simply asserting that because someone is earning a high rate of return that this must somehow demonstrate that they must earn such a return in order to continue to grow is to just define the problem out of existence. And I am afraid that is what Professor Whitney might have been doing in his exchanges with you; at least that was my impression as I read the testimony last night. Mr. GROSSMAN. Do you give any stock to his point about the rail- roads, how they prospered for 30 years? Dr. MUELLER. I think he is factually correct, that they did profit for 30 years. Mr. GROSSMAN. Then, I assume that there is a risk, and we will have to wait for 50 years before we can ever know? Dr. MUELLER. I would not wait on the problems in the drug indus- try for 50 years to find out whether or not Professor Whitney's theory is correct. With your permission, we will provide a brief supplemental mem- orandum to the committee should we have any additional observations after reviewing these data. PAGENO="0268" 1818 COMPETITIVE PROBLEMS IN THE DRUG INDUSTRY Senator NELSON. If we thought it would `be of value, would YOU be prepared to submit the supplemental data to the committee at a sub- sequent date? Dr. MUELLER. Certainly. I would be happy to.' You will recall that Conrad and Plotkin attempted to test the hy- pothesis that the level of an industry's profit rate is positively cor- related with the degree of risk faced by firms withrn the industry. In other words, the more risky an industry, the higher its average profits. The concept of risk in investment decisionmaking theory refers to situations where it is impossible to predict with certainty the outcome of particular economic events. The presence of uncertainty is assumed to affect investors' decisions. A common assumption is that investors must be paid a "risk premium" if they have an aversion to assuming risks. "Risk aversion" has been an underlying assumption in a number of recent. theoretical works, particularly in the areas of portfolio selec- tion and monetary theory. However, the assumption of "risk aver- sion" is not a universal economic law. One need only view the crowds at the racetrack paying for the privilege of taking a gamble to infer that some persons regard risktaking as furnishing positive rather than negative satisfaction. These individuals may be viewed as "risk lovers." As a group, these risk lovers lose money at the racetrack. This is also the case with persons gambling in commodity futures markets. In order to explain why individuals will both purchase insurance to guard against large losses and undertake gambles with remote possi- bilities of achieving high returns, Friedman and Savage, have argued that some persons regard risk-taking as furnishing positive rather This, in a. nutshell, is what risk aversion theory is all about. But note two important points. First, the size of the risk premium is an empirical question. The theory tells us nothing about. the amount. of the premium, nor even whether it is positive or negative. Second, central to the hypothesis that is necessary to offer `a. positive premium to in- vestors in order `to attract. adequate co~pital into a risky industry is the idea that. risk may cause firms to incur losses, as well as to enjoy ab- normally high profit rewards. Hence, risky industries would be char- acterize.d by the presence of both firms with `abnormally high profits and firms with abnormally low profits. It would be inconsistent with risk theory if nearly all firms in an industry made very high profits and few or none ever suffered losses. The Conrad-Plotkin measure of risk misses this point. Risk is quan- tified by Conrad and Plotkin by measuring the variance of individual companies' rates of return about the industry average in a given year and computing a simple average of these values for the 16-year period 1950-65. This measure assmues that the greater the variation in the profit. rates of firms about the industry `average, the riskier the indus- try. The chief conceptual shortcoming of this measure is that it does not necessarily tell us anything about the probability `of incurring losses. In truth, using this measure a.n industry may be defined as risky even though all firms in it earn excessively high profits; on the other hand, this measure may define an industry as having very low risk even though all firms are making little or no profit. An example will illustrate this point. By the Conrad-Plot.kin measure, the drug I See supplemental ~taternent beginning at p. 184S. lnfra. PAGENO="0269" COMPETITIVE PROBLEMS IN THE DRUG INDUSTRY 1819 industry is a high risk industry and the aluminum industry is a low risk industry. Conrad and Plotkin's estimates of the average rate of return, standard deviation, and variance for the two industries are as follows, and I show that in my statement in tabular form. The drug companies in their sample experienced an average rate of return of 17.5 percent over the period 1950-65. The standard devia- tion in profits around `this average was 8.6 percent. This means that the profit ratio of roughly two-thirds of the companies in the industry fell in `the range, 8.9 percent to 26.1 percent. On the other hand; `the average rate of return of a group of alumi- num companies was 7.8 percent, with `a standard deviation of 1.3 per- cent. This means that two-thirds of the time aluminum company profits fell in the range, 6.5 percent to 9.1 percent. Thus, according `to Conrad and Plotkin, `the `drug industry is riskier than the aluminum industry because `of the greater standard deviation in the profit rates of drug manufacturers. Just what do these facts concerning the variation in profits tell an investor about the relative profit expecta~tions in these two industries? They say, in effect, that there is a 2-to-i chance that profit rates in the drug industry will fall in `a range from 8.9 percent `to 26.1 per- cent, whereas there is a 2-to-i chance in the minimum industry that profit rates will fall in a range from 6.5 percent to 9.1 percent. Can anyone seriously argue that investOrs would prefer to place new capital in the aluminum industry rath'er than in the drug industry? The only risk that the aluminum investor is saved from is `the h'igh probability that aluminum companies will earn less than 9 percent- there is only 1 chance in 6 of getting more than 9 percent. On the other hand, the risk t'he investor in the drug industry faces is that the chances are poor that drug companies will earn a rate of return as low as the `average return in the aluminum industry. In fact, there is only 1 chance in 6 that they will earn a rate of return of below 9 percent, whereas there are 4' chances out of 6 that they will earn between 9 and 26 percent, and 1 chance in 6 that they will earn over 26 percent. In other words, five-sixths of the time the drug companies would be well above the `aluminum companies' average return. Clearly, then, it is nonsense to infer from the Conrad-Plotkin variance measure of risk `that the drug industry is riskier than the aluminum industry in `terms of attracting new capital. Losses, or even low profits, are practically un'heard of among large drug companies. In this respect the `drug industry is practically unique among important American industries. Figure 2 shows for 22 maj or industries the percent of the time the eight largest companies fell in various profit rate categories during the period 1954-66. Large drug companies not only earned a higher return than any other of the major manufacturing industries shown, but none of the drug companies ever experienced losses during the period, nor did any com- pames experience profit rates below 5 percent. Only two other indus- tries enjoyed this distinction, petroleum refining and cigarettes. I might say parenthetically that I used' this 5 percent figure because it represents `the approximate rate that someone would receive when purchasing very `secure bonds during this period. So, in effect, this r the upper level of very secure investn'ients. And when you receive PAGENO="0270" 1820 COMPETITIVE PROBLEMS IN THE DRUG INDUSTRY below 5 percent, you are receiving less than you would earn in sort of an insured, risk-free investment. Companies in most manufacturing industries had profit rates rang- ing between 5 percent and 15 percent. By considering this range to `represent a sort of "norm," we can visualize the extent to which large `drug manufacturers departed from it. Over the period 1954-66, the eight largest drug manufacturers were in this range about 25 percent of the time and none ever fell below it. On the other hand, 75 percent of the time the leading drug ma.nufacturers earned profits exceeding 15 percent, and fully 17 percent of the time they had profit rates exceeding 25 percent. No other industry matched drugs in the frequency with which companies had profit rates exceeding 15 percent. In only three other industries-motor vehicles, computing machines, and aircraft-d~d large firms have profit rates exceeding 15 percent more than 50 percent of the time. On the other hand, all but four of the remaining 18 industries had profit rates exceeding 15 percent less `than one-fourth of the time. Finally, some companies in all but two of the 21 industries outside drugs earned below 5 percent at least part of the time. These contrasting patterns cast serious doubt on the proposition that large drug companies face a serious risk of incurring losses. I might add parenthetically that if we take the 29 drug companies used in the Conrad-Plotkin study, their profit performance was not a great deal different than the eight used in that chart. The 29 com- panies earned below 5 percent only 0.4 percent of the time. If we had used as many companies, say 29 companies, in those other 21 industries, the red area shown on the right, these other industries- in other words, the below 5 percent area-would be considerably wider. Figure 3 illustrates the profit experience of leading drug companies in each year during the 1954-66 period. Most importantly, it shows a pattern of persistently high profits. None of these large companies *earned below 5 percent in any year, and in only 4 years did any `company earn between 5 percent and 10 percent; 3 of these years ~were in the beginning of the period-1954, 1955, and 1956. Since then, in only 1962 did a company earn below 10 percent. And in the last 3 years only a small percentage earned below 15 percent. This actual profit experience seems to fly in the face of the Conrad- Plotkin-Markham-Cootner inference that drug manufacturing is a uniquely risky business. The explanation, of course, is to be found in their definition of risk. Using a different definition of risk, Dr. Irving N. Fisher and Dr. George R Hall of the Rand Corp. concluded 4hat risk accounts for a very small portion of the high profits of `drug companies. The findings are shown in table 3. They show that for the period 1959-64 drug companies earned an average return `of 18.32 percent. Fisher and Hall attributed 1.68 percent of this to risk. They concluded that the "risk premiums" for drugs are "very low" and that the explanation for high profits "must be sought in factors other than risk." 1 Mr. GORDON. Mr. Plotkin testified here last month. I quote from his Ttestimony on page 2743 of the transcript: `The report "Risk and Corporate Rates of Return," by Dr. I. N. Fisher and Dr. G. N. Hall, appears as Appendix II, p. 2120, infra. PAGENO="0271" COMPETITIVE PROBLEMS IN THE DRUG INDUSTRY 1821 Management, it Is our contention, forms its risk expectations not solely on the past profit history, of its own company, but also on the diversity of histories that other companies undertaking similar ventures, to use the term again, other companies within its industry undertake. Would you comment on that, please? Dr. MUELLER. Well, I think that is essentially true, that they would look at the experience in their industry. And applying that to drugs, they would find that if they were going to go into the industry at the level of success of the leading concerns, that the profit experience is very good. On the other hand, they would look at the profit experience of the generic producer, say, who does not sell the differentiated prod- uct, does not have a monopolistic product due to a patient, say, and the profit experience would demonstrate a very poor prospect. So, one must understand what it is he is looking at. And they have in their measure tried to capture the risk factor by looking at. the (hf- ference in the profit rates of t.he leading 29 companies. If they had taken the top eight companies, for example, the profit variance would be considerably less. In some other industries they have as few as five companies. So, it turns out, that their measure is very arbitrary. It is true that Conrad and Plotkin have found a statistically sig- nificant relationship between their measure of risk and industry profits. It is statistically significant in the sense that the correlations that they came up with could not have been due simply to chance. And in this sense it is significant statistically. This does not imply that there is necessarily a causal relationship between the two variables they were measuring. In truth, they have misinterpreted the causal factors responsible for their statistical relationship. Upon close analysis, the Conrad-Plotkin measure of risk turns out to be a better proxy of relative market power than of risk. Their meas- ure assumes the existence of "homogeneous" industries; that is, "in- dustries in which all the firms product similar products, compete in the same markets, and, in general, face the same elements of risk and uncertainty." In fact, however, when broad industry definitions are used, such as those in the Conrad and Plotkin study, the constituent firms within each "industry" are frequently highly differentiated from one another by a variety of factors. Hence, each firm in the industry may face different risks and other factors having a bearing on profits. This is particularly true in consumer, service, and other so-called differentiated product industries. Because of advertising and other factors, some firms in such industries have a pronounced and persistent advantage over others. As a result, the most advantaged firms earn persistently higher profits than the less advantaged firms. Such a dif- ference between the profits of the most advantaged and least ad- vantaged firms in an industry may provide a rough measure of the height of the entry barriers into the industry. Economic theory pre- dicts and empirical analysis verifies that the higher an industry's entry barriers, the higher its profits. Hence, if intra-industry profit variance measures the height of entry barriers, we may expect a positive statis- tical association between industry variance and `average industry profit rates. Thus, it is not surprising that Conrad and Plotkin find some statistical association between intra-industry profit variance and average industry profit rates. Unfor1~unately, they misinterpreted the significance of their own findings. PAGENO="0272" 1822 COMPETITIVE PROBLEMS IN THE DRUG INDUSTRY The. effect of product-differentiation-created int.ra-mdustry profit variance on the Conrad-Plotkin analysis is obvious by inspection of specific industries with a high degree of product differentiation. I should apologize, Senator, for this rather theoretical and, I am sure, obscure explanation of risk. Senator SCOTT. That is all right. We do not understand it either. Dr. MUELLER. But I think the following illustrations explain what I am getting at here. Automobiles, for example, show up as a very "risky" industry in the Conrad-Plotkin study. This is the case because of the wide disparity in profits between the strongest, most entrenched firms and the weak- est, marginal ones. For example, during the last 5 years automobile companies enjoyed average profits as follows: General Motors, 21 per- cent; Chrysler, 15 percent; Ford, 14 percent; American Motors, 6 per- cent. During most of the 1050's while Studebaker was in the industry it operated in the red. I think nearly all students of industria.l orga- nization will agree with Bain that the reason for the high average profit rates in the automobile industry is the high degree of market concentration and the very great barriers confronting potential en- t.rants. Thus, the persistently high average profit rates of the auto- mobile industry are primarily due to the structure of the industry, not its risk, as measured simply by the difference in the profit rates of the low and high companies. Nor are automobiles the exception. On the contrary, of the indus- tries included by Conrad and Plotkin, eight of the nine with average profit rates exceeding 14 percent were industries characterized by sub~ stan~t.ia.l differentiation advantage among even the largest firms, and in each case the most advantaged firms held a substantial and persistent profit advantage over the less advantaged firms. The drug industry is an especially poor candidate for the explicit assumption of the Conrad-Plotkin model that induStries must be homo- geneous. There are great product differences among even the 2.9 drug companies they studied. They produce varying mixes of ethical and proprietary drugs, varying proportions of branded and generic drugs, and then enjoy varying degrees of differentiation for their branded drugs. All of these factors, as well as a number of others, result in persistently higher profits for some drug companies than others. Amer- ican Home Products, for example, not only earned average profits well above all other drug companies, but over the entire period 1954-1966, it had profits higher than every other firm. On the other hand, over the same period, R.exa.li Drug had the lowest profits among the top eight companies in all but 2 years, when it was second lowest. Although a number of factors affect the profit differential among drug companies, the degree of advertising-achieved-product differen- tiat~on plays a big role. Table 4 classifies the 29 drug companies used in the Conrad and Plotkin study by the volume of their advertising outlays. The five companies with advertising outlays in excess of $50 million in 1966 enjoyed an average rate of return of 29.2 percent during 1961-1965; those with advertising outlays between $10 million and $50 million had an average rate of return of 19.7 percent; and those spending less than $10 million earned 17.3 percent. Significantly, all of the top five advertisers earned in excess of the average return of those spending between $10 million and $50 million. PAGENO="0273" COMPETITIVE PROBLEMS IN THE DRUG INDUSTRY 1823 The preceding reveals that much of the profit variance which Con- rad and Plotkin found among leading drug companies is the result of the product differentiation advantage held by some firms in the industry. If we are correct in believing that differences in intraindustry profit variance actually measures differences in the degree of product dif- ferentiation rather than risk, then Conrad and Plotkins' correlation results may have been heavily influenced by the inclusion of highly differentiated industries. We now turn to an examination of this possibility. To test the hypothesis that product differentiation caused profit variance is largely responsible for the statistical association uncovered by Conrad and Plotkin, we have analyzed separately their consumer goods and producer goods manufacturing industries. Product dif- ferentiation, of course, shows up primarily in consumer goods. Figure 4 shows all the industries used by Conrad and PIotkin. They lind a modest degree of correlation between intraindustry profit vaii- ance and profits using one measure of profits rates. Using a number of other measures they found less close relationships. Figure 5A shows that the plotted observations of the consumer goods industries used in the Conrad-Plotkin analysis, and figure 5B shows the plotted observations of the producer goods industries. Among consumer goods industries, you will dbserve, there is a quite strong positive relationship, whereas among producer goods industries the relationship is very weak, and is not statistically significant. In consumer goods, 76 percent of the variation among average indus- try profit rates is associated with the variance of intraindustry profit rates. Additionally, the slope of : the regression line fitted to these observations is quite steep, which means industry profit rates rise sharply with high intraindustry profit variance. On the other hand, when only producer goods industries are used in the analysis, the statistical relationship is extremely weak. Only 8 percent of the variation in industry profit rates is associated with variation in intraindustry profit variance. Moreover, the regression line is much less steeply inclined, indicating that average industry 1)rofit rates increase very slightly with increases in intraindustry profit variance. These findings are extremely significant. They demonstrate that the statistical relationship found~ *by~Conrad and Plotkin was due almost entirely to the consumer goods industries in their sample. The fact that no significant statistical relationship remains when only producer goods industries are used to test their model is especially damaging to the Conrad-Plotkin analysis. A basic assumption of their method of measuring intraindustry risk is that the industries analyzed be homo- geneous. Producer goods manufacturing industries are, of course, much more homogeneous than are consumer goods industries. Hence, accord- ing to their assumptions the "purest" relationship between "risk" and profits should have been uncovered in the analysis of producer goods industries. And, of course, there was none. The close statistical relationship existing in consumer goods indus- tries very probably results because intraindustry profit variance in consumer industries is a rougth proxy for the height of entry barriers. Thus, Conrad and Plotkin unwittingly have made a case for the 81-280-68-pt. 5-18 PAGENO="0274" 1824 COMPETITIVE PROBLEMS IN THE DRUG INDUSTRY inference that a substantial part of the high profits earned by drug companies is really due to advertising- and promotion-created barriers to entry, rather than risk. This, of course, coincides with the conclu- sion of nearly every economist who has carefully studied the drug industry. Upon completing our analysis of the Conrad-Plotkin-Markham-. Cootner explanation of risk and profits in the drug industry, I recalled the admonition once given by the great classical economist `and `logician John Stuart Mill. Mill cautioned economists against the pitfall of the multiplicity of causes. We `must always be skeptical of `simple `statistical associations among complicated economic phenomena. Professor Ken- neth Bould'ing put it well when he said: Some of us, perhaps, still have to learn that arithmetic is a complement to, not a substitute for, thought, and that what my spy in IBM calls the "gigo prin- ciple"-that is, garbage in, garbage out-is a sound approach even to the most elegantly computerized simulation. This more or less capsules my findings in reviewing the analysis of drug profits and their possible association with risk. I find, to be very brief, that the high `profit experience of the drug industry is related `only minimally to risk and uncertainty in a casual way. On the other hand, the high profits of the drug industry are more closely associated with high barriers to entry of new competition. In other words, in the classic tradition, the market power enj'oyed by drug firms has `been achieved primarily because the leading drug companies have been able to fence themselves off from effective competition, and in this `sheltered position t'hey have garnered extremely high profits-profits whi~h the economist would label as "abnormal" or "excessive," profits substan- tially above the competitive norm. (The complete prepared statement and supplemental statement sub- mitted'by Dr. Mueiler follows:) STATEMENT OF DR. WILLARD F. MUELLER, DIRECTOR, BUREAU OF Ecoxo~rics, FEDERAL TRADE Co~rMIssIoN Mr. Chairman and members of the Committee. It is a privilege and a pleasure to appear before this committee. I am accompanied today by two members of the staff of the Bureau of Economics, my assistant, Dr. Russell C. Parker, and Mr. William H. Kelly. My appearance today is in response to the request of your chairman `that I sub- mit testimony on the subject of profits in the drug industry, as well as present an independent analysis `of the study Risk and Rctnrn in American Industry- an Econometric Analysis, presented to this committee on December 19, 1967. The study was spoDsored by the Pharmaceutical Manufacturers Association and pre- pared by Gordon R. Conrad and Irving H. Plotkin of Arthur D. Little, Inc., in con- sultation with Professor Jesse W. Markham and Professor P. J. Cootner. Here- after we shall refer to the study as the Conrad-Plotkin study. Before turning `to the `study, we shall first place in perspective the profits of drug Inanufacturers by comparing them with those earned by business enterprises in other American industries. PROFITS IN THE DRUG INDUSTRY Figure 1 shows for 1966 the average rate of return on stockholders investments of leading firms in 22 important American manufacturing industries.' Profit 1 This information is based on the Federal Trade Commission reports on Rates of Return for Identical Companies in ~5elected Manufacturing Industries. The industry aver- age is based on the 12 leading companies in each industry. The 22 industries shown in Figure 1 are those where the 8 largest corporations had combined assets of $1 billion or more in 1966, thereby excluding 15 smaller industries appearing in the FTC Report. None of the excluded industries had profits as high as did the drug industry. PAGENO="0275" COMPETITIVE PROBLEMS IN THE DRUG INDUSTRY 1825 FIGURE 1 RATES OF RETURN OF 12 LEAD~NG HRMS ~N SELECTED !NDUSTR~ES, 1966 DRUGS & MEDICINES PERIODICALS RADIO & TELEVISION EQUIPMENT MOTOR VEH1CLES COMPUTING MACHINES* GLASS CONTAINERS AIRCRAFT NON-FERROUS METALS MOTOR VEHICLE PARTS* CIGARETTES* BAKERY PRODUCTS PETROLEUM REFINING DAIRY PRODUCTS RUBBER PRODUCTS MALT LIQUORS PULP PAPER KNIT APPAREL DISTILLED LIQUORS STEEL PLUMBING FIXTURES CEMENT MEAT PRODUCTS 0Excopt for Motor Vehicler and Computing Machines, based on S firms; and Cigarettes, based on 4 firms SOURCE: Federal Trade Commission - Rates of Return For tdenti:cslConippnies In Selected Manufacturing Industries, 1957-1966. 15 20 NET PROFITS AS PERCENT OF STOCKHOLDERS' INVESTMENT PAGENO="0276" 1826 COMPETITIVE PROBLEMS IN THE DRUG INDUSTRY rates of leading drug manufacturers exceeded those of large firms in the 21 other industries. In fact, drug industry profits were twice as great as one-third of the remaining industries; were 44 percent or more above those of all but 4 other in- dustries; and they exceeded even such traditionally high profit industries as motor vehicles and computing machines. Nor was 1966 an exceptional year. Table 1 compares over the period 1950-1966, the average profits of large drug companies and large companies in the 22 indus- tries shown in Figure 1. Several points are of special interest. First, in the early years, 1050-1955, average drug company profits were about equal to or somewhat above the average of other large manufacturers. Second, beginning in 1956, how- ever, average profit rates of drug companies were well above the average of other large companies. Finally, since 1956 drug companies have consistently ranked either first or second among all large manufacturing industries. This indicates that during the last decade large drug companies have occupied an especially advantaged position relative to large companies in other American industries. Table 2 summarizes profit data of all drug companies and all manufacturing companies for the period 1956-1967. Over the period covered it shows essentially the same picture as Table 1. Since 1956 drug manufacturers have failed to oc- cupy first place in only one year. TABLE 1-RATES OF RETURN OF LEADING DRUG MANUFACTURERS AND ALL LEADING MANUFACTURERS, 1950-66 Ranking of leading Year Drug industry 1 All manufacturing 2 drug companies among all leading manufacturing companies 1950 19.6 17.3 6 1951 15.7 14.6 7 1952 12.7 12.7 11 1953 12.3 13.2 12 1954 12.8 12.8 8 1955 15.4 15.5 8 1956 18.2 13.8 2 1957 21.5 12.8 1 1958 20.2 9.3 1 1959 20.3 10.8 1 1960 18.4 10.3 1 1961 17.6 9.8 1 1962 17.1 10.6 2 1963 17.8 11.5 2 1964 18.9 12.3 2 1965 21.0 13.4 2 1966 21.1 13.3 1 1 Based on 8 largest companies from 1950 to 1953 and 12 largest companies from 1954 to 1966. Based on the 8 largest companies from 1950 to 1953 and the 12 largest companies from 1954 to 1966 in each of the 22 industries shown in fig. 1 aSh the exception of the computer and motor vehicle industries in which the 8 largest firms~ were used for all years and cigarettes in which the 4 largest firms were used for all years. Note: Rate of return after taxes as a percent of stockholders' investment. Sources: Report of the Federal Trade Commission, "Rate of Return for Identical Companies in Selected Manufacturing Industries, 1955-66" and Moody's Industrial Manual, 1952-54. This pattern of persistently high profits indicates that large drug companies occupy a unique position in the American economy. And they appear to have become increasingly unique since the mid-1950's. Before turning to an analysis of the Conrad-Plotkin-Markham-Cootner expla- nation of high profits in the drug industry, I shall review briefly w-hat appears to be a virtual concensus of opinion among researchers in the field of industrial organization concerning the causes of high profits of drug manufacturers. In this connection I w-ould like to emphasize the crucial role which Con- gressiosoal hearings have played in developing the facts necessary for scholars to study the organization and performance of the drug industry. Prior to the Kefauver drug hearings on administered prices in 1959-1961, not a single article concerning the American pharmaceutical industry had appeared in a professional economic journal.2 2 In each case they ranked second to automobile manufacturers. a Hugh Douglas Walker, "Market Power and Price Levels in the Ethical Drug Industry," Unpublished Ph.D. dissertation, Vanderbilt University, June 1967, pp. 2-3. PAGENO="0277" COMPETITIVE PROBLEMS IN THE DRUG INDUSTRY 1827 TABLE 2--RATES OF RETURN OF DRUG MANUFACTURERS AND ALL MANUFACTURING INDUSTRIES, Year Profits after taxes as a percent of stockholders' equity Profit rank of the drug industry among all manufacturing industries I All drug All manufacturers manufacturers 1956 17.6 12.3 2 1 1957 18.6 11.0 1 1958 17.7 8.6 1 1959 17.8 1 1960 1961 1962 16.8 9.2 16.7 8.8 16.8 9.8 1 1 1 1963 1964 1965 1966 16.8 10.3 18.2 11.6 20.3 13.0 20.3 13.5 1 1 2 1 1967(3quarters) 18.6 11.5 1 Rank among the 26 industries for which profits are reported separately in Quarterly Financial Reports. Source: Federal Trade Commission and Securities and Exchange Commission, Quarterly Financial Report. Since the first such article appeared in 1002, there has been a growing volume of research literature on the subject, all of which has drawn heavily on Lhe Kefauver and subsequent Congressional hearings. The facts developed by this committee have made another enormous contribution to the fund of knowledge concerning the drug industry. I am confident that scholars will be sifting and winnowing the facts for years. THE CAUSES OF HIGH PROFITS IN TIlE DRUG INDUSTRY The preponderance of economic evidence argues that the persistently high profits of the drug industry are the result of the absence of effective price competition in the sale of many products. Price competition in drugs hi ineffective for several reasons. Concentration in the production of many drugs is high because of the patent privilege.1 And even where there are relatively many sellers, as well as many potential sellers (for example, in the case of unpatented drugs sold under generic names), effective prjce competition often is muted by vast advertising, promotion, and other selling effort which differentiates in the minds of consumers the products of the largest drug manufacturers selling under their own brand or trade names from those of other manufacturers.° Hence, manufac- turers selling chemically identical drugs under generic names frequently have difficulty in selling them at any price. The resulting wide price spread between advertised and generic drugs often applies to unpatented as well as patented The Kefauver committee staff analyzed concentration for 51 products in the major drug groupings: antibiotics, hormones, diabetic drugs, sulfas, vitamins, and tranquilizers. These products represented at least two-thirds of the total value of ethical drugs in 1958. The 115 leading drug companies controlled the production of these important products as follows: "In 27 of the products, or more than one-half the entire U.S. output is produced by 1 of the 15 companies. . . . In sulfa drugs, one company accounts for 100 percent of the output in eight of the nine products. In tranquilizers the condition of monopoly prevails in six of the seven products. In antibiotics (other than penicillin) the total output is produced by one company In five of the nine products, and in hormones and vitamis, each, in three out of the nine. In 8 addItional products concentration takes the form of "die- opoly"-control by 2, while in 10 others the entire output is produced by 3 companies. Against the typical structure of concentration in manufacturing industries, it is Indeed remarkable that In only 6 of the 51 products are there as many as 4 producers." Report No. 448 of the Committee on the Judiciary, U.S. Senate, made by its Subcom- mittee on Antitrust and Monopoly, 87th Cong., First Sees., Study of .4dniinisterecl Prices in the Drug Industry, June 27, 1961, pp. 68-69. It well recognized that advertising and promotion effort in the drug industry is greater than it Is In nearly all other large American industries. See, for example, William S. Comanor and Thomas A. Wilson, "Advertising, Market Structure, and Performance: An Empirical Analysis," Review of Economics and Statistics, November 1967, Appendix Table 2. Of the 41 industries studied by Comanor and Wilson, all but two had lower adver- tising-to-sales ratios than did drug manufacturers. Comanor and Wilson further point out that advertising outlays represent less than' half of the total selling expenditures of drug companies. PAGENO="0278" 1828 COMPETITIVE PROBLEMS IN THE DRUG INDUSTRY drugs. These factors shelter the leading concerns, or any concern with a highly differentiated drug product, from effective price competition.6 This explanation of high drug profits is not novel. Nearly all researchers who have analyzed the drug industry in detail have come essentially to the same conclusion.7 Similarly, empirical studies which cut across many industries have identified the elements of market structure that are primarily responsible for high noncompetitive profits.8 These elements are high seller concentration, high barriers to entry, and product differentiation. One or more of these factors are present in the sale of all drugs where price competition is ineffective. Per- haps the most pervasive factor blocking effective price competition in drugs is the presence of substantial product differentiation of branded drug items. A recent econometric study demonstrates that advertising- and promotion-created barriers to entry are the single most important explanation for differences in profit rates in American industry.0 The drug industry was among the indus- tries included in that study. Professor Seymour Harris of Harvard University pretty well summarizes the conclusions of academic scholars concerning the organization and per- formance of the drug industry: "Many are concerned that an industry which comes close to being a public utility achieves the highest profits in relation to sales and investment of any industry; is highly concentrated in its control of the market; reveals serious monopolistic trends; increases the cost to consumers by differentiating the product at a dizzy pace, with the differentiated product usually similar to or identical with existing products; and greatly inflates the cost through record expenditures on selling. The competition among companies to overwhelm the doctors by repetitious and often misleading advertising, and a failure to give as much publicity to the bad side effects as to the immediate beneficial effects, are unfortunate. Thus competition1° forces even highly moral firms to become less ethical in their behavior. In the drug industry the relation of labor to total costs is minimal; and like the soap and tobacco industries, using similar sell- ing techniques, their relation of labor to value added is a minimum-selling expenditures and profits are the large items in gross receipts. C In a highly important study, Professor Hugh Douglas Walker of the University of Indiana measures the extent to which drug prices have been raised by the market power created by patents and brand advertising. He estimates that the gross benefits of the re- moval of both brand names and patents would amount to $617 million per year. Since the removal of patent protection might have an adverse effect on research effort, he esti- mates that the net benefits to society after allowing an additional research subsidy of $192 million (the total amount financed by Industry in 1961) would be $425 million. Professor Walker summarized his finding in a paper before the Econometrics Society meetings, December 29, 1967. His complete analysis appears in his unpublished doctoral dissertation, "Market Power and Price Levels in the Ethical Drug Industry," Vanderbilt University, June 1967. Report No. 448 of the Committee on the Judiciary, U.S. Senate, made by its Sub- committee on Antitrust and Monopoly, 87th Cong., First Sess., t~tudy of Administered Prices in the Drug Industry, June 27, 1961. Federal Trade Commission, Economic Report on Antibiotics Manufacture, June 1958. Leonard G. Schifrin, "The Ethical Drug Industry: The Case for Compulsory Patent Licensing," The Antitrust Bulletin Fall 1967, pp. 893- 915. Henry Steele, "Patent Restrictions and Price Competition in tile Ethical Drugs In- dustry," Journal of Industrial Economics, July 1964, pp. 198-223. Henry Steele, "Mo- nopoly and Competition in the Ethical Drugs Market," The Journal of Law d Economics, October 1962, pp. 131-164. Seymour Harris, The Economics of American Medicine, 1964. William S. Comanor, "Research and Competition Product Differentiation in the Pharma- ceutical Industry in the United States," Economica, November 1964; William S. Comanor, "Research and Technical Change in the Pharmaceutical Industry," The Review of Economics and ~Statistics, May 1965. Frank Cacciapaglia. Jr. and Howard B. Rockmam. "The Proposed Drug Industry Antitrust Act-Patents, Pricing, and the Public," The George Washington Law Review, Vol. 30, June 1962, pp. 875-949. For a dissenting view see Jules Backman. "Economics of Propietary Drugs," Annals of the New York Academy of sciences, July 14, 8 See, for example, Joe S. Bain, "Relation of Profit Rate of Industry Concentration: American Manufacturing, 1936-40," Quarterly Journal of Economics, August 1951; Joe S. Bain, Barriers to New Competition, Harvard University Press, 1962; L. W. Weiss, "Average Concentration Ratios and Industrial Performance," Journal of Industrial Economics, July 1963. Norman R. Collins and Lee Preston, "Concentration and Price Margins in Food Manufacturing Industries," The Journal of Industrial Economics, July 1966, p. 226. A report by the staff of the Federal Trade Commission, The structure of Food Manufacturing, Technical Study No. 8, National Commission on Food Marketing, June 1966, pp. 202-210. H. Michael Mann, "Seller Concentration, Barriers to Entry, and Rates of Return in Thirty Industries, 1950-1960," Review of Economics and Statistics, August 1966, pp. 296-307. Tjnpublished study by Norman R. Collins and Lee Preston, "Concentration and Price-Cost Margins in Manufacturing Industries," April 1, 1966. William S. Comanor and Thomas A. Wilson, "Advertising, Market Structure, and Market Performance," Review of Eco- nomics and Statistics, Nov. 1967. 8 William S. Comanor and Thomas A. Wilson, ibid. 10 Professor Harris is referring to nonprice rather than price competition. PAGENO="0279" COMPETITIVE PROBLEMS IN THE DRUG INDUSTRY 1829 "The cost of drugs is too high. I say this, though I am aware that the research contributions of the industry are important and that the lives saved, the suffering averted, and the acceleration of recoveries are worth more than the $4 billion spent on drugs. But the cost could, be substantially less." ~` Does this mean that risk plays no role in high drug profits? Not necessarily. Although these high profits can be explained by the structural characteristics of the industry-high concentration, high entry barriers, and a high degree of product differentiation-it is conceivable that risk also played some part. Con- rad, Plotkin, Markham and Cootner testified that they believed high drug prices and profits were due primarily to uniquely high risks assumed by large drug manufacturers, and that the Conrad-Plotkin study measured the magnitude of this risk. The relevant question, of course, is how much of total profits can be attributed to the risk factor. Let us therefore turn to the empirical evidence on this subject. THE RELATIONSHIP BETWEEN RISK AND HIGH DRUG PROFITS The investment analyst's view of the drug industry One way of gaining insight into the question of "risk" is to look at what investment analysts tell investors about the drug industry. This may seem to be a rather homespun approach to the problem, but after all it is what investors believe about an industry that determines investment decisions. A perusal of studies by investment analysts indicates that they generally ad- vise investors that the drug industry is a rapid growth, high profit industry where established firms hold a strong position relative to small companies and potential entrants. The industry is freqUently described as "depression re- sistant" because, as one analyst put it, "Illness is no respecter of business cycles, and Americans have shown that they will buy the medicine they need regardless of economic conditions." 12 These characteristics are considered to make drug stocks good "defensive" investments. As a result, drug stocks sell at relatively high price-earnings ratios, indicating that investors are confident of a high future payout. A Standard & Poor's analysis of drugs summed up the factors affecting drug stocks as an investment as follows: "Shares of drug equities have historically sold at relatively high price-earn- ings ratios, owing to the industry's recission-resistant characteristics, its above- average earnings growth rate, and its strong underlying position. Moreover, it is difficult to enter the drug field." ~ Investment analysts generally emphasize that the high earnings of drug com- panies make drug stocks a good buy. This is not to imply, of course, that invest- mnent analysts view the industry as completely riskless. The staff of Moody's Jfl~restors Service., after explaining a number of reasons why drug stocks were a good investment, stated: "The drug industry cannot be risk-free. The postwar years have seen periods of slowdown, and individual companies have suffered temporary setbacks. The causes have been many. Competition has led to price-cutting in popular products, such as penicillin, where capacity has been overexpanded. Occasionally, a profit- able new drug is found to have unsuspected and unfavorable side effects. This, however, is less of a problem than product obsolescence or the expiration of patents on major drugs that have been exclusive with one company. In recent years, government regulation has been tightened at the drugmakers' expense. Finally, the ebb and flow of respiratory diseases often causes sharp fluctuations in drug sales." 14 But after enumerating the above points, the Moody's analysts continued, "The impact of such development has caused only temporary deviations in a growth curve that has pointed strongly upward." 15 Thus, while the drug industry faces uncertainties and problems, from an investor's standpoint these "risks" appar- ently are no greater than those found in many other industries. On the contrary, drugs are considered a sound growth investment. 11 Seymour Harris, The Economies of American Medicine, The MacMillan Company, 1964, ~ "No Cycle for Drugs," analysis prepared by the staff of Moody's Investor Service for Dun's Review, October 1967, p. 127. 12 Standard & Poor's Industry Surveys, Drugs, Cosnietmcs-Basmc Analyses, May 4, 1967, p. D 24. 14 "No Cycle for Drugs," analysis prepared by the staff of Moody's Investor Service for Dun's Review, October 1967, p. 127. 15 Ibid. PAGENO="0280" 1830 COMPETITIVE PROBLEMS IN THE DRUG INDUSTRY Investment analysts frequently make mention of the fact that drug profits may be adversely affected by factors that will increase price competition and thereby erode high profits. The most frequent reference of this sort is the obser- vation that anything threatening to increase the use of generic drugs as opposed to brand name drugs threatens high profits. An analysis by Value Line of the possible effects of `medicare" on drug profits is typical of investment analysts' views on the subject. After explaining that "medicare" very probably would in- crease drug sales, T7alue Line concluded that drug profits would not go up by a corresponding amount because, "Hospitals and institutions usually, wherever possible, buy generic name drugs rather than brand names in order to reduce costs. The most profitable busi- ness for the drug manufacturers is that which comes through drugstores, where drugs are prescribed on a brand name basis." 16 [Emphasis added.] This observation, of course, is concerned with how medicare might affect competition because of the increasing use of generic drugs; it is not explaining profits associated with risks. In sum, there is no reason to conclude, on the basis of advice being given investors by investment analysts, that the drug industry is a uniquely risky industry. On the contrary, the generally glowing reports of investment analysts suggest that large drug companies should have little difficulty obtaining adequate capital should they choose to go into the market for it. Actually, however, their profits are so large that drug companies seldom need go to the capital market for equity capital. And there is no reason to expect that drug companies w-ould have difficulty getting adequate capital even if they en- joyed profit rates comparable .to most other American industries. But perhaps this is a too prosaic approach to the problem. Let us, therefore, turn to the Courad-Plotkin-Markham-Cootner "econometric" explanation of high profits in the drug industry. Conrad-Piotlcin study of risk and profit rates My comments today concerning the Conrad-Plotkin analysis will be limited to an evaluation of the testimony presented to this Committee last month. Arthur D. Little, Inc.. has promised to provide us w'ith the underlying data used in their analysis. With your permission, we will provide a brief supplemental memo- randuin to the Committee should we have any additional observations after reviewing these data. You will recall that Conrad and Plotkin attempted to test the hypothesis that the level of an industry's profit rate is positively correlated with the degree of risk faced by firms within the industry. ifl other words~, the more risky an~ indus- try, the higher its average profits. The concept of risk in investment decision making theory refers to situations where it is impossible to predict with certainty the outcome of particular eco- nomic events. The presence of uncertainty is assumed to affect investors' deci- sions. A common assumption is that investors must be paid a "risk premium" if they have an aversion to assuming risks. "Risk aversion" has been an under- lying assumption in a number of recent theoretical works, particularly in the area of portfolio selection and monetary theory.17 However, the assumption of "risk aversion" is not a universal economic law-. One need only view the crowds at the race track paying for the privilege of taking a gamble to infer that some persons regard risk taking as furnishing positive rather than negative satisfaction. These individuals may he viewed as "risk lovers." As a group, these risk lovers lose money at the race track. This is also the case with persons gambling in cam- modity futures markets. In order to explain why individuals will both purchase insurance to guard against large losses and undertake gambles with remote probabilities of achiev- ing high returns, Friedman and Savage have argued that the same individual may be both a risk averter and a risk lover.18 This, in a nut shell, is what risk aversion theory is all about. But note two im- portasit points. First, the size of the risk premium is an empirical question. The theory tells us nothing about the amdunt of the premium, nor even whether it is 1~ "Medicare: Bad for the Drug Makers ?", The Value Line Investment Survey, Edition 4, February 12, 1965 p. 426. 17 See H. M. Markowitz, Portfolio Selection, John Wiley and Sons, New York, 1959, and Jamcs Tobin, "LioniCity Preference as Behavior Towards Risk," Review of Economic Studies, February 1958. 18M. Friedman and L. J. Savage, "Utility Analysis of Choices Involving Risk," Journal of Political Economy, Vol. 56, August 1948, pp. 279-304. PAGENO="0281" COMPETITIVE PROBLEMS IN THE DRUG INDUSTRY 1831 positive or negative.59 Second, central to the hypothesis that it is necessary to offer a. positive premium to investors in order to attract adequate capital into a risky industry is the idea that risk may cause firms to incur losses as well as to enjoy abnormally high profit rewards. Hence, risky industries would be characterized by the presence of both zruns with abnormally high profits and firms with ab- normally low profits. It would be inconsistent with risk theory if nearly all finns in an industry made very high profits and few or none ever suffered losses. The Conrad-Plotkin measure of risk misses this point. Risk is quantified by Conrad and Plotkin by measuring the variance of individual companies' rates of return about the industry average in a given year and computing a simple average of these values for the sixteen-year period 1950 to 1965.20 This measure assumes that the greater the variation in the profit rates of firms about the industry aver- age, the riskier the industry. The chief conceptual shortcoming of this measure is that it does not necessarily tell us anything about the probability of incurring losses. In truth, using this measure an industry may be defined as risky even though all firms in it earn excessively high profits; on the other hand, this measure may define an industry as having very low risk even though all firms are making little or no profit. An example will illustrate this point. By the Conrad-Plotkin measure, the drug industry is a high risk industry and the aluminum ionTustry is a low risk industry. Conrad and Piotkin's estimates of the average rate of return, standard deviation, and variance for the two industries are as follows: [In percentl Average rate of return Standard deviation 1 Risk (profit variance) 2 Drugs Aluminum 17.5 7.8 8.6 1.3 74.2 1.6 1 The standard deviatian is defined as the positive square roit of the variance. In a normal distribution 68 percent of the observations fall within 1 standard deviation, plus or minus, of thu average; 2 standard deviations about the average include 95 percent of thu obsurvatines. 2 Variance measures the dispersion of obsnrvations about an average. It is computed by taking a simple average of the squared deviations of the observatioes from the avurage. The drug companies in their sample experienced an average rate of return of 17.5 percent over the period 1950~1965.m The standard deviation in profits around this average was 8.6 percent. This means that the profit ratio of roughly two-thirds of the compasuies in the industry fell in the range, 8.9 percent to 26.1 percent. On the other hand, the average rate of return of a group of aluminum companies was 7.8 percent, with a standard deviation of 1.3 percent. This means that two-thirds of the time aluminum company profits fell in the range, 6.5 percent to 9.1 percent. Thus, according to Conrad and Plotkin, the drug indus:try is riskier than the aluminum industry because of the greater standard deviation in the profit rates of drug manufacturers. Just what do these facts concerning: the variation in profits tell an investor about the relative profit expectations in these two industries? They say, in effect, that there is a two to one chance that profit rates in the drug industry will fall in a range from 8.9 percent to 26.1 percent, whereas there is a two to one chance in the aluminum industry that profit rates will fall in a range from 6.5 percent to 9.1 percent. Can anyone seriously argue that investors would prefer to place new capital in the aluminum industry rather than in the drug industry? The only risk that the aluminum investor is saved from is the high probability that aluminum companies will earn less than 9 percent- there is only one chance in six of getting more than 9 percent. On the other 10 See, for example, Professor Bain's discussion of the effect of risk on average profit rates of an Industry. He concludes that, "a weighted average profit rate for all firms In the economy or In the industry (all losers as well as all winners being included) should in- clude a true net risk return of roughly zero-and there should be no obvious risk reward explanation of group-average excess profits." Joe S. Bain, Industrial Organization, 1959, p. 375. ~° Gordon It. Conrad and Irving H. Plotkin, Risk and Return in American Industry, p. 12. 21 Conrad and Plotkin's computations of Industry variance and profits in 59 industries laave been reproduced in Appendix Table 1. PAGENO="0282" 1832 COMPETITIVE PROBLEMS IN THE DRUG INDUSTRY hand, the "risk" the investor in the drug industry faces is that the chances are poor that drug companies will earn a rate of return ~g low as the average return in the ~l1iffliflh1ffl lflthisfr~~ Iii fact, there is only one chance in six that ~m caril a rate of return of below 9 percent, whereas there are four chances out of six that they will earn between 9 and 26 percent, and one chance in six that they will earn over 26 percent. Clearly, then, it is nonsense to infer from the Conrad-Plotkin variance measure of risk that the drug industry is riskier than the aluminum industry in terms of attracting new capital. Losses, or even low profits, are practically unheard of among large drug companies. In this respect the drug industry is practically unique among impor- tant American industries. Figure 2 shows for 22 major industries the percent of the time the 8 largest companies fell in various profit rate categories during the period 1954-1966.~ Large drug companies not only earned a higher return than any other of the major manufacturing industries shown, but none of the drug companies ever experienced losses during the period, nor did any companies experience profit rates below 5 percent. Only two other industries enjoyed this distinction, petroleum refining and cigarettes. Companies in most manufacturing industries had profit rates ranging between 5 percent and 15 percent. By considering this range to represent a sort of "norm," we can visualize the extent to which large drug manufacturers departed from it. Over the period 1954-1966, the 8 largest drug manufacturers were in this range about 25 percent of the time and none ever fell below it. On the other hand, 75 percent of the time the leading drug manufacturers earned profits exceeding 15 percent, and fully 17 percent of the time they had profit rates exceeding 25 percent. Xo other industry matched drugs in the frequency with which companies had profit rates exceeding 15 percent. In only three other industries-motor vehicles, computing machines and aircraft-did large firms have profit rates exceeding 15 percent more than 50 percent of the time. On the other hand, all but four of the remaining 18 industries had profit rates exceeding 15 percent loss than one-fourth of the time. Finally, some companies in all but two of the 21 industries outside drugs earned below 5 percent at least part of the time. These contrasting patterns cast serious doubt on the proposition that large drug companies face a serious risk of incurring losses. Figure 3 illustrates the profit experience of leading drug companies in each year during the 1954-1966 period. Most importantly, it shows a pattern of persistently high profits. None of these large companies earned below 5 percent in any year, and in only four years did any company earn between 5 percent and 10 percent; three of these years were in the beginning of the period-1954, 1955, and 1956. Since then, only in 1962 did a company earn below 10 percent. And in the last three years only a small percentage earned below 15 percent.23 22 This Information was obtained from the Federal Trade Commission reports on Rates of Return for Identical Companies in Selected Manufacturing Industries. The same industries were used as for Figure 1; namely, those in which the 8 largest companies had combined assets of $1 billion or more in 1966. The total assets of the 8 largest firms In each industry were grouped in 5 profit categories for each year, 1054 through 1966. The combined assets of all firms within each category were totaled over the 13-year period and divided by the total combined assets in all categories over the period. The result was a weighted per- centage of the number of companies among the 8 largest firms in each industry, earning profits in each of the 5 profit categories during the period 1954-1966. 23The profit experience of the 29 drug companies used In the Conrad-Plotkin analysis does not differ materially from that of the eight leaders. In only two years (one company in each case) did any of the companies eara less than 5 percent. Over the period 1954-1965, the 29 companies fell In the various profit rate categories as follows: Percent of time companies Profit Rate (Percent) : in category 25 and over 22. 5 20 to 25 13. 6 15 to 20 31. 8 10 to 15 19. 3 5 to 10 12. 7 Below 5 0. 04 PAGENO="0283" COMPETITIVE PROBLEMS IN THE DRUG INDTJSTRY 1833 COMPUTING MACHINES MOTOR VEHICLES AIRCRAFT PERIODICALS GLASS CONTAINERS CIGARETTES MOTOR VEHICLE PARTS CEMENT BAKERY PRODUCTS PETROLEUM REFINING NON-FERROUS METALS MALT LIQUORS KNIT APPAREL PLUMBING FIXTURES DISTILLED LIQUORS STEEL DAIRY PRODUCTS RUBBER PRODUCTS PULP PAPER MEAT PRODUCTS SOURCE: Federal Trade Commission - of Return For Identical Connpanias In Selected Manufacturing tndustrios. II.o~ 15.2~ 13.3~' FIGURE 2 DISTRIBUTION OF RATES OF RETU:RN OF 8 LARGEST FIRMS BY PROFIT RATE CATEGORY 22 MAJOR MANUFACTURING INDUSTRIES, 1954 - 1966 Industry :cd;~i:Ae~c:~cun Percent Distribution of Company Annual Profit Ratios DRUGS & MEDICINES- 18.5~ ________________ 15.5~ ~7~4 RADIO & TELEVISION EQUIPMENT 19 1~ ~!~* //~//!!! I 11 riIrzsrIJf,A 13.4%1'~//~~ 13.3%/~,~If#% 10.9%IU~/ 12.3%'W~~4 9.5~ 1~ I~-~ 11.6~ 10.3% - 9.2%~j 7.7% 7.O%i 6.7%1 I' 9.2%J 11.4%U 1O.8%~ 10.2%~ 5.6%~ 80 100 0 20 40 60 KEY:U Above 25% Ota5% ~15%to25% ~Loss 5% to 15% PAGENO="0284" 1834 COMPETITIVE PROBLEMS IN THE DRUG INDUSTRY ~tv~r~ t~c~ ~ StA ///f//~ ~ ~Wa~~!P~ ~aS a -, - ~t1~ r~ 111 II air /,42á4//~/i/// lilt ivtai~~~~ ~JJj/~ ~! VAr FIGURE 3 DISTRIBUTEON OF RATES OF RETURN OF S LARGEST FIRMS IN THE DRUG INDUSTRY, 19541966 Veer Industry Averege Pe:ceet Distribution of Corupony Aenuol Profit Pete Prof it Rote 1954 12.6% 1955 15.4% 1956 18.2% 1957 21.5% 1958 20.2% 1959 20.3% 1960 18.4% 1961 17.6% 1962 17.1% 1963 17.8% 1964 18.9% 1965 21.0% _______ ~pyyyy~J rp,,,jj /f/~~f 1966 21.1% ___________________________________ ________________ ~SI4~ ~ F/YVJ~~i1 ~f7I~/%' 20 40 60 80 KEY: Above 25% fl 10% to 15% SOURCE: ofRethrnFo~d:~~Cpg~ryruo. 20% to 25% 5% to 10% teOotodedMoeotootorieo_tvdostva 15% to 20% ~avi too ~iA/ PAGENO="0285" COMPETITIVE PROBLEMS IN THE DRUG INDUSTRY 1835 This actual profit experience seems to fly in the face of the Conrad-Plotkin- Markbam-Cootner inference that drug manufacturing is a uniquely risky business. The explanation, of course, is to be found in their definition of risk~ Using a different definition of risk, Dr. Irving N. Fisher and Dr. George R. Hall of the Rand Corporation concluded that risk accounts for a very small portion of the high profits of drug companies.24 The findings are shown in Table 3. They show that for the period 1959-1904 drug companies earned an average returis of 18.32 percent. Fisher and Hall attributed 1.08 percent of this to risk. They concluded that the "risk premiums" for drugs are "very low," and that the explanation for high drug profits "must be sought in factors other than risk." It is true that Conrad and Plotkin have found a statistically significant relationship between their measure of risk and industry profits. But they have misinterpreted the casual factors responsible for their statistical relationship. TABLE 3-FISHER & HALL ESTIMATES OF AVERAGE INDUSTRY RISK PREMIUMS [In percent] Average observed Risk-adjusted rate Average risk Industry group rate of return of return premium Drugs 18.32 16.64 01.68 Aerospace 15.70 13.35 02.45 Chemicals 14.09 11.31 02.78 Petroleum 11.47 10.26 01.21 Rubber 10.96 10.21 00.75 Fond 10.72 09.15 01.57 Electrical machinery 11.96 08.57 03.39 Automotive 14.77 07.54 07.23 Office machinery 14.08 07.24 06.84 Steel 08.25 07.03 01.22 Textiles 07.89 05.94 01.95 Source: Irving & Fisher and George R. Hall, "Risk and Corporate Rate of Return," paper presented before the Econo- metrics Society, Dec. 29, 1967. Upon close analysis, the Conrad-Plotkin measure of risk turns out to be a better proxy of relative market power than of risk. Their measure assumes the exist- ence of "homogeneous" industries; that is, "industries in which all the firms produce similar products, compete in the same markets and, in general, face the same elements of risk and uncertainty." 20 In fact, however, when broad industry definitions are used, such as those in the Conrad and Plotkin study, the constitu- ent firms within each "industry" are frequently highly differentiated from one another by a variety of factors.27 Hence, each firm in the industry may face dif- ferent risks and other factors having a bearing on profits. This is particularly true in consumer, service, and other so-called differentiated product industries. Because of advertising and other faètors, some firms in such industries have a pronounced and persistent advantage over others. As a result, the most ad- vantaged firms earn persistently higher profits than the less advantaged firms. Such a difference between the profits of the most advantaged and least advan- taged firms in an industry may provide a rough measure of the height of the entry barriers into the industry.n Economic theory predicts and empirical anal- ysis verifies that the higher an industry's entry barriers, the higher its profits.25 24 Hall and Fisher, "Risk and Corporate Rate of Return," paper presented at the meet- ings of the Econometrics Society, December 30, 1967. Their complete study appears in Risk and the Aerospace Rate of Return, The Rand Corporation, Santa Monica, California, December 1967. Hall and Fisher measure risk as the variance of the profit rates of com- panies overtime taking into account trends in profit rates. 22 Hall and Fisher, op. cit., p. 16. ~° Fisher and Hall, Risk and the Aerospace Rate of Return, op. cit., p. 31. FIsher and Hall conclude that not only does Conrad and Plotkin's measure of risk Involve serious practical measurement problems, but that it also "does not fully agree with a reasonable theoretical notion of risk." Ibid. 27 Ibid. 28 Joe S. Bain, Barriers to New Competition, 1962. The difference In the profits of the most advantaged and least advantaged firms most accurately measures the height of entry barriers when the least advantaged firms earn only a "normal" profit. 20 Ibid. PAGENO="0286" i~3b COMPETITIVE PROBLEMS IN THE DRUG INDUSTRY Hence, if intra-industry profit variance measures the height of entry barriers, we may expect a positive statistical association between industry variance and average industry profit rates. Thus, it is not surprising that Conrad and Plotkin find some statistical association between intra-industry profit variance and aver- age industry profit rates. Unfortunately, they misinterpreted the significance of their own findings. The effect of product-differentiation-created intra-Industry profit variance on the Conrad-Plotkin analysis is obvious by inspection of specific industries with a high degree of product differentiation. Automq1jjIe~, f9P 9~Ample, show up as a very "risky" industry in the COfl~a~-Plotkin stuCy. This is the case because of the WhO dieparity in proflts between the strongest, most entrenched firms and the weakest, marginal ones. For example, during the last five years automobile companies enjoyed average profits as follows: General Motors, 21 percent; Chrys- ler, 15 percent; Ford, 14 percent; American Motors, 6 percent. During most of the 1950's while Studebaker was in the industry it operated in the red. I think nearly all students of industrial organization will agree with Bain ~° that the reason for the high avcraçje profit rates in the automobile industry is the high degree of market concentration and the very great barriers confronting potential entrants. Thus, the persistently high average profit rates of the automobile indus- try are primarily due to the structure of the industry not its risk. Nor are automobiles the exception. On the contrary, of the industries included by Conrad and Plotkin, eight of the nine w-ith average profit rates exceeding 14 percent w-ere industries characterized by ~ubstantia1 differentiation advantage among even the largest firms, and in each case the most advantaged firms held a substantial and persistent profit advantage over the less advantaged firms.3° The drug industry is an especially poor candidate for the explicit assumption of the Conrad-Plotkin model that industries must be homogeneous. There are great product differences among even the 29 drug companies they studied. They produce varying mixes of ethical and proprietary drugs,32 varying proportions of branded and generic drugs, and they enjoy varying degrees of differentiation for their branded drugs. All of these factors, as well as a number of others, result in persistently higher profits for some drug companies than others. American Home Products, for example, not only earned average profits well above all other drug companies, but over the entire period 1954-1966, it had profits higher than every other firm. On the other hand, over the same period, Rexall Drug had the lowest profits among the top eight companies in all but twe years, when it was second low-est. Although a number of factors affect the profit differential among drug com- panies, the degree of advertising-achieved product differentiation plays a big role. Table 4 classifies the 29 drug companies used in the Conrad and Plotkin study by the volume of their advertising outlays. The five companies with adver- tising outlays in excess of $50 naillion in 1966 enjoyed an average rate of return of 29.2 percent during 1961-1965; those with advertising outlays between $10 million and $50 million had an average rate or return of 19.7 percent; and those spending less than $10 million earned 17.3 percent. Sigiilficantly, all of the top 5 advertisers earned in excess of the average return of those spending between $10 million and $50 million.83 The preceding reveals that much of the profit variance which Conrad and Plotkin found among leading drug companies is the result of the product differ- entiation advantage held by some firms in the industry. If we are correct in believing that differences in intra-industry profit variance actually measures differences in the degree of product differentiation rather than risk, then Conrad and Plotkins' correlation results may have been heavily in- fluenced by the inclusion of highly differentiated industries. We now turn to an examination of this possibility. 3° Ibid., p. 169. 31 The industries are: radio-TV broadcasters. book publishing, drugs, cosmetics, automo- biles. radio-TV manufacturers. confectionary and soft drinks. 32 It is generally recognized that profits are more stable for highly differentiated proprietary drugs than for ethical drugs. See, for example, James Bolog, "Forecasting Drug Earnings." Financial Analysts Journal, July-August 1966, p. 39. 3° Some companies below the top five earn persistently high profits because they enjoy a strong position in one or two products. Smith, Kline and French has consistently earned high profits, in recent years averaging over 30 percent. Its two specialty items are Thorazine and Compazine, made and sold under license from Rhone-Paulenc of France.~ Sehifrin. op. cit., p. 911. PAGENO="0287" COMPETITIVE PROBLEMS IN THE DRUG INDUSTRY 1837 TABLE 4-ADVERTISING OUTLAY AND AVERAGE EARNINGS OF 29 DRUG MANUFACTURERS Advertising expenditures in 1966 Number of companies Advertising as percent of sales Average earnings on stockholder investment 1 (percent) Standard deviation (percent) Over $50,000,000' $10,000,000 to $50,000,000 2 Under $10,000,000 Total 5 9 15 29 24.1 11.9 (3) 29.2 19.7 17. 3 20.1 5.8 6.6 8. 3 8.6 1 Simple average of after-tan earnings for the perind 1961-65. 2 Drug companies included in the Conrad-Plotkin study with advertising expenditures of $10,000,OJ0 or more in 1966, an reported in Advertising Age, Aug. 28, 1967, p. 36. Not available. Relationship between product differentiatiOn and profits To test the hypothesis that product differentiation caused profit variance is largely responsible for the statistical association uncovered by Conrad and Plotkin, we have analyzed separately their consumer goods acid producer goods manufacturing industries. Product differentiation, of course, shows up prisnarily in consumer goods. Figure 4 shows all the industries used by Conrad and Plotkin. They find a modest degree of correlation between in'tra-industry profit variance and profits using one measure of profit rates.33 Figure 5A shows the plotted observations of the consumer goods industries used in the Conrad-Plotkin analysis, and Figure SB shows the plotted observations of the producer goods industries34 Among consumer goods industries there is a quite strong positive relationship, whereas among producer goods industries the relationship is very weak, and is not statistically significant. ~ Conrad and Plotkin test their theoretical analysis against a number of measures of rates of return and risk. They report the results of fitting eight distinct models in Appendix Tables D-1 and D-2 of their report and refer to others in tine text. Most of the discussion and the conclusions of tine report, however, are based primarily on the two measures which resulted in tine l)est fitting models. The first and most important of these is a model in which rate of return is defined as net profit plus fixed charges over total capitalization. The risk coefficient is the average yearly intra-industry variability of company profit ratios for the period 1950-1965. The mechanics of computation were to calculate a variance of company profit ratios about the industry average profit ratio in each year of the 16-year period and then compute a simple average. Conrad and Plotkln used the coefficient of determination (R2) to evaluate the goodness of fit of their simple correlation-regression models. The R2 for the first model was .46. This means that 46 pecent of the difference in industry profit levels was associated with the intra-industry variance in profits. Corresponding R2 values for two other models using book value rates of return were .31 (net income as a percent of total assets) and .18 (net income as a percent of common equity). In addition to the models using book value measures of rates of return, Conrad and Plotkin also apply their analysis to returns to stockholders calculated as a percent of the market value of outstanding stock. This ratio is defined as the total of dividends, fixed charges, and the change in market value of shares of common stock during tine year divided by the market vnlue of common shares outstanding at the beginning of the year. The risk coefficient computed on the basis of this earnings ratio is suggested as a measure of the market risk to present and prospective purchasers of outstanding shares of stock. There is apparently no claim that it is an indicator of the company's ability to attract new investment financing or that it indicates tine way in which real resoinrces in the economy should be allocated efficiently between industries. We concur with this. The relevant comparison in evaluating an industry's ability to attract new capital should be expected earnings relative to cost of capital. The market value of outstanding stock may bear little relationship to a company's earning on invested cnpital because the market value of outstanding stock has capitalized into it expected future income resulting from monopoly profits. Because of this, a firm enjoying the benefits of monopoly and earning profits well in excess of the cost of capital may exhibit only a normal rate of return per dollar of equity measured at market prices. ~ We have excluded the following industries from our analyses because none, strictly speaking, is a manufactured product: Radio-TV broadcasters, book publishing, publishing, trucking, eating places, department stores, apparel retailers, air transport, shipping, variety retailers, food retailers, and financial institutions. Leading firms in all of these industries, with the possible exception of shipping and trucking, enjoy varying degrees of product differentiation. Therefore these industries do not meet the homogeneity assumption of the Conrad-Plotkin model. They therefore most appropriately should be grouped with the consumer goods industries shown in Figure 5A. When the above 12 industries are included with the 15 consumer goods industries plotted in Figure 5A, the resulting R2 is 0.63. PAGENO="0288" [NDWTRY Figure 4--Ratcof return and intra-industry variance RATE in rate of return, 59 industries, 1950-1965 of 20 RETURN C) C s * Y~8.6÷,QS3X E~p1ainec1 variance F .46 2 o t .._._.l ` -L Z ID 20 30 CC' 50 60 70 80 90 100 110 120 130 ~ Industry rate of retui-n = (net income + fixed charges)/total capitalization Iiitra-industry variance in rate of CI) return Intra-industry variance is the 16-year simple average of the variances of company profit rates about their industry views in each of the years. PG represents position of pharmaceutical industry. Source: Gordon R. Conrad and Irving H. Plotkin, Risk and Return in American Industry Figure 1, p. 1. The plotted observationsThppear in Appendix Table 1 of t~it'statement. PAGENO="0289" COMPETITIVE PROBLEMS IN THE DRUG INDUSTRY FIGURE 5 Relationship between industry average rate of return and intra-industry variance in rate of return A. 15 Consumer Goods industries 10 20 2(~ Source: Appendix Table 2. 0(5 00 ItS `Ott SU ontra-industry variance in rate of retttrn 1839 industry mate of return I Industry rate of return I B. ~2 Prod~~er Goods Tnddutrie~ 15 10 -~J------~-- - * Regression line ~ I 0~ Y~9.6t.03K 0.~ , f Iigplained variance R~,08 81-280-68-pt. 5-19 PAGENO="0290" 1840 COMPETITIVE PROBLEMS IN THE DRUG INDUSTRY In consumer goods, 76 percent of the variation among average industry profit rates is associated with the variance of intra-industry profit rates. Additionally, the slope of the regression line fitted to the observations is quite steep, which means industry profit rates rise ~harp1y with high intra-industry profit variance. On tile ~th@r liflncl, when on'y producer goods industries are used in the analy- use statistical relationship is extremely weak. Only 8 percent of the variation in industry profit rates is associated with variation in intra-industry profit vari- ance. Moreover, the regression line is much less steeply inclined, indicating that average industry profit rates increase very slightly with increases in intra-indus- try profit variance. These findings are extremely significant. They demonstrate that the statistical relationship found by Conrad and Plotkin was due almost entirely to the consumer goods industries in their sample. The fact that no signifi- cant statistical relationship remains when only producer goods inftustries are used to test their model is especially damaging to the Conrad-Plotkin analysis. A basic assumption of their method of measuring intra-industry risk is that the industries analyzed be homogeneous. Producer goods manufacturing industries are, of course, much more homogeneous than are consumer goods industries. Hence. according to their assumptions the "purest" relationship between "risk' and profits should have been uncovered in the analysis of producer goods industries. The close statistical relationship existing in consumer goods industries very probably results because intra-industry profit variance in consumer industries is a rough proxy for the height of entry barriers. Thus, Conrad and Plotkin unwit- tingly have made a case for the inference that a substantial part of the high profits earned by drug companies is really due to advertising- and promotion- created barriers to entry, rather than risk. This, of course, coincides with the conclusion of nearly every economist w'ho has carefully studied the drug industry. CONCLUSION Upon completing our analysis of the Conrad-Plotkin-Markham-000tfler expla- nation of risk and profits in the drug industry, I recalled the admonition once given by the great classical economist and logician, John Stuart Mill. Mill cau- tioned economists against the pitfall of multiplicity of causes. We must always be skeptical of simple statistical associations among complicated economic phe- nomena. Professor Kenneth Boulding put is well when he said, "Some of us, perhaps, still have to learn that arithmetic is a complement to, not a substitute for, thought, and that what my spy in IBM calls the `gigo irinciple,' (that is, garbage in, garbage out) is a sound approach even to the most elegantly com- puterized simulation." This more or less capsules my findings in reviewing the analysis of drug profits and their possible association with risk. I find, to be very brief, that the high profit experience of the drug industry is related only minimally to risk and uncer- tainty in a causal way. On the other hand, the high profits of the drug industry are more closely associated with high barriers to entry of new competition. in other words, in the classic tradition, the market power enjoyed by drug firms has been achieved primarily because the leading drug companies have been able to fence themselves off from effective competition, and in this sheltered position they have garnered extremely high profits-profits w-hich the economist w-ould label as "abnormal" or "excessive," profits substantially above the competitive norm. ~ Kenneth Boulding, "The Economics of Knowledge and the Knowledge of Economics" American Economic Review, May 1906, p. 10. PAGENO="0291" Variance (risk) 116. 859 99. 085 78. 783 74. 2 13 69. 936 67.284 59. 901 58. 127 57. 631 41. 800 34. 965 34. 338 33. 259 32. 866 29. 705 28. 321 28. 217 27. 486 27. 426 26. 777 25. 412 22. 822 21. 306 20. 535 19. 873 19. 580 19. 528 18. 637 18. 599 18. 312 16. 111 15. 485 15. 476 14. 330 14. 170 13. 321 13. 320 12. 497 11. 812 9.259 9.219 8.948 8.769 8.708 8.682 8.477 8.157 8.061 7.804 7.494 7.255 6.018 5.331 5.080 5.014 4.521 3.709 2. 899 1.579 Mean (return) 18. 929 15. 477 8.797 17. 524 10. 552 18. 726 12. 231 17. 983 15. 042 14. 699 8.754 15. 214 15. 275 8.532 7. 114 12. 949 11. 950 10. 959 10. 121 11. 348 9.831 12. 949 12. 144 8.698 11. 107 9.019 9. 493 11. 852 10. 002 8. 547 10. 396 8.007 9. 199 7.878 12. 453 9.242 7.209 9.849 6.224 10. 764 11. 137 6.625 12. 177 11.666 8. 591 7.519 9. 443 10. 082 10. 819 10. 634 9.689 9.531 8. 304 9.673 8. 534 7.410 8.091 9. 546 7.778 COMPETITIVE PROBLEMS IN THE DRIJG INDUSTRY 1841 APPENDIX TABLE 1.-INDUSTRY VARIANCE AND RETURN BASED ON BOOK VALUE Industry 1. Radio-TV broadcasters 2. Publishing books 3. Gold mining 4. Drugs 5.Publishing 6.Cosmetics 7. Aerospace 8.Automobilu 9. Radij-TV manufacturers 10. Confectionery 11. Building materials, heat 12. Beverages, soft drinks 13. Lead and zinc 14. Miscellaneous metalwork 15. Watches 16. Auto parts accessories 17. Oil, crude, producers 18. Electrical products 19. Machinery 20. Metals, miscellaneous 21. Beieruges, brewers 22. Electronic products 23.Chomicals 24.Shoss 25. Trucking 26. Machinery, metal fabrication 27.Copper 28. Eating places 29. Retail, department stores 30. Retail, apparel chains 31. Container, paper 32. Home furnishing 33. Food products 34. Air transport 35. Office and business equipment 36. Auto trucks 37. Beverages, distillers 38. Textile apparel manufacturers 39.Shipping 40.Tobacco 41. Building materials, rouf 42. Retail, variety stores 43. Vending machines 44. Building materials, cement 45. Blast furnaces 46. Textile products 47. Paper 48. Abrasive products 49. Retail, food chains 50.Oil 51. Forest products 52. Paint 53. Cnal, bituminous 54. Tire and rubber 55. Steel 56. Railroad equipment 57. Containers, metal and glass 58. Financial 59. Aluminum Source: Gordon R. Conrad and Irving H. Plotkin, Risk and Return in American Industry, table F-i, p. 78. PAGENO="0292" 1842 COMPETITIVE PROBLEMS IN THE DRUG INDUSTRY APPENDIX TABLE 2-A--CONSUMER GOODS INDUSTRIES RATES OF RETURN AND INTRA-INDUSTRY VARIANCE IN PAT~Or RETUIST4, 1550-65' Industry Ave rage rate of return I 1950-65' ntra-ifldustry variance in rate of return 2 Drugs Cosmetics Automobile Radio-TV manufacturers 17.5 18.7 18. 0 15.0 74.2 67.3 58. 1 57.6 Confectionery Beverages, soft drinks Watches Beverages, brewers Shoes Home furnishings Food products Beverage, distillers Textile apparel manufacturers Tobacco Tire and rubber 14. 7 15. 2 7. 1 9. 8 8.7 8.0 9.2 7.2 9. 8 10.8 9. 7 41. 8 34. 3 29. 7 25. 4 20.5 15.5 15.5 13.3 12. 5 9.3 5. 1 I Rate of return is defined as net income plus fixed charges divided by total capitalization. 2 The variance is the simple average of intra-industry variances for each year of the period, 1950-65. Source: Gordon R. Conrad and Irving H. Plotkin, "Risk and Return in American Industry," Arthur D. Little, Inc.; appen- dix table F-i, p. 78. APPENDIX TABLE 2-B-PRODUCER GOODS AND MINING INDUSTRIES RATES OF RETURN AND INTRA- INDUSTRY VARIANCE IN RATE OF RETURN, 1950-65 Industry Average rate of return, 1950-65' Intraindustry variance in rate of return 2 Gold mining Aerospace Building materials, heat Lead and zinc Miscellaneous metalwork 8. 8 12.2 8. 8 15.3 8. 5 78. 8 59.9 35. 0 33.3 32. 9 Auto parts, accessories Oil, crude, producers Electrical products Machinery Metals, miscellaneous Electronic products Chemicals Machinery, metal fabrication Copper Container, papnr Office and business equipment Autotrucks Building materials, roof Vending machines Building materials, cement Blast furnaces Textile products Paper Abrasive products Oil 12.9 12. 0 11.0 10.1 11.3 12. 9 12.1 9. 0 9.5 10.4 12. 5 9.2 11.1 12.2 11.7 8.6 7.5 9.4 10. 1 10.6 28. 3 28. 2 27.5 27.4 26.8 22. 8 21.3 19. 6 19.5 16. 1 14. 2 13.3 9.2 8.8 8.7 8.7 8.5 8.2 8.1 7.5 Forest products Paint Coal, bituminous Steel Railroad equipment Containers, metal and glass Aluminum 9.5 8.3 8.5 7.4 8. 1 7. 8 6.0 5.3 5.0 4. 5 3. 7 1. 6 I Rate of return is definod as not incornn plus fixed chargns divided by total capitalization. 2The variance is the simpla avoraga of intra-industry variances for each year of the period, 1950-65. Source: Gordon R. Conrad and Irving H. Platkin, "Rink and Return in American Industry," Arthur D. Little, Inc.; appundix table F-i, p. 78. PAGENO="0293" COMPETITIVE PROBLEMS IN THE DRUG INDUSTRY 1843 SUPPLEMENT TO STATEMENT OF Da. WILLARD F. MUELLER, DIRECTOR, BUREAU OF ECONOMICS, FEDERAL TRADE C01IMIssI0N In my statement on January 18 before the Monopoly Subcommittee I raised a number of questions concerning the theoretical and empirical credibility of a study to Messrs. Conrad and P'lotkin of Arthur D. Little, Inc.1 Most im- portantly, whereks their study concluded that differences in the level of profit rate among industries are explainable by differences in the degree of risk faced by firms within the industries, my analysis demonstrated that their statistical findings wTere the result of the inclusion of consumer goods industries with a high degree of product differentiation. At that time I promised to provide a brief supplemental memorandum analyz- ing the underlying data used by Conrad and Plotkin in their study. The results of this further analysis give added support to the finding that the differences in profit rates among industries result almost exclusively from the inclusion of dif- ferentiated consumer goods industries in the sample. Such industries are char- acterized both by high average profits and by high intra-industry variance in profits. EFFECT OF GROUPING SOME INDUSTRIES Conrad and Plotkin, in choosing the sample of industries to be used in their analysis, grouped together a number of industries which are defined separately in both the Census of Manufacturers and in the data source from which they computed their regression variables.1 One is immediately puzzled by the reason- ing behind such an apparently arbitrary grouping procedure.8 This gPouping procedure is particularly questionable with regard to the "Food Products" and "Machinery" industry groups. In the case of "Food Products" seven 4-digit SIC industries are grouped together; yet several other industries which Census defines as food industries are excluded from the grouping and included in their analysis as separate industries.4 Data for the seven industries were provided separately in the banic data source used by Conrad and Plotkin, and six of the seven had numbers of companies equal to or greater than the numbers of com- panies which were included as separate industry observations. Therefore, to be consistent these six should be included as separate observations in the sample rather than combined into a single observation. The seven industries included in the fOod group are listed in Table 1 along with their individual intra-industry profit variances and average profit rates. As Table 1 shows, there are wide differences in both profit variances and average profit rates among the seven food products industries. Average profit rates ranged between 5.561 and 12.481, and the variances of industry profit rates ranged between 2.190 and 32.698. In the case of "Machinery" Conrad and Plotkin arbitrarily combined eight 4-digit SIC Industries into a single observation. Table 2 shows the variances in profit rates and average profit rates of these eight machinery industries. These industries are clearly distinct and should be treated as separate observations. As is the case with food products, both the profit variance and the average profit rate vary greatly among the various machinery industries. Intra-industry profit variance ranged between 9.372 and 41.005 and average profit rate ranged be- tween 8.212 and 14.552. 1 Gordon R. Conrad and Irving H. Plothln, Risk and Return in American Industry, Arthur D. Little, Inc., May 1967. 2 Ibid., Appendix C, pp. 34-35. Data were not submitted which could be used to determine time effect that the grouping of certain industries had on the regression and correlation models reported by Conrad and Plotkin. There is a suggestion, however, that the grouping may have produced higher correlation coefficients by the fact that observations computed on an unweighted basis had that effect. From the data submitted it was possible to compute simple or unweighted variance coefficients. Using these, the explained variance or the R2 of tile equation with grouped data was slightly higher than an equation with only ungrouped observations. Conrad and Plotkin combined the seven industries shown in Table 1 as one observation, and used confectionery, soft drinks, beer arid distilled liquor as separate observations. For unexplained reasons Conrad and Photkin excluded the sugar industry from their rinalysis although information was available on the industry. In addition to sugar refining, Conrad and Plotkin excluded three other industries for which the Standard & Poor's Compustat Tape of 1967 contained five or more companies and ten Industries which con- tained either three or four companies. These latter ten industries are significant because in some cases Conrad and Photkin combined industries that contained as few as two companies into grouped observations. PAGENO="0294" 1844 COMPETITIVE ?I~OELEMS I~ THE D1UJG INDUSTRY TABLE 1.-INTRAINDUSTRY PROFIT VARIANCES AND AVERAGE PROFIT RATES OF 7 FOOD PRODUCTS INDUSTRIES Industry category I Number of companies 2 Industry profit rate Intraindustry profit variance 4 Food products combined Packaged foods Meatpackers Dairy products Cannedfoods 43 9.075 24.463 9 7 7 7 12.481 5. 561 9. 190 7.215 23.878 12. 590 2. 190 18.004 Bread and pastries Cookies and crackers 6 5 9.458 10.134 32. 698 9.112 Corn refiners 2 9. 964 8. 632 tIndustry category as defined in the basic source used by Arthur D. Little, which was Standard & Poor's compustat industrial tape. 2 Number of companies in the Arthur D. Little sample. 3 Net income plus fixed charges as a percent of total capitalization (total assets minus current liabilities). The ratios shown are the 16-year (1950-65) simple averages of profit ratios for firms included industries. 4 The 16-year simple average of annual unweighted intraindustry profit variances. Weighted intraindustry profit vari- ances could not be computed from the data submitted by Arthur D. Little. Source: Data provided by Arthur D. Little, Inc. TABLE 2.-INTRAINDUSTRY PROFIT VARIANCES AND AV ERAGE PROFIT RATES OF 8 MACHINERY INDUSTRIES Industry category 1 Number of companies 2 Industry average profit rate 3 Intraindustry profit variance Machiaery combined Spacialty machinery Industrial machinery Machine tools 62 16 10 8 11. 165 11.450 12. 181 10. 906 33. 954 41.005 21. 386 21. 054 Construction and materials handling machinery Oil well machinery Agricultural machinery General industrial machinery Steam generating machinery 7 6 6 5 4 10. 580 13.096 8.212 14. 552 9.912 9.372 14.280 33. 349 38. 282 17. 478 1 Industry category as defined in the basic source used by Arthur D. Little, which was Standard & Poor's compustat industrial tape. 3 Number of companies in the Arthur D. Little sample. 3 Net income plus fixed charges as a percent of total capitalization (total assets minus current liabilities). The ratios shown are the 16-year (1950-65) simple averages of profit ratios for firms included in industries. 4 The 16-year simple average of annual unweighte'J intraindustry profit variances. Weighted iatraindustry profit vari- ances could not be computed from the data submitted by Arthur D. Little. Source: Data provided by Arthur D. Little, Inc. The grouping procedure used by Conrad-Plotkin cannot be jiastified because of the fewness of firms in the separate 4-digit SIC industries within these two groups. The criterion which they used for including an industry in their analysis as a separate observation w-as that the industry contain at least five companies. Nine of their sample industries contained the accepted minimum of five coin- panies. As Tables 1 and 2 show, in only one of the seven grouped food products industries w-ere there fewer than five companies, and in only one of the eight grouped machinery industries were there few-er than five companies. These arbitrary groupings conflict with Conrad and Plotkin's testimony before the Committee that they had not "massaged" the data.5 Their arbitrary groupings also violate one of the basic assumptions of their theoretical m'od~l that the in- dustries analyzed be homogeneous, i.e., that the companies within each industry grouping are subject to the same demand and supply conditions. Whatever their rationale for grouping these industries, the i~ocess of ungroup- ing provides an additional test of the hypothesis that the positive relationship be- tween intra-industry variance and industry average profit rate results pri- ° Plotkin testified on this point, "We have not changed one iota of their data, including- their industry classifications as to which companies belong to which industries." Transcript, Tuesday, December 19, 1967, p. 2745. PAGENO="0295" COMPETITIVE PROBLEMS IN THE DRUG INDUSTRY 1845 manly from the inclusion in the sample of consumer goods industries character- ized by product differentiation.6 Turning first to consumer gdods, the separation of the food products group into its separate 4-digit SIO industries enables us to expand the number of con- sumer goods observations from 15 to 20. TJsing data supplied by Arthur D. Little, two regression equations were computed expressing the relationship between intra-industry profit variance (X) and industry average profit rate (Y). The first regression was computed using the sample of 15 consumer goods industries included among the industries analyzed by Conrad and Plotlcin.7 The resulting regression equation is: Equation I Y=9.3+0.038X with explained variance R2=0.26.8 A second regression equation was computed based on a sample of 20 indus- tries. The increased number resulted from replacing the food products group by the six food products industries having at least five sample companies.9 The regression equation using the adjusted sample is: Equation II Y=8.9+0.042X with explained variance R2=0.30. Increasing the size of the consumer goods sample raises from 26 to 30 percent the variation among industry average profit rates associated with the variance of intra-industry profits. In addition, the expansion of sample size increases the slope of the regression line from 0.038 to 0.042, indicating a greater response of industry profit rates to changes in intra-industry l)rOfit variance.10 These re- sults substantiate our earlier finding that in the consiuner goods sector a strong relationship between intra-industry profit variance and industry average profit rates does in fact exist. Turning now to the producer goods and mining sectors, the separation of the machinery group into its 4-digit SIC industries enables us to expand the size of the producer goods and mining sample in the same way as we did for consumer goods. Using the Arthur D. Little data two regression equations were computed-the first using essentially the same producer goods and mining industries as were included in the Conrad-Plotkin sample, and the second using an expanded sample adjusted by ungrouping the machinery industries and replacing the single machinery group observation by the seven machinery industry observations containing at least five sample companies.11 6 The process of expanding the number of observations increases the reliability of a statistical relationship. That is, all other things equal, a larger sample size reduces the probability that the relationship could have been due to chance. If, in addition, the increase in the size of the sample also increases the strength of the relationship between I)rofit variance and average profit rate, one can even be more confident that such a rela- tionship exists. If the relationship is weakened when the sample size is increased, the likelihood that such a relationship actually exists may be more suspect. Observations are listed in Appendix Table 1A. S Intra-industry variance data used In this equation were computed on an unweighted basis. Data were not submitted by Arthur D. Little, Inc., to permit computing the addi- tional values of the variables for Equation II on the weighted basis used by Conrad and Plotkin for Risk and Return in American Industry. The use of weighted variables by Conrad and Plotkin results in a higher R2 for consumer goods than the ones shown in Equation I. (See Figure 5A of my January 18, 1968, statement before this Subcommittee.) There is no reason to believe that the use of weighted observations should reverse any of the relationships between Equation I and Equation II discussed below. Observations are listed in Appendix Table lB. 10 The statistical reliability of these relationships was Increased because of both the higher correlation coefficient and the enlarged size of the sample. By adding tile addi- tional industry observation the statistical probability that the relationship could have occurred by chance was reduced from 2.5 percent to about .5 percent. 11 See Appendix Tables 2A and *2B. Arthur D. Little did not furnish the underlying data on gold mining, lead and zinc, miscellaneous metals, or forest products. Hence, these observations were excluded from the two samples. PAGENO="0296" 1846 COMPETITIVE PROBLEMS IN THE DRUG INDUSTRY The regression equation for the unadjusted sample containing 26 observations is: Equation III Y=9.8+O.018X with explained variance, R2=0.17.12 Computing a similar simple regression equation for the adjusted producer goods and mining sample containing 32 observations, the resulting equation is: Equation IV Y=10.2±0.016X with explained variance, R2=0.10.13 Whereas the expansion of the sample size improved the statistical relationship between intra-industry profit variance aiid industry average I)rOfit rate in the consumer goods sector, exactly the opposite results occurred in the producer goods and mining sectors. The explained variance, R2, decreased from 0.17 to 0.10, indicating that an even smaller percentage of the differences in industry average profit rates resulted from the variance in intra-industry profits.14 More- over, the slope of the regression line, which shows the response of industry profit rates to changes in industry profit variances, decreased from its already low level of 0.018 to 0.016. In sum, the strong statistical relationship between intra-industry profit vari- ances and industry average profit rates in the consumer goods sector proved to be even stronger when the size of the consumer goods sample was increased, while the already weak statistical relationship between variances in industry profit rates and industry profits in the producer goods and mining sample became still w-eaker when the size of the producer goods and mining sample w-as ex- panded. This finding is significant. Together w-ith the findings I reported in may January 18, 1968 statement to this Subcommittee, it demonstrates conclusively that the statistical relationship reported by Conrad and Plotkin was due entirely to the differentiated consumer goods industries in their sample. As pointed out in my original statement, a basic assumption of their hypothesis is that the in- dustries analyzed be homogeneous. The producer goods industries in their sample most closely approximate this assumption. Yet, when only these industries are analyzed, no significant statistical relationship exists between industry variance and profit rates. PROFIT RATES OF LARGEST COMPANIES AND OTHER COMPANIES The underlying data furnished by Arthur D. Little provide additional facts with which to test our hypothesis that the intraindustry variance in profit rates is really a measure of the advantage which some firms enjoy over others by reason of their success in differentiating their product. If we are correct that in many consumer goods industries product differentiation creates advantaged positions for the largest companies relative to smaller ones, then the largest companies may be expected to enjoy persistently higher profits than other com- panies.15 On the other hand, we generally would not expect the largest firms in producer goods industries to enjoy a significant product differentiation advantage over other companies in the industry. So, unless such firms enjoyed some other advantage we would not expect the largest companies in such industries to earn persistently higher profits than the remaining companies in the industry. To test the above hypothesis, the Arthur D. Little sample of companies in each industry was divided into two size classes on the basis of relative company size. 12 Values for X and Y in Equations III and IV are unweighted with respect to company size. Unweighted variables had to be used in measuring the effects of adding the 7 addi- tional ungrouped observations because the company size weights employed by Conrad and Plotkin were not made available by Arthur D. Little. The effect of using unweighted values rather than weighted values can be seen by comparing the R2 and the slope of Equation III with comparable values for the equation shown in Figure 5B of my January 18, lOGS. statement before this Subcommittee. This comparison shows that the relationship between intra-indlustry profit variability and average industry profit rates is even less significant when Conrad and Plotkin's weighted values are used. 11 Ibid 14 The statistical probability that the relationship could have occurred due to chance increased from 2.5 to 5 percent. 15 Ideally, we should know the degree of product differentiation enjoyed by each firm in the industry. Lacking this information, we have assumed that the largest companies possess the most highly differentiated products. This assumption is generally valid becauce the leading companies in consumer products industries are also by and large the leading advertisers. PAGENO="0297" COMPETITIVE PROBLEMS IN THE DRUG INDUSTRY 1847 In each industry, the four largest companies based on total assets were grouped together to form one size class, while the remaining sample companies were grouped in a second size class. If there were fewer than eight sample companies in an industry, the sample was divided evenly between the two classes. The simple average of profit rates for the period 1963 through 1965 was computed for each size class of each industry. Table 3 summarizes the results of these tabulations. TABLE 3.-SUMMARY OF RELATIVE PROFIT RATES OF LEADING COMPANIES AND OTHER COMPANIES, 54 INDUSTRIES Typeof industry' Total md number of ustries2 The number of md ustrius in which the earned 3- leading companies - Higher profits than other companies The same as ether companies Lower profits than other companies Consumer goods Producer goods 19 35 15 16 0 3 4 16 `Industries classified according to Federal Reserve Board listing for the Index of Industrial Production. 2 Arthur D. Little did not submit data for several industries which Conrad and Plotkin used in their analysis reported in "Risk and Return in American Industry." For this reason the total number of industries is slightly fewer. 3 Net income plus fixed charges over total capitalization. This is the profit ratio used by Conrad and Plotkin. Source: Tables 3A and 3B. For the period 1963 to 1965, the leading companies in fifteen of the nineteen consumer goods industries earned higher profit rates than the other companies in the idnustry as a group. In sixteen of the thirty-five producer goods and mining industries the group of leading companies had higher profit rates than the group of smaller companies; in another sixteen industries the group of smaller companies averaged higher profit rates than the group of leading companies; and in the remaining three industries both groups averaged exactly the same profit rates. Clearly, no great advantage accrued to producer goods or mining companies by reason of their being leading companies in their respective industries. The relationship between relative firm size and average profit rates was random. The results of this test again support the proposition that the leading com- panies in consumer goods industries possess some unique advantages which do not acme to the leading companies in producer goods and mining industries. It is our hypothesis that these advantages are the result of the largest companies' greater capacity to differentiate their products. As a result the companies which are most successful in achieving a highly differentiated product are able to charge higher prices and make higher profits than the less advantaged com- panies. In contrast, the products of producer goods and mining industries are more homogeneous. Consequently, relatively small firms'° enjoy profit rates about equal to industry leaders. DIFFERENCES IN PROFIT RATES AMONG CONSUMER GOODS INDUSTRIES Economic theory posits that industries with a high degree of product differen- tiation will experience both higher average industry profits and greater differ- ences in profits between the leading companies and the remaining companies in the industry than industries with a moderate to low degree of product differentia- tion,12 We are now able to test this theory by employing the underlying data used in the Arthur D. Little study. In Table 4 the consumer goods in ustries listed in Appendix Table 3A have been grouped into three categories on the basis of the total advertising expendi- tures of the four leading advertisers in the industry in 1964. It shows that there is a strong positive relationship between the absolute amount of an industry's advertising and its average profit rates. In the top category, containing in- dustries whose products were highly advertised, the average profit rates of both `° The companies included in the Arthur D. Little sample generally included only large and medium size companies. Very small companies were excluded from the sample since it included only companies whose stocks were registered on principal stock exchanges. `~ Joe S. Bain, Barriers to Neov Competition, Harvard University Press, 1956. PAGENO="0298" 1848 COMPETITIVE PROBLEMS IN THE DRUG INDUSTRY leading companies and all other companies were substantially higher than those in either of the other two categories. In fact, the smaller companies in the top category earned higher average profits than the leading companies in the second category containing industries characterized by moderate levels of advertising. As the level of industry advertising declined, the average industry profit rates also declined, dropping from 10.9 percent for the leading companies in the top category to 11.0 percent for the industry leaders in the second category and to 8.5 percent for the leading companies in industries in the lowest category. TABLE 4.-RELATIONSHIP BETWEEN ADVERTISING, PROFIT RATES AND INTRAINDUSTRY PROFIT VARIABILITY FOR 19 CONSUMER GOODS INDUSTRIES Advertising expenditures of 4 leading advertisers, 1964 Number of industries Average of industry profit rates of leading corn- panies, 1963-65 Average of industry profit rates of other companies in 1963-65 Absolute difference in profit rates Average of intraindustry variance in profit rates 1950-65 Over $150,000,000 $50,000,000 to $150,000,000 Under $50,000,000 `3 2 ~1 2 5 16. 9 11. 9 8. 5 12. 9 10. 0 6. 8 4. 0 1. 9 1. 7 89. 8 43. 1 29. 8 1 Include drugs, automobiles, and tobacco. 2 Include cosmetics, beer, soft drinks, confectionery, canned foods, liquor, tires, radio and TV, dairy products, bread and pastries, and biscuits and crackers. 2 Include watches, shoes, home furnishings, apparel, and meat products. Source: Tables lB and 3A. Advertising data were tabulated from Advertising Age, Jan. 3, 1966, p. 46. Greater absolute expenditures on advertising not only lead to higher average industry profit rates but to greater variances in intra-industry profit rates. This occurs, as explained above, because the leading companies enjoy an ad- vantaged position relative to those companies less capable of sustaining large advertising expenditures. Table 4 show-s the difference in average profit rates betw-een the class of leading companies and the class of other companies in each advertising category. In the top advertising category the profit rates of the leading companies averaged 4.0 percent higher than the profit rates of smaller companies. In the second category the leading companies' profit rates exceeded those of the smaller companies by an average of only 1.9 percent. In the third category this difference declined to 1.7 percent. The persistently higher profits of leading concerns resulted in greater variances in profits. As shown in Table 4, the average of intra-industry variance in profit rates was 89.9 for the top category. For the second category the variance dropped to 43.1, and it declined still further to 29.8 for the lowest category. These results again verify our earlier findings that the advertising-achieved product differentiation advantages held by leading firms in an industry affect both the higher dispersion of profit rates w-ithin the industry and industry average profit rates. CONRAD AND PLOTKIN'S UNDISCUSSED FINDING Conrad and Plotkin did not discuss before the Subcommittee their statistically most significant finding.'8 The regression models appearing in Appendix Table D-2, page 40, Risk and Return in. American Industry, show that company profit- ability is negatively related to profit variability. That is to say, the more profit- able a company, the more stable are its profits. This finding is in direct con- tradiction to their principal hypotheses that high profit firms earn such profits as a reward for the "risk" associated with varying profit levels. Conrad and Plotkin gave no explanation for their strange silence regarding this potentially significant finding. CONCLUSIONS Our analysis of the underlying data for the Conrad-Plotkin study supplied by Arthur D. Little provides further support for the hypothesis that the dif- 18 The explained variances (R2) and the simple regression coefficients for temporal vari- ances in Conrad and Plotkin's models analyzing returns on total capitalization and net income to common equity (Model II, page 40, .Rmsk and Return in American Industry) are statistically more significant than those discussed in their text. PAGENO="0299" COMPETITIVE PROBLEMS IN THE DRUG INDUSTRY 1849 ferences between average profit rates among industries studied resulted pri- marily from the impact of product differentiation on industry profit rates. The intra-industry variance in profit rates is really a rough measure of the size of the product differentiation barrier to entry in an industry and not a measure of industry risk. It measures the differenèe in profit rates between the most advantaged firms and the least advantaged firms, this difference increasing as the level of product differentiation increases. Hence, the Conrad-Plotkin study unwittingly supports the inference that in consumer goods industries high industry profits are caused by product differentiation-created barriers to entry. On the other hand, careful statistical analysis reveals that their risk hypothesis has no empirical foundation. TABLE 1A.-RATES OF RETURN AND INTRAINDUSTRY VARIANCE IN RATE OF RETURN FOR 15 CONSUMER GOODS INDUSTRIES, 1950-65 Average Intra- rate of industry Industry return variance 1950-65 1 in rate of return 2 Average Intra- rate of industry Industry return variance 1950-65 1 in rate of return2 Automobiles 11. 129 168. 889 Cosmetics 17. 852 101. 583 Drugs 17. 998 90. 827 Radio-TV manufactures 13. 721 90. 226 Watches 7.335 70. 392 Beverages, brewers 11.966 67. 558 Confectionery 13. 018 50. 358 Beverages, soft drinks 12.424 45. 950 Tire and rubber 9.914 31. 367 Shoes 9.528 26. 143 Beverages, distillers 9.057 25. 591 Food products 9.075 24. 463 Textile apparel manufactures 10. 655 23. 115 Home furnishings 7.874 16. 878 Tobacco 9.627 9. 824 ~ 1 Rate of return is defined as net income plus fixed charges divided by total capitalization. 2 The variance is the simple averagn of unweighted intraindustry variances for each year of the period, 1950-65. Source: "Underlying Data for Study" by Conrad and Plotkin, `Risk and Return in American Industry." TABLE lB-RATES OF RETURN AND INTRAINDUSTRY VARIANCE IN RATE OF RETURN FOR ADJUSTED SAMPLE CONTAINING 20 CONSUMER GOODS INDUSTRIES, 1950-65 Average Intra- rate of industry Industry return variance 1950-65 1 in rate of return 2 Average Intro- rate of industry Industry return variance 1950-65 1 in rate of return 2 Automobile 11. 129 168. 889 Cosmetics 17.852 101.583 Drugs 17.998 90. 827 Radio-TV manufacturers 13. 721 90. 226 Watches 7.335 70. 392 Beverages, brewers 11.966 67. 558 Confectionery 13. 018 50. 358 Beverages, soft drinks 12.424 45.950 Food, bread and pastries 9.458 32. 698 Tire and rubber 9. 914 31. 367 Shoes 9. 528 26. 143 Beverages, distillers 9.057 25.591 Food, packaged foods 12.481 23. 878 Textile apparel manufacturers 10. 655 23. 115 Food,canned foods 7.215 18. 004 Home furnishings 7.874 16. 824 Food, meatpackers 5.561 12. 590 Tobacco 9.627 9.824 Food, cookies and crackers 10. 134 9.112 Food, dairy products 9. 190 2. 190 1 Rate of return is defined as net income plus fixed charges divided by total capitalization. 2 The variance is the simple average of unweighted intraindustry variances for each year of the period, 1950-65. Source: "Underlying Data for Study" by Conrad and Plotkin, "Risk and Return in American Industry." PAGENO="0300" 1850 COMPETITIVE PROBLEMS IN THE DRUG INDUSTRY TABLE 2A.-RATES OF RETURN AND INTRAINDUSTRY VARIANCE IN RATE OF RETURN FOR 26 PRODUCER GOOD AND MINING INDUSTRIES, 1950-65 Average Intra- rate of industry Industry return variance 1950-65 1 in rate of return 2 Average Intra- rate of industry Industry return variance 1950-65' in rate of return 2 Electrical products 11. 292 174. 584 Aerospace 12.628 113.246 Building materials-heating, air conditioning 11. 886 71. 494 Paper 10. 753 64. 410 Oil, crude producers 12. 608 47. 690 Office and business equipment 12. 871 46.056 Miscellaneous metal evork 8.447 35. 774 Machinery, metal fabricating 9.360 34. 644 Auto parts and accessories 12. 749 34. 287 Machinery 11. 165 33. 954 Abrasive prodacts 11.814 32. 161 Containers, paper 10. 638 19. 228 Chemicals 10. 492 18. 759 Textile products 7. 843 18. 754 Copper 9.842 15.673 Auto trucks 10. 301 14. 543 Building materials, cemest 12. 558 13. 214 Steel 8.830 12. 418 Building materials, roof and walL -- 10. 609 10. 839 Point 10. 096 9.855 Vending machines 12. 493 8.406 Oil 10. 087 7. 937 Containers, metal and glass 8.066 7. 871 Coal, bituminous 8.368 7.449 Railread equipment 8.149 6.318 Aluminum 7.938 2.812 `Rate of return is defined as net income plus fixed charges divided by total capitalization. 2The variance is the simple average of unweighted intraindustry variances for each year of the period, 1950-1965. Source: "Underlying Data for Study" by Conrad and Plotkin, "Risk and Return in American Industry." Electrical productu 11. 292 174. 584 Aerospace 12.628 113. 246 Building materials-heating, air conditioning 11.886 71. 494 Paper 10.753 64.410 Oil, crude producers 12. 608 47. 690 Office and business equipment 12. 871 46. 056 Machinery, specialty 11. 450 41. 005 Machinery, general industrial 14. 552 38. 282 Miscellaneous metalwork 8. 447 35. 774 Machinery, metal fabricating 9.360 34. 644 Auto parts and accessories 12. 749 34. 287 Machinery agricultural 8. 212 33. 349 Abrasiveproducts 11.814 32.161 Machinery, industrial 12. 181 21. 386 Machine toals 10. 906 21. 054 Containers, paper 10. 638 19. 228 Average Intra- rate of industry return variance 1950-65' in rate of return 2 Chemicals 10. 492 Textile products 7. 843 Copper 9.842 Auto trucks 10. 301 Machinery oil viell 13. 096 Building materials, cement 12. 558 Steel 8.830 Building materials, roof and wall. 10. 609 Paint 10. 096 Machinery, construction and ma- terialhandling 10.580 Vending machines 12. 493 Oil 10.087 Containers, metal and glass 8. 066 Coal, bituminous 8.368 Railroad equipment 8. 149 Aluminum 7. 938 TABLE 2B.-RATES OF RETURN AND INTRAINDUSTRY VARIANCE IN RATES OF RETURN FOR ADJUSTED SAMPLE OF 32 PRODUCER GOOD AND MINING INDUSTRIES, 1950-1965 Industry Average Intra- rate of industry return variance 1950-65 1 in rate of return 2 Industry 18. 759 18. 754 15. 673 14. 543 14. 280 13. 214 12. 418 10. 839 9.855 9.372 8. 406 7.937 7.871 7. 449 8. 318 2. 812 `Rate at return is defined as net income plus fixed charges divided by total capitalization. 2 The variance is the simple average of unweighted intraindustry variances for each year of the period, 1950-65. Source: "Underlying Data far Study" by Conrad and Plotkin, "Risk and Return in American Industry." PAGENO="0301" COMPETITIVE PROBLEMS IN THE DRUG INDUSTRY 1851 TABLE 3A.-PROFIT RATES OF LEADING COMPANIES AND OTHER COMPANIES IN 19 CONSUMER GOODS INDUS- TRIES, 1963-65 Size classes 1 Leading companies All other companies Industry Number of Average rate Number of Average rate companies or return companies or return 1953-65 2 1963-65 2 Cosmetics 4 22.7 7 14.3 Drugs 4 20.4 25 19.3 Automobiles 2 19. 0 3 10. 0 Confectionery 2 16.2 2 9.8 Radio-TV manufacturers 4 15.5 4 8.8 Soft drinks 3 13.4 3 19.8 Tobacco 4 11.4 6 9.5 Apparel 4 10.8 7 13.2 Dairy products 3 10.5 4 9.0 Canned foods 3 10.3 4 4.0 Malt liquors 4 9.9 13 9.5 Shoes 3 9.2 4 9.1 Biscuits and crackers 2 9. 1 3 10. 1 Home furoisfrings 4 8. 7 5 7. 8 Tires aod rubber 4 8.3 7 7.8 Distillers 4 7.8 7 11.1 Bread and pastries 3 7. 3 3 5. 7 Watches 2 6.9 3 .3 Meat products 3 6. 7 4 3. 4 I The group of leading companies consists of the 4 largest companies except when fewer than 8 companies are included in the sample for an industry. In this case, the sample companies are divided evenly between the 2 groups with the largest firms going into the group of leading companies. If there is an odd number of companies in industries having less than 8 sample companies, the median company is placed in the all other company category. 2 A 3-year simple average of company profit ratios. Note: The packaged food products category was excluded from the group of consumer goods industries for this analysis because it was too broadly defined, and hence contained a number of sample companies which did not compete with one another. Source: `Underlying Data for Study" by Conrad and Plotkin," Risk and Return in American Industry-An Econometric Analysis.' PAGENO="0302" 1852 COMPETITIVE PROBLEMS IN THE DRUG INDUSTRY TABLE 3B.-AVERAGE PROFIT RATES OF LEADING COMPANIES AND ALL OTHER COMPANIES IN 35 PRODUCER GOOD AND MINING INDUSTRIES, 1963-55 Leading companies ` All other companies Industry Number Average of rate of companies return 1963-652 Number of companies Average rate of return 1963-652 Chemicals 4 15.7 39 10.0 Aerospace Machinery, construction, and material handling Office and business equipment Machinery, general industrial Electronic products Electrical products Oil Vending machines 4 3 4 2 4 4 4 3 15.5 13.7 13.0 12.9 12.2 12.0 12.0 12.0 13 4 10 3 26 26 21 3 11.1 9.6 10.9 10.9 10.6 12.0 9.1 11.7 Miscellaneous metalwork 3 11.9 4 7.2 Machinery, oil well 3 11.7 3 11.7 Auto trucks 2 11.5 3 12.5 Oil, crude producers Machine tools 4 4 10.7 10.6 5 4 19.7 11.4 Auto parts and accessories Machinery, steam generating Abrasive products Textile products Building materials, roof Copper Machinery, industrial 4 2 3 4 3 4 9.9 9.6 9.6 9.4 9.4 9.0 9.0 1 1 3 2 4 1 6 3 7 12. 3. 8. 9. 8. 10. 10. 7 9 7 6 8 7 8 Contuiners,puper Cool, bituminous 4 3 8. 8. 7 7 7 4 10.2 7.9 Paint 2 8. 6 3 10.6 Machinery, metal fabricating Railroad equipment Building materials, cement Paper Containers, metal and glass Machinery, agricultural Steel 4 4 6.6 8.4 8.4 7.9 7.9 3 7.8 4 7.6 9 7 6 13 3 3 19 9.2 9.2 7.7 10.7 10.0 14.1 8.0 Machinery, specialty Building materials, heat Blastfurnaces 4 2 7.5 7.0 7.0 12 11 2 12.3 11.3 5.6 Aluminum 2 6.3 3 6.0 I The group of leading companies consists of the 4 largest companies except when fewer than 8 companies sre included in the sample for an industry. In this case, the sample companies are divided evenly between the 2 groups with the largest firms going into the group of leading companies. If there is an odd number of campanins in industries having less than 8 sample companies, the median company is placed in the eli other company category. 2 A 3-year simple average of company profit ratios. Source: "Underlying Data for Study" by Csnrad and Plstkin, "Risk and Return in American Industry-An Econometric Analysis." Senator NELSON. Thank you very much, Dr. Mueller, for your very instructive presentation. Do you have any questions, Senator Scott? Senator SCOTT. If your conclusions are correct, that the high profits vf the drug industry are due primarily to the high entry barriers- your table 1 shows three industries where the industry variance in return based on their book value exceeds drugs-would you therefore conclude that the profits of radio broadcasting, book publishing, and gold mining are also high because of the high entry barrier charac- teristic of those industries? Dr. MUELLER. First, I do not think gold mining was among these, but in these other industries I think the high entry barriers go a long way toward explaining the level of their profits. But I have not studied them individually as I have drugs. Senator SCOTT. There is an old quotation in the publishing business, "Making many books, there is no end." PAGENO="0303" COMPETITIVE PROBLEMS IN THE DRUG INDUSTRY 1853 And there are new publishing companies in the field all the time. I know of one within the last 6 months. Is it so difficult to enter the publishing field? Dr. MUELLER. I think we would have to make a distinction here. These are the profits of leading concerns which Arthur D. Little have included, and it is hard to enter, I would say, magazine publishing and make a profit return comparable to Time and Look and the others in the industry that have a very strong position. If you write a book, you can find someone who will publish it for you. You would not very likely end up with these industry leaders, however. Senator SCOTT. Over the last 20 years, perhaps the easiest industry to get in has been the i'adio broadcasting industry, as an example. Would your parallels app'y to this? Because if they do, it is some sort of a criticism of the Federal Communications Commission, it seems to me. Dr. MUELLER. I \vouldl not want. to be in the position of being up here criticizing another regulatory agency. I would rather be criti- cizing my own bureau, which I know best. And we all can stand some criticism, I am sure. But, again, I think what this measure that I am referring to here, the product differentiation entry barrier, captures the fact that the in- dustry leaders that have a very strong position-say the networks. I do not have the numbers with me., but I looked at them in making this analysis. You have somethiiig comparable to automobiles where you have very strong firms at the top of the heap which earn persistently high profits, and then you have a. range coming down to the weaker ones, and then there are a large number of firms which are making rather modest profits. I simply took the Arthur D. Little data, which showed profits for leading firms essentially and not for all firms in the industry. It is completely consistent with this analysis to find a number of enterprises which have very strong positions because of product differentiation. In drugs a number of enterprises can enter the easy-entry part of the business, say, producing penicillin under a generic name. Here you do not make much money because it is extreme- ly competitive. Senator SCOTT. You have said that the concentration of the drug in- dustry is high because of the patent privilege. Would you suggest to the committee an alternative to the existent pat.ent policy? Dr. MUELLER. I am not up here today making any such suggestion on patent policy. I might say on that point, after listening to the testi- mony of Professor Markham and several of his colleagues last month, in which they spoke of the short-life cycle of drugs, one might draw the inference that 17 years is an extremely long period to enjoy the patent privilege in this industry. Senator Scorr. Because if they were testifying before the Patent Committee their life cycle might gro.w longer. Dr. MUELLER. I expect they might have. In the case of specific drugs, such as tetracycline with which I am somewhat familiar, the patent privilege is still valuable, even though it has only a few years to run. PAGENO="0304" 1854 COMPETITIVE PROBLEMS IN THE DRUG INDUSTRY Senator Scorr. Would you recommend that Congress impose. limita- tions on advertising in the drug industry, or for that matter in any other area? Dr. MUELLER. I would not make this recommendation for any area in American industry. I would hope there are other alternatives to this. Senator SCOTT. I am glad you said that, because if there were to be limitations recommended on the drug industry, it would seem to me they would apply to other industries similarly situated, such as motor vehicles, computer machines, and aircraft. I do not know whether or not-I realize that the purpose of your testimony is a. reply to other testimony, but are you inferring that you, yourself, in your official position, would be happier if the drug companies were making less profits than they are? Is that the point you are making? Dr. MUELLER. I am certainly never in the position of being happy just because someone is making less money. Our system is built on the profit motive, and we use profits as an allocator of resources; but my own response on that point is that I would like to see the drug industry exposed to more of the competitive pressures that are so commonplace in the great bulk of American industry. They are an exception, and I certainly would not want to be interpreted as im- plying somehow that this is characteristic of American industry. It is an exception. And there are very few industries like this. Senator SCOTT. What I am thinking here is this: The Government runs on taxes. Taxes are the result of income and profits. And the President last night asked for more taxes, which would imply that he would like to see more income in the country, and I am wondering whether the desire for more income and more taxes on the one hand is wholly consistent with the criticism of those industries which are producing the income which provides the taxes which pay your salaries and mine and even the President's. I just wonder whether or not the purport of some of the hearings in Congress do not say on the one hand "We want to stop some people from making as much income", and immediately when they do, the people who would be hurt would be the Government which takes a great part of it. That is not very good economics, perhaps, but it is certainly awfully good common- sense. Dr. MUELLER. My only answer would be that we have a tax policy which takes 50 percent of the profits of corporations that are earning, say, 5 percent, or 10 percent. The average of American industry is 12 percent. And to argue that some industries should be immune to any kind of public policy criticism just because, in addition to ex- ploiting of consumers, they are giving half of this back to the Gov- ernment.. So, I think that in a sort of a. commonsense way, this is not a very good rationalization for not. doing anything about the industries which present a public policy problem. Senator SCOTT. Then, in 3-our judgment, is an automatic public Policy problem only part of which industry makes a profit, a. high profit? PAGENO="0305" COMPETITIVE PROBLEMS IN THE DRUG INDUSTRY 1855 Dr. MUELLER. I think we have to make a distinction between in- dustries and individual companies making very high profits. Some industries earning below average industry profits, say in meat pack- ing which happens to be a very competitive, low return industry averaging 6 percent. on equity in many years. In meat packing there are some companies, however, that are extremely efficient and better managed and more aggressive, and they make high profits. Iowa beef packers is the prime example. They are up here with the average of drug companies, 20 percent. On the other hand, other companies make considerably less, the ones for one reason or another just are not able to cut it. But that is the way things are in a competitive industry. And as long as the average return in the industry is adequate to at- tract capital, the system works well. So it is not tha.t some companies are making a profit, no one objects to that., but it is the persistently high profits for the entire industry that are an indicator of poor performance. Senator SCOTT. I am probing to find out wh~ther there is any law or regulation which says that a company can't make a profit, even a high profit. Dr. MUELLER. Certainly not.. Senator SCOTT. Therefore, why is a Government a.gency anxious to show that it is critical of some conwa.nies making profits higher than other companies? Why does that become a matter of Government policy? Dr. MUELLER. The only point that is relevant here is that high profits are indicative of the presence of market power, and upon analysis of the reason for the high profits we find that they are due to a certain clog on competition. Senator SCOTT. Now, if there is any fault in the area of market power, is that not covered by the ahtitrust statutes? Dr. MUELLER. I t.hink not. I think, as I explained in the beginning of my statement, the consensus of economists is that the source of market power is a combination of the patent law and the vast amounts spent on promotion and advertising, which in this industry create the so-called product differentiation : entry barrier. So that insofar as there are solutions, it is something outside of the existing antitrust law. Senator SCOTT. This is the strongest country economically in the world, and the strength of this cOuntry economically is built, in part, upon advertising, upon promotional genius, and upon management.. Why, then, would a Governrnenta.gency come in and be critical of the consequences of these elements which in other areas we rush up to iraise? Dr. MUELLER. First, since this area of a.dvertising is especially a sensitive one for anyone in public office, I want to make clear that I am not here speaking for five Federal Trade Commissioners who have emphasized many times that they are not critical of advertising. I am here as Chief Economist, the Bureau of Economics. And I am not being critical of advertising as such either. Public policy has long been concerned with advertising, as has the advertising profession. The advertising profession now claims the major credit for helping to get the Federal Trade Commission started, and I think they did play some S1-2S0-65---pt. ~i-2O PAGENO="0306" 1856 COMPETITIVE PROBLEMS IN THE DRUG INDUSTRY small part. The point is that we have responsibilities now to keep ad- vertising honest. And we put up with a lot of puffing, and so on, but there are already safeguards in this respect. But I am here simply to try to help identify the problem in this industry, and I am not at all in disagreement with you that advertising can play and has played an extremely important role in many indus- tries in the American economy. But recognizing this, I can't bind myself to the possibility that in certain areas it creates problems. Senator SCOTT. What kind of problems does advertising create? Dr. MUELLER. I think it mutes the voice of competition. The fact that it has prevented the purchase of drugs on the basis of their in- herent qualities. It has prevented the purchase on this basis, and the advertising has prevented, in effect, competition from working as we ordinarily hope it would. It has gone far beyond the informational aspects that we are most interested in. Senator Scorr. Let us see if the corollary of your argument holds water-I do not know whether a corollary is intended to hold water- if you are arguing that advertising in the drug industry is so used as to muffle competition, and to muffle the opportunity for the public to have more information on the nature of the drugs it is `buying, would it therefore follow that if you would advertise less the public would know more? That is a hard one to answer. Senator NELSON. May I interrupt? I have an answer for that. It is not a question of the public know- ing more or less. The public, in fact, knows nothing, and advertise- ments in this industry are not addressed to the public; they are addressed to the man in the medical profession exclusively. I have an example which I think explains the point and may `be helpful to the Senator. We have a case with regard to the drug prednisone where adver- tising very clearly was not helpful to the public at all, nor to the doctor himself, in terms of his patient's welfare. The company that discovered prednisone and got it patented is the Schering Corp., which gave it the trade name "Meticorten." A large number of companies went into the marketplace with the drug under license from Schering. The prices on prednisone, according to the Medical Letter of June 2, 1967, ranged from 59 cents for 100 tablets to the pharmacist by one company to $17.90 a 100 to the pharmacist by another company. The latter is the highest priced precl- nisone on the market. The Medical Letter stated that these drugs were of comparable therapeutic value, and advised the doctors to prescribe the lower priced generic named drug. Now, here is a case where tremendous advertising by one company convinced the doctor-not the public, because the public had no notion about it-convinced the doctor that he ought to continue `to prescribe Meticorten even though there were 21 other drugs in the marketplace that the Medical Letter said were of equal therapeutic value, one sell- ing for 59 cehts a 100, one for 72 or 73 cents a 100, Deltra for $2.20 a 100, Meticorten for $17.90 a 100, and Paracort for $17.88 a 100. So, here is a case where the advertising is to the medical profession and not to the public and has not enlightened anybody, `but rather PAGENO="0307" COMPETITIVE PROBLEMS IN THE DRUG INDUSTRY 1857 resulted in tremendous expenditures by patients in this coimtry for a drug that was availaible on the market at a fraction of the price. And I think the classic result of the publicity which was brought out before this committee occurred ~when one company reduced its price dramatically 2 months after we examined their pricing structure and showed them the Medical Letter. They reduced their price from $17.90 a 100 to $10.50 a 100, and the second company reduced theirs from $17.88 a 100 to about $3.45 a 100. Now, that is a clear case where the advertising was doing a disservice to the medical profession and the public in terms of economics. Senator SCOTT. Isn't Dr. Goddard currently looking into the whole question of whether drugs are ôhemically and therapeutically equivalent? That has not been finally established, has it? Senator NELSON. He is examining that, drug by `drug, as I under- stand it. Senator Scorr. If the investment: opportunities in the brand name drug manufacturing industry are so fabulous, would you tell me why the prices for the stock of these companies are not `constantly sky- rocketing? And is not `that because of the c.onstant threat of congres- sional investigation and threats of further regulatory acti'on? Dr. MUELLER. You mean they keep going up, skyrocketing? Senator Scorr. If their profits are so big, why doesn't the stock keep going up? Dr. MUELLER. Why does it not? Senator SCOTT. Why does it not. skyrocket if profits are as fabulous as you have indicated? Dr. MUELLER. Well, first, the stock earnings ratios on drug stocks are extremely high, they are among the highest of any American industry, indicating that stockholders and investment analysts expect drug stocks to keep going up-of `drug companies to continue to make high profits in the future. Now, as an investor, of course, you may buy into the market at any p'oint, and you are not necessarily going to make vast amounts on your investments in stocks because the market for drug stocks is a very competitive one. How well an individual does in `the market is not any indication of how profitable the industry is. Were you referring to the variation? Senator `Soorr. I am not familiar with these figures, but were you to take, again, the eight leading companies, would their securities, their common stock, on the market today be `higher or lower than it was a year ago or the year before? Dr. MUELLER. I misunderstood you. I thought when you said sky- rocketing, that it was going up. But you are talking about it coming dow-n as well, and whether this is associated with- Senator Score. What I am talking about is the feeling that I have- not the knowledge-that actually drug securities in the market have not been skyrocketing or increasing in their market prices recently. I am not sure, though. Dr. MUELLER. They `are not, of course, entirely unique. The market has been `a very mixed kind' of one `this year. Bu't is response to your- PAGENO="0308" 1858 COMPETITIVE PROBLEMS IN THE DRUG INDUSTRY Senator SCOTT. Machinery stocks, IBM. Xerox, and others like, that have gone up fantastically. Dr. MUELLER. In response to your question as to whether the stock market prices are responsive to public, polic actions in the drug inclus- try, I would say they are. As I indicated in my discussion of the invest- ment analysts' views of the drug industry, one thing `they discuss is that if there is increasing use of generic drugs, profits will come down. In other words, what they are predicting is: If your industry becomes more competitive profits are going to come down, and drug stocks are coming down, and this is true, I suppose, with respect to any public policy action `that. is going to affect drug prices. So, I would not be at all surprised if this association exists. I ran across such comments in preparing this statement, as I looked at the way in which investment analysts viewed the industry. This is some- thing they talked about. Senator SCOTT. I notice there is no chart here which would tend to bear out or theorize-there is no chart of a steady rise in the valua- tion of these drugs. WTe are discussing only profits. Dr. MUELLER. As I pointed out in the beginning, the stocks over this period have-stock price.s relative to earnings have been very high. These ratios are far above the average of American industry. And I did not really think it was important for my comments to show the trends in stock prices. Over this period, however, they reflected the profit situation very accurately. Now, it is possible-and it has happened-that there are variations- Senator SCOTT. Maybe I have not made my point clear in these hearings. Because if this industry or any other is doing anything they ought not to be doing, we want to know about it. If there is anything' that is statutorily wrong, we want to know about it. If corrective leg- islation is needed, we want to know about. it. But I get a little bit fed up, after 25 years of residence down here in this cave of winds, with the constant plea. of Government to indus- try to produce more, make more money, give us more taxes, and then have other agencies of the Government beating them over the head. all `the time they are doing it. Now, it seems to me t.hat if the people are doing something wrong, let us know about it. But I do not think it is wrong to make a profit in this industry, and I have not yet heard a single witness in 25 years that convinced me that it was. Dr. MUELLER. Well, I certainly will say nothing here today to t.ry to convince you that it is improper. I am not carrying a club to beat any American industry. I have a job at the commission, and my part of it is to try to make our system work fair and competitively; that is all. Se.nator SCOTT. I only wish some Government agencies would come before some senatorial committees occasionally in praise of industry, in praise of business, and not always appear in such an antibusiness aspect. Dr. MUELLER. I agree with you 100 percent., Senator. And before the full committee- Senator SCOTT. Because business is jobs, and if you praise business, you are praising job-making. PAGENO="0309" COMPETITIVE PROBLEMS IN THE DRUG INDUSTRY 1859 Dr. MUELLER. And before the full committee last June., with Sen- ator Long and Senator Morse presiding, I appeared along with Mr. Turner of the Justice Department, Professor Adams of Michigan State University; we had a dialog with Professor Gaibraith. At that hearing I went to considerable length to explain as accurately as I could that our competitive system is working extremely well. People have been selling it short. There are exceptions. But Professor Gal- braitli has exaggerated their frequency. In that dialog, he kept emphasizing the automobile industry. He kept telling Mr. Turner that, "Unless you solve automobiles, you have to concede my point that the system has had it." Senator SCOTT. That is the affluent Mr. Gaibraith; isn't it? Dr. MUELLER. The ver affluent. And the point, though, is that the problem industries that are discussed most before congressional com- mittees are the exceptions in our economy. The great bulk of the economy works very well and, unfortunately, we do not come up here and tell you "This industry is working perfectly," because you do not hold hearings on industries that do not have problems. But our silence should not be interpreted as implying that we do not think the system is-the greater part of it-is working very effectively. Senator SCOTT. I am glad you are always proud of it, because I think we ought to be proud of `the system. I have seen some of `the others. I have been in England quite recently, and when I see the mess they have made of things it has made me more inclined to defend our Own. Dr. MUELLER. And the more we see of other country's problems the more tolerant we become of the imperfections in our system. Senator SCOTT. Thank you. Senator NELSON. I am sure the Senator knows that Dr. Mueller came here at the request of the committee. Senator SCOTT. I did not mean to irnj)ly that Dr. Mueller came up here to wage a vendetta against industry. I was speaking broadly and generally about wanting more Government witnesses to find occasion to praise the capitalistic system1 because it needs all the friends it can get, especially around income tax time. Senator NELSON. We asked him specifically to analyze the testimony of other economists who had testified here and give his viewpoint. And I think part of his testimony was directed specifically at the question of factors that make it possible for prices to be held artifically high. Mr. GORDON. Dr. Mueller, at our hearings last month, when Dr. Markham was here, our chairman stated to Dr. Markham on page 2775, line 18 of the transcript: Senator NELSON. As one who majored in economics when I was in school and listened to the arguments in those days, I just learned enough to discover that you could find any poillt of view you wanted from any set of economists any place in the United States. And from my discussions with my friends today, I find that this is still true. Dr. MARKHAM. If I may intrude :on the dialogue, Senator, I think you are quite correct when you went to school, and that was true when I went to school, too. If this analysis means anything, it means I think, what precisely Professor Cootner and the rest who have testified here have said, and that is as Pro- fessor Plotkin has said, this takes the issue really out of the area of opinion; it puts it in the area of quantitative fact. PAGENO="0310" 1860 COMPETITIVE PROBLEMS IN THE DRUG INDUSTRY First of all, how new is this, really? Is this really new? Dr. MUELLER. Tho oconomefric approach which they used? ~\Te11, actually, all econometrics involves the application of mathe- matical and statistical teclmiques in testing economic relationships. Economists have been doing this for many years. The chief advantage of this technique today is that it permits us to analyze vast volumes of data with modern electronic computers in a short period of time, test many relationships. Actually, the Arthur D. Little study is a very simple one. It involves a study of only two relationships, the so-called risk factor and profits. It does not introduce any other variables such as market structure or competitive variables, as have some other studies of risk. These teclmiques have been used long before I was born; they came into prominence in the twenties as simple and multiple regression techniques. So they are not anything new or novel. The whole point of it is that you can't take, as Professor Boulding pointed out-you can't take data that are very crude and send them through a machine and come out with something that is better than the input. In addi- tion, you have to know what it is you are putting in. So, simply using a sophisticated technique does not give you any good new answers to difficult questions unless you have a good con- ceptual framework and you have good data. And, as we point out, when you really understand what Conrad and Plotkin are measuring, their data show nothing at all in the so-called producer goods industries; there is no significant statistical relation- ship present. And, as a result, when we test their theory with, in effect, better data than they used, we find no relationship. That is what the scientific method is all about. As Cootner explained in his opening statement, he had hoped this is the way science operated, the testing of alternative hypotheses, and so on. And this is what we have tried to do in our analysis, really test what it was they thought they had come up with. Mr. GROSSMAN. Dr. Mueller, two questions: First of all, I would like to ask you just as an economist to comment on one of the statements that Dr. Whitney made at our last meeting. He said: "If a $5 prescription, or six of them, will keep a patient from losing a couple of days' pay or spending a night in the hospital, the price is reasonable." \~\That do you think of this "market-will-bear" type of philosophy? Dr. MUELLER. It, in effect, permits a rationalization of anything that you see. It is like saying: "We would all be willing to pay more for our newspapers." Certainly we buy the Star, the Post, and the New ~fork Times, and we pay only 10 or 15 cents for these fine newspapers. We would be willing, however, to pay more; but competition pushes the price down; and, as a result we get more than our moneys' worth. The economist refers to this as a consumer surplus. So, merely because there is a great value received, a greater value received than someone would be willing to pay, does not prove anything in terms of the effec- tiveness of our market system, our competitive system. Senator NELSON. As I understand that testimony given a few weeks ago, what the economist was really saying was that if you are scuba PAGENO="0311" COMPETITIVE PROBLEMS IN THE DRUG INDUSTRY 1861 diving and are under 50 feet of water and run out of oxygen, any price you pay for it is good. Dr. MUELLER. That is right, and you can't argue with that. Mr. GROSSMAN. One final question, and this is more in the way of summation- Dr. MUELLER. If 1 may interject, I think you have devised a new theory of value, Senator. We will call it the "scuba theory of value." Mr. GROSSMAN. In the way of summation, when I asked you about what you thought a fair and profitable return was, you talked in terms of more competition rather than a specific figure. And when Senator Scott asked you about changes in patent policy, you said: "No", per- haps, but you had nothing specific to recommend. With regard to ad- vertising, you had nothing specific- Dr. MUELLER. I never answered the question, I do not think. Mr. GROSSMAN. I wondered whether you could tell us, since we have to rely on your expertise on this, what exactly do you recommend? In other words, you talk about competition. And I would like to know specifically what you think should `be done. Dr. MUELLER. My testimony and my part in these hearings is a very small one. I was asked primarily to analyze the significance of the Conrad-Plotkin study. But this committee and its predecessors which h'ave investigated the drug industry have put their fingers on the key problem, which is the beginning' of understanding. Advertising- crea'ted-product differentiation is certainly a. key point. And some- thing-I am not recommending anything particularly today-but there should be something that can' be done to erode this power. One possibility is the greater use of generic drugs-anything that helps to bring this development about may help; perhaps something designed to simply give consumers, in this case doctors, more information `about what it is they are prescribing for their patients. There are different ways of doing this, some private and some public perhaps. On the patent question, I am sure that you will get other witnesses who would be more willing than I to take a particular position on that point.. But that is the o'ther key point. And many other countries take a.' different position on this than does the United States. Mr. GROSSMAN. Thank you. Senator NELSON. Thank you very much, Dr. Mueller, for your thoughtful presentation to the committee and the very valuable refer- ence work you did. And if you wish to present supplemental material in the future, as you stated earlier that you may, we will be happy to look at it. And, as I understand yOu, you said if you thought it worth- while, you would be happy to come back before the commit'tee and pre- sent it personally? Mr. MUELLER. Yes, sir. Senator NELSON. Thank you very much. We will adjourn until 10 o'clOck tomorrow morning. (Whereupon, at 11 :55 a.m., the. subcommittee adjourned, to recon- vene at 10 a.m., Friday, January 19, 1968.) PAGENO="0312" PAGENO="0313" COMPETITIVE PROBLEMS IN THE DRUG INDUSTRY FRIDAY, JANUARY 19, 1968 U.S. SENATE, MONOPOLY SLTBCOMMITTEE OF THE SELECT COMMITTEE ON SMALL BUSINESS, Washington, D.C. The subcommittee met, pursuant to notice, at 10 :15 a.m., in room 318, Old Senate Office Building, Senator Gaylord P. Nelson (chairman of the subcommittee) presiding. Present: Senator Nelson. Also present: Benjamin Gordon, staff economist; James H. Gross- man, minority counsel; Susan H. Hewman, research assistant; and William B. Cherkasky, legislative director, staff of Senator Nelson. Senator NELSON. We will open the hearings of the Subcommittee on Monopoly with Prof. Leonard Schifrin, head of the department of economics at the College of Williams and Mary. Professor, we appreciate very much your taking time to come here thismorning. I have read your statement. It is a very well documented, very well reasoned statement. We appreciate the time you have taken to assemble this material. You may present your statement in any fashion you see fit. If you find it most effective to read it, you may. If you wish to extemporize or add to it as you go along, that is perfectly satisfactory. I trust you will have no objection if we occasionally interrupt you with some questions. STATEMENT OF LEONARD G. SCHIFRIN, PH. D., DIRECTOR, DEPART- MENT OF ECONOMICS, COLLEGE OF WILLIAM AND MARY, WIL- LIAMSBURG, VA. Dr. SCHIFRIN. I would like to read my statement, Senator Nelson, because it represents a general condensation of a very vast quantity of material. Senator NELSON. Very well. Go ahead, Professor. Dr. SCHIFRIN. My name is Leonard Gerald Schifrin, and I am presently associate professor of economics and head of the department of economics at the College of William and Mary in Williamsburg, Va. I received my bachelor's and master's degrees from the University of Texas, and my doctorate from the University of Michigan. I taught at the University of Michigan and at Yale University before coming to William and Mary in 1965. Within the field of economics my areas of specialization are industrial organization, problems of competition and monopoly, and governmental policy toward business. My particu- lar research interest in recent years has been the economics of the ethical drug industry. 1863 PAGENO="0314" 1864 COMPETITIVE PROBLEMS IN THE DRLTG INDUSTRY I thank the members of this committee for affording me the oppor- tunity of presenting to them some of the salient aspects of my studies of the operation of this industry. First, let me speak to what the economist calls industry perform- ance-a concept of many facets, but one that essentially deals with the effectiveness with which the industry, in its operations, serves the goals of society. Industry performance thus deals with such matters as tech- nological progress, the development of new products, changes in pro- duction levels to correspond to changes in the pattern and size of con- sumer demand, the efficiency with which development, manufacture, and marketing are done, and the reasonableness of prices charged consumers. The different dimensions of performance in the ethical drug industry fall under two headings-"product performance" and "market per- formance." Product performance may be measured by such things as (a) the magnitude and quality of the industry's effort to develop ne~v and better products, (b) the tangible results of this effort, i.e., the num- ber of discoveries and new products flowing from research and develop- ment, a.nd (c) the human and economic impact-better health, longer life, and greater productivity-resulting from these tangible results of industry research and development. In this regard, it is apparent that the drug industry is a highly re- search conscious industry, in basic research as well as in applied re- search and development. It has made available to the public over the years a large number of new and better products; and these products, together with advances in other areas of health, have had dramatic impact on our mortality rates, our longevity, and on our general well- being. Many dangerous illnesses have been brought under control; much of the discomfort and even the hopelessness of illness have been checked. The drug industry, along with other sectors of the health industries, deserve credit for the contributions it has made in this respect. But the record also shows clearly that the product performance of the industry needs qualification and contains serious flaws. A number of authorities have demonstrated that the research performance of the industry is exaggerated by industry officials so as to justify the very large profits of the large drug firms (an illogical argument, by the way). Furthermore, the H. & D. performance is quite small compared to the promotional and advertising outlay, which generally runs four times as large. Beyond questioning its magnitude, critics have con- tended convincingly that much of the research is imitative, wasteful, and aimed at patent procurement rather than progress. Further, many of the so-called "new drugs" coming to market represent duplications of existing drugs, combinations of drugs representing no therapeutic improvement over their components taken separately, or new items that are the result of molecule manipulation rather than substantive therapeutic advancement. We have heard or seen cases of ineffective drugs, harmful drugs, drugs without adequate warnings, dangerous drugs coming to the market as part of the "new drug flow" for which the industry claims credit. Senator NELSON. May I interrupt a moment? Dr. SCHIFRIN. Yes, sir. PAGENO="0315" COMPETITIVE PROBLEMS IN THE DRUG INDUSTRY 1865 Senator NELSON. You say beyond questioning its magnitude, critics have contended convincingly that much of the research performed by the industry is imitative, wasteful, and aimed at patent procurement rather than scientific progress. Do you have any specific examples of people who have made this criticism? Dr. SCIIIFRIN. Yes, sir; I do. Dr. Walter Modell in his own profes- sional articles and in testimony before several Senate committees has made this point rather convincingly. Dr. Claude Forkner has published in several medical journals, including the New England Journal of Medicine, about the drug mixtures in that respect, the "shotgun therapy," as he puts it. Dr. Dale Console, formerly with Squibb Lab- oratories, has talked about such questionable practices of the drug industry. Senator NELSON. You have the articles in which they made their criticisms? Dr. SCHIFRIN. Yes, sir; I have these and other articles and I would be glad to submit them for the record.: Senator NELSON. If you would submit those to the committee coun- we will print them in the record. Dr. SCHIFRIN. Yes, sir. (The material referred to follows:) [From the New England Jounal of Medicine 259 :438-439 (Aug. 28), 1958] MEDICAL INTELLIGENCE-DRUG MIXTIJRES 1 (By Claude E. Forkner, M.D.') Much of the success in modern treatment of disease resides in advances in chemistry, physiology and pharmacology. This has resulted in the isolation of many chemical and biologic substances that have specific and sometimes powerful effects; For example, twenty years ago there was 1 sulfonamide available to the medi- cal profession-namely, sulfanilaiaide. Today, according to the sixth edition of the Modern Drug Encyclopedia, 200 different products are listed as sulfonamide preparations. Many of these contain 1 or more of the sulfonamides in combina- timi with other drugs, often without any indication in the name of what is actu- ally in the drug. For example, few people would guess that Eskadiamer is a com- bination of 2, Neotrizine a combination of 3, Ray-Tri-Mides a combination of 3 and Terfonyl a combination of 3 sulfonamides, Powdalator is a combination of peni- cillin G and sulfanilamide and Thizodrin is a combination of 3 drugs, 1 of which is sulfathiazole. I suspect that few physicians use more than 3 or 4 of the sulfonamides and that they would like them to be marketed under standardized names rather than under a series of arbitrary combinations of letters making meaningless and non- existent words that add nothing but confusion, add greatly to the cost of medical care and promote seric~1~ errors in treatment. Who is it that wants all this nonesense on the market? Is it the doctors? No. Is it the patients? No. Is it the drugstores? No. Is it the hospitals? No. A couple of decades ago only a handful of antispasmodics were available, and only 2 or 3 of these were very useful. Today over 200 differently named products are on the market. A few examples are Alubelap, containing a mixture of alumi- num hydroxide gel, phenobarbital and belladonna, An1esec, containing amino- phylline, epherdrine hydrochloride and amobarbital, Atralose, containing homa- tropine methylbromide, phenobarbital and methylcellulose, Bunesia, containing 1 Presented at the convention of the American Medical Association, New York City, June 4, 1957. 2 Professor of clinical medicine, Cornell University Medical College; attending physician, New York Hospital. PAGENO="0316" 1866 COMPETITIVE PROBLEMS IN THE DRUG INDUSTRY homatropine methyibromide, butabarbital and magnesium hydroxide, and Kol- antyl, containing 5 drugs, none of which have any relation to the name of the product. I have no doubt that physicians would rather use a few drugs of known com- position sold under their own names. such as belladonna, phenobarbital, ami- nophylline, curare and atropine, to 200 different combinations of products sold under names that have little or no relation to the chemical compositions of the drugs. A few years ago one had available 3 or 4 good antihistaminics, which are as useful today as any of the 130 marketed products, which for the most part con- tain a variety of mixtures with which physicians cannot hope to become familiar. There are today over 100 antacid compounds on the market for sale in drug- stores, most of which contain combinations of drugs. One of these drugs picked at random is called X. It contains magnesium carbonate, calcium carbonate, col- loidal kaolin, tricalcium phosphate, sodium bicarbonate, bismuth subcarbonate, papain and diastase. I do not believed that any good.gastroenterologist, any good internist or any good scientist would find use for such a drug. It is like the me- dieval prescriptions written hundreds of years ago. Such concoctions, of which there are hundreds in drugstores, should have no place in modern therapy. Fifteen years ago there were only 2 or 3 antibiotics. Today, a dozen or slightly more have been shown to be of special value, and more than 270 different prep- arations are on the market, many of which are combinations and duplications. A few decades ago, before medicine became as precise and scientific as it is today, doctors' prescriptions often consisted of 0, 8 or more ingredients, many of which w-ere more or less inert. The prescriptions were written in Latin, and the mystery of the ingredients constituted part of their virtue. Gradually, most of this sort of unscientific and meaningless procedure was abandoned; professors of medicine and of pharmacology taught students to use simple drugs, for pre- cise reasons and for definite periods. Gradually. w-ith the intense developments of chemotherapy, of antibiotic therapy, of antihistaminic therapy, of endocrinologic therapy and of vitamin therapy there has been a mushrooming of drug manu- facturers who are in desperate competition. The commercialistic factor has crept into therapeutics to such an extent that physicians everywhere are confused and misled by the literally thousands of drugs increasing in number daily. Dozens of expensive commercial brochures, sample drugs and elegant preparations reach one's office daily and are promptly disposed of in the wastebasket. Many of these modern preparations are mixtures of drugs, some of which are dangerous, some of which are useless, and most of which would be more intelligently given as separate drugs rather than in a shotgun capsule. There are today over 300 preparations on the American market listed as hem- atinics designed to increase the red-cell count and hemoglobin. This, of course, is ridiculous. It is well known to every hematologist that not more than 8 or 10 useful drugs are necessary to treat anemia. For the vast majority of cases only I of 2 or 3 drugs is necessary or desirable. Shotgun therapy with multiple drugs usually is unscientific, often means that the doctor does not know what he is doing. invariably is more expensive for the patient and not infrequently results in tragedy both for the patient and for the doctor. An example of such shotgun therapy is the use of drugs containing vita- mins, including B12 and fohic acid, along w-ith iron, thereby masking bleeding from an otherwise asymptomatic neoplasm until the tumor has become incurable. Many students believe that folic acid administered to patients with pernicious anemia may precipitate an acute and serious exacerbation of the neurologic symptoms. There are about 450 vitamin preparations currently on the market. Many of these are so-called fortified vitamins. In addition to this, milk, bread, fruit juices amid other foods are today fortified with vitamins. Many of the preparations con- tam a dozen or more items, including copper, iron, cobalt, iodine, phosphorus, cal- cium, manganese, molybdenum and zinc, in addition to 8 or 10 vitamins. Every year I am told that Americans buy over the drugstore counters about $250000000 worth of vitamins. It is safe to say that at least $240,000,000 of this is wasted. No reason whatever exists for the taking of vitamins by any healthy adult American on an adequate diet. There is good reason often for correcting PAGENO="0317" COMPETITIVE PROBLEMS IN THE DRUG INDUSTRY 1867 the diet of people who have faulty eating habits. The giving of vitamills in no sense is a substitute for a faulty diet in an otherwise normal person. Vitamin Br~ is being pumped into people by the gallons all over the country by doctors who are not aware of the fact that vitamin B12 is of no value wliatcver except in one group of rare diseases, the macrocytic anemias. My plea is merely the restatement of a very old and sound principle: no mcdi- cine, and that includes vitamins, should be prescribed unless there is a good reason for doing so. When drug mixtures are prescribed, one should remember that one is getting into dangerous territory. Fifty years ago there were over 300 medical schools in this country. A study was made. More than 200 of these schools were found to be poor and were forced to close. The American Medical Association did much to promote that improve- inent in medical education. Today, many thousands of useless drug and vitamin preparations exist, thousands being duplicates under misleading names. Doctors, patients, the proprietors of legitimate drugstores, the people generally and hos- pitals deplore this situation. Exploitation of the public by the existence of such a situation constitutes an important item in the high cost of medical care. Who is going to devise a remedy for the insidious disease? [From the New York Herald Tribune, Thursday, April 14, 1960] Ex-Dauo OFFICIAL CHARGES "QUESTIONABLE" PRACTICES WASHINGTON, April 13 (UPI).-The former medical director of a major drug firm accused the drug industry today of profiteering and other questionable prac- tices and urged Congress to crack down with restrictive legislation. Dr. A. Dale Console, of Princeton, N.J., told Senate investigators he saw little hope of the drug companies dropping practices he said they engaged in. Dr. Console, medical director for Squibb Laboratories until he left to re-enter private practice, said that "unless sweeping reforms are instituted, a truly ethical (drug) house cannot survive in the present competitive wrangle." CHARGES LISTED In a severe indictment of the industry, Dr. Console charged before the Senate Anti-trust subcommittee: That doctors and the public are subjected to a constant "barrage" of new drugs, some of which are worthless and others which have "a greater potential for harm than for good." He said that "since so much depends on novelty, drugs change like women's hemlines." That more than half of the drug companies' research e~ort is directed toward projects that are really not worthwhile but "are pursued simply because there's profit in it." That the industry has high-pressure sales techniques based on the maxIm, "If you can't convince them, confuse them." That most medical leaders and educators "face the problem with denial, com- placency or a sense of futility" because the industry "is unique in that it can make exploitation appear a noble purpose." The subcommittee also heard Dr. Frederic H. Meyers, of San Francisco, a University of California expert on drugs, challenge the American industry's con- tention that it leads the world in discovering new drugs. "Far from leading in drug progress," Dr. Meyers said, "it appears that our industry has usually followed and often after a clear `lag." He said that much of the laboratory work by American drug firms was centered on "exploiting and marketing" foreign discoveries. Because of `this, he belittled the American industry's assertion that the cost of its research justifies high drug prices in `this country. Dr. Meyers also criticized practices used in advertising new drugs. He said many drug ads were "at best incomplete and at worst dishonest." "Some ads become so expensive that they approach `payola'," he said. PAGENO="0318" 1868 COMPETITIVE PROBLEMS IN THE DRUG INDUSTRY [From Time Magazine, May 26, 1961] Too ~L&xv DRUGS Prescription drugs would be cheaper and more effective if manufacturers would market fewer of them, says Dr. Walter Modell of Cornell University Medical College, one of America's foremost drug experts. He also believes that this is the way to bigger profits for the companies. Writing of pharmaceutical chemists in Clinical P1~armacology and Tlicra- pea tics, Dr. Modell asked: "Will they realize that there are too n1any drugs for the patient, for the physician, and, surprisingly enough, for the pharmaceutical industry?" No fewer than 150,000 preparations are now in use, of which 90% did not exist 25 years ago, and 75% did not exist ten years ago. About 15,000 new mixtures and dosages hit the market each year, while about 12,000 die off. These figures, says Dr. Modell, reflect the fact that new drugs are often intro- duced not because they are better than existing drugs or because there is a real need for them, but "to horn in on a market which has been created by someone else's discovery." He denounces as "structural roulette" the game of making a minor change in the molecule of a competitor's drug, to get around patent restric- tions, and rushing the resultant analogue to market. He points to one manufac- turer "who sells one drug entity in this country and a congener [close chemical relative] in another country," and argues that "each is the best for the same purpose. Since more than one drug cannot be the best for the same indication, we simply don't have enough diseases to go around. At the moment the most helpful contribution is the new drug to counteract the untoward effects of otl1er new drugs; we now have several of these." Dr. Modell recommends that manufacturers exercise self-restraint by making and marketing only the single best drug for each purpose, and cross-license one another to spread both risks and profits. One of the most successful of all U.S. companies, he says, introduces the smallest number of new drugs and does the least "molecule manipulation." The others, Dr. Modell suggests, should do the same--to their own advantage as well as that of bewildered doctors and patients. Senator NELSON. sow, you comment in the next statement that there are many so-called new drugs coming to the market which represent duplications of existing drugs, combinations, and so forth. Do you have some specific examples of these duplicative drugs and those resulting from molecule manipulation? Dr. SOHIFRIN. May I answer that question this way: The usual sta- tistics cited on this come from tile product survey put out by Paul cle Haen. In this survey that he updates annually, he not only lists the total number of so-called new drugs coming to the market, but a break- down as to those that represent duplicates of drugs existing on the market, those that represent combinations of drugs already on tile market, and then how many of these are clearly new drugs-that is, items containing drugs that have not appeared on tile market previ- ously. The large majority of tile total so-called new drugs come from the duplications and the combinations category. I would be glad to submit the data I have on the De Haen product survey along with this other material. Senator NELSON. Please hand it to the reporter. PAGENO="0319" COMPETITIVE PROBLEMS IN THE DRUG INDUSTRY 1869 (The material referred to follows:) TABLE 6-PHARMACEUTICAL PRODUCTS NTRODUCED NATIONALLY Number of firms New single chemicals Duplicate single products Cow- pounded products (a) Total new products (b) New dosage forms Cumulative total (a and b) 1966 1965 1964 1963 1962 1961 1960 1959 1958 1957 1956 1955 52 65 82 89 108 111 109 106 126 127 126 124 13 23 17 18 28 41 45 63 44 51 42 31 16 23 34 43 47 33 64 49 73 88 79 90 53 73 111 152 180 191 202 203 253 261 280 282 82 119 162 213 255 265 311 315 370 400 401 403 26 22 41 52 84 106 98 104 109 96 66 96 5,618 5,510 5,369 5,166 4,901 4,562 4,191 3,782 3,333 2,884 2,388 1,921 Note: New single chemicals indicates products that are new single chemic I entit es not previously known and de- veloped by one manufacturer. Duplicate single products include products such as dexamethaaone or griseofulvin put out by various manufacturers. Compounded products comprise any product having mare than one active ingredient. New dosage form: if a product has originally been marketed in tablets and is now offered in ampules, suppositories, etc., the latter are considered new dosage forms. Source: Paul de Haen, New York City. Senator NELSON. Now, you also stated in this same paragraph that we have heard abotit or seen cases of ineffective drugs, harmful drugs, drugs marketed without adequate warnings, and dangerous drugs coming to the market as part of tile "new drug flow" for which the industry claims credit. Do you have specific examples of such harmful drugs, those without adequate warnings, and so forth? Dr. SCHIFRIN. Yes, sir; I think the most harmful example of a new drug coming to tile market was MER 29, which was withdrawn from the market because of its harmful effect on people taking it. Also in these hearings, the question of inadequate advertising has come up. I believe tile Chlormycetin case is probably the most im- portant one there. A highly promoted drug came to the market, a potent drug, with dangerous side effects, and the information as to the side effects was fully understood only after the problems of use of the drug had been so serious that it was withdrawn from the market for a year for a very careful study. So I cite that as the most Outstanding example of a drug, a dangerous drug, with inadequate information. There are quite a number of other examples. The Merrell Co. was cited for false claims with MER-29. This is an article from the New Haven Register of December 20, 1963, that I happen to have here. That is just one of several examples I could name. Senator NELSON. If you would submit that last article from it, we will print it in tile record. (The material referred to follows:) [From the New Haven Register, Dec. 20, 19631 DRUG FIRM INDICTED ON FALSE CLAIMS WASHINGTON (AP).-The William S. Merrell Co. was indicted by a Federal Grand Jury today on charges of making false statements to the government about the anti-cholesterol drug, MER/29, and tests made to determine its safety. PAGENO="0320" 1870 COMPETITIVE PROBLEMS IN THE DRUG INDUSTRY The indictment returned here said the company and three of its research scienti~t~ `concealed aiid coeered up, by trick and scheme, material facts imimportant to the Food and Drug Administration." The scientists named are Harold W. Werner, Evert F. Van Maanen and William M. King. Mr. GROSSMAN. Could I ask you a question here? On Page 4, you talked about the proliferation of products that confuses rather than proves drug selection or drugs inadequately tested or whose side effects are minimized in the race to market for the consumer's dollar. Are you criticizing the FDA here, as well? Dr. SCHIFRIN. I am not criticizing- Mr. GROSSMAN. In other words, in order for these drugs to get to the market, they have to somehow get past the FDA, do they not? Dr. SCHIFRIN. Let me read that part of my statement that covers that point you raised there. That discusses the point. Mr. GROSSMAN. Surely. Senator NELSON. We are not above criticizing a Federal agency, you know. Dr. SCHIFRIN. I consider many of these criticisms valid. I offer them as important qualifications of the industry's product perform- ance. But many aspects of this product performance warrants praise, and this must be recognized, criticisms and qualification notwith- standing. The other half of the picture is what I refer to as "market per- formance," and deals with the efficiency with which the industry uses society's scarce resources; that is, the extent to which economic re- sources are used to enhance consumer well being and other economic goals. In this regard, the important questions that must be answered are these: Is there enough competition to place a premium on efficiency and penalize waste? Is there enough competition to compel firms to pursue only those activities that benefit consumers? Can firms incur costs for activities that do not benefit consumers yet charge consumers, in the prices they pay, for such activities? Is there. enough competition to keep prices in realistic relation to costs, providing profits adequate for maintaining or expanding desirable a.ctivities but not profits derived from the exercise of monopoly power? To generalize from the vast quantities of evidence available, I be- lieve that the characteristics of the drug-product marketing and dis- tribution systems are such that effective competition does not pre- vail. As a result, firms are free to engage in many practices-most notably in promotion and advertising but in research and development also-that serve their own profit goals but provide no benefit to so- ciety. A largely wasteful promotional effort costing in the hundreds of millions of dollars per year; misdirected research; rivalry in nov- elty, in capturing the attention of physicians; all this represents costs of large magnitude passed on in full to the consumer, but without any corresponding benefit-and perhaps some harm, such as a prolifera- tion of products that confuses rather than improves drug selection or drugs inadequately t.ested, or whose side effects are minimized, in the race to market for the consumer's dollar. Yet these practices and the factors contributing to them have become part of the industry-woven deep into the design of its fabric. PAGENO="0321" COMPETITIVE PROBLEMS IN T~) DRUG INDUSTRY 1871 Now, if I may comment on that question, I would say that to a considerable extent, in recent years, the Food and Drug Administra- tion has been doing much, much better in watching out for the ad- vertiser than trying to do something about the proliferation of drugs that may not really contribute to drug therapy. The Kefauver-Harris amendment in 1962 made a great contribution in this respe'ct and I think the FDA has administered that law very capably and effec- tively. My main criticism is not so much that the Food and Drug Ad- ministration can cure the problems of this industry as much as that a system prevails,, an informational system, a distributional system- the whole process by which information about products comes onto the market; that is, the information largely comes from the drug firms and it is aimed at the physicians. Thus the new products and the claims about them, the information regarding their use all come from the same source. I believe this is a cause of many of these abuses in product development and product promotion. Now, as far as the FDA, there is `one-there are certain changes in what the FDA can do that will be very helpful and if the FDA has not done the job that I would like it to do, its failure is in not pressing hard enough, I believe, for certain further changes that would make it more effective. Mr. GROSSMAN. There could not: be this proliferation of products, certainly, without FDA approval. if the FDA did not approve it, they just would not be marketing it. Dr. SOHIFRIN. That is true, the FDA-as far as the proliferation of products goes, that is true. The FDA is playing a permissive role in this, certainly. But understand me, when I cite here the large number of new products, I am not really criticizing the industry's performance just because of the quantity of its new products here. I am criticizing the new products in this respect; the industry takes credit for putting out a tremendous number of new products and representing these things as therapeutic advancements. Most of these new products are not therapeutic advancements. They are new items duplicating those already existing on the market. These may be useful for competitive purposes, but it represents no advancement of drug therapy. Thus if the industry is saying, "Look, one of the manifestations of our research and development effort is this tremendous number of new products," we are concerned with the industry's operations and must understand just how much of a contribution these are or are not. Senator NELSON. If I may interrupt the minority counsel, the Kefauver-Harris Amendment of 1962 for the first time gave the FDA the authority to disapprove a New Drug Application unless the drug was proved efficacious; is that correct? Dr. SOHIFRIN. Yes. Senator NELSON. Now, if there is a molecular manipulation that gives you a product, which has the same therapeutic result as an already existing drug, does the .FDA presently have the authority to deny the company the right to market that drug? Dr. SOHIFRIN. No, sir. Before 1962, the FDA could deny a New Drug Application only if the product were harmful-that is, if it did not hurt the patient, that product could be marketed. Now a product has to be efficacious. But it does not need to be of superior efficacy to prod- ucts already on the market to be cleared by the FDA. 81-280-68-pt. 5-21 PAGENO="0322" 1872 COMPETITIVE PROBLEMS IN THE DRUG INDUSTRY Senator NELSON. If it. is just a product differentiation, as the phrase is used- Dr. SCHIFRIN. Yes, sir. Senator NELSON. And if it does duplicate a product on the market and is efficacious, does the FDA have the authority to prevent the firm from marketing it? Dr. Scrn~mN. No, sir. Mr. GROSSMAN. One further point on this: If it is efficacious and if the FDA is agreeable to let it. go on the market, you pointed out when you used the word "it will provide com- pet.it.ion"-isn't that something we are trying to do here? In other words, if it is going to be a competitive drug, does it not have some value as far as our inquiry is concerned? Dr. SCHIFRIN. That may be a contribution. Mr. GROSSMAN. That is a pretty important point. Dr. SCHIFRIN. Yes, but it may be offsetting negative contributions. In other words, these drugs that. are duplicated may be very heavily promoted. I think that is a waste of resources that could be used for other purposes. Secondly, they may be promoted under brand names, which creates confusion and the fact that they are duplicates may be obscured in the promotional literature. They may be represented as new therapeutic factors when in fact, a large majority are not. Senator NELSON. The only way in which a company can come onto the market with a duplicate but differentiated product is if a patent has expired on the one they are duplicating. Is that not. correct? Dr. Scrnr~IN. I would say that is the usual way. There are rare circumstances under which it could obtain a license, usually as the result of some antitrust activity. Senator ~ELSON. But that is a license from the patentholder? Dr. ScrnFRIx. The patentholder, yes. Senator NELSON. But usually in that case, if it is a differentiated product, it is a product, as I understand it, of the same chemical com- position with some insignificant difference and then is put on the mar- ket and advertised to the medical profession as something better or something------ Dr. SCHIFRIN. Yes, sir, the differentiation is nominal. It is not in its chemical composition and many opinions maintain that it is not in its therapeutic action, either. It is a nominal differentiation, and therefore the professional thrust is to make a claim that this is a new and better product. But it is really an identical product to many products already on the market. Senator NELSON. So that the reason for doing it really is to benefit the company in a competitive situation? Dr. SOHIFRIN. Yes, because if the product is made~ from a drug whose patent has expired or is not patented, there are likely to be many such items on the market and an item comes out to fill out a company's cata- log and will get promoted. The company, hopefully, will want this thing selected by its brand name. They will put a high price on it., very often; but it does not give the doctor additional medicine to use for his patients. Senator NELSON. What I do not have clear in my own mind is whether product differentiation by definition is another product with PAGENO="0323" COMPETITIVE PROBLEMS IN THE DRUG INDUSTRY 1873 the same chemical composition, as was testified yesterday-I assume that is correct. Is that what a product differentiation is? Dr. SCHIFRIN. I would not define it that narrowly, as an economist. Senator NELSON. But in any event, as to the chemistry involved, I think the testimony was that a differentiated product was of the same chemical composition as the one that isbeing duplicated. Dr. SCHIFRIN. That is a common usage of the term, yes. That is the way it has been used. Senator NELSON. What I do not have clear is why should the com- pany go to the bother of creating a product differentiation, whatever that really means, when all they `would have to do is just duplicate the product on the market, the same compound, and give it their brand name and advertise it on the market under their brand name? `What is the reason not following that procedure instead of going to what is called product differe.ntiation~? Dr. SCHIFRIN. `Well, Senator, I think there are two types of product differentiation. One is an actual difference in the product, a physical difference. That may be a small difference. That is what the term "molecule manipulation" refers to, to differentiate a product, chem- ically, and that is done very often, tO get a patent on a product that is very much like the product of some other company that is already patented. That is a way of entering their market. The second way of differentiating a product is really to produce something just like the other firms are producing but to differentiate it by calling it some other name, a trade name differentiation. Why would a firm do either of these? `Well, to have the physical differentiation, they would do it for the sake of acquiring a patent and then having an exclusive item that they could promote under a trade name, hopefully have it prescribed heavily and have these pre- scriptions filled, as they are at a very high price. Perhaps a company duplicates a drug that is already on the market; first of all if it were a small firm, in the hope that through being an active price competitor, it could gain large sales. A large firm would do it in hopes of making that trade name stick and thus having PIIY sicians prescribe it by its trade name designation and those prescrip- tions would be filled at much higher prices and thus it would be profit- able for the large firm to attempt this nominal differentiation, if it could make that trade name stick for this differentiated product. Senator NELSON. So in this case, where they simply duplicate the drug already on the market, the differentiation involved is only the differentiation so fa.r as identification by a brand name is concerned, is that correct? Dr. ScmFiuN. Yes. Senator NELSON. And the hope in that case is that by effective ad- vertising, they will be able to get into the market and compete. Is that it? Dr. ScmFiiIN. Well, to get into a market and compete, but also, here we get into the whole practice of trade names, by promoting under a trade name, of course, they are popularizing the specific trade name they are promoting. They are also engaging in promoting the practice of prescribing by trade names Now, the promotion serves both of these purposes. There may be many, many firms producing that i~roduet. But only the larger ones PAGENO="0324" 1874 COMPETITIVE PROBLEMS IN THE DRUG INDUSTRY that can popularize these trade names become the effective ones and thereby become insulated against the generic firms who are in the market and selling at a lower price. So there are very large profits to be gained by a larger firm entering the market, even for an unpatented product. Senator NELSON. So then you are describing a situation where a company, because of its distinction, can then effect a nominal differ- entiation simply by the use of a brand name and then, by effectively advertising its brand name to the medical profession-because that is where they advertise-convince the majority of the profession to pre- scribe that brand name. Once a firm is able to do that, it can artificially set the price higher. Is that what you are saying? Dr. SOHIFRIN. Yes, sir. Senator NELSON. And continue to compete because there is no price competition. There is merely promotional brand name competition in the head of the physician who knows the name? Dr. S0HIFRIN. That is right. Senator NELSON. Is that what you are saying? Dr. SCHIFRIN. Yes, sir. The people who have appeared here, many of them, have referred to this, the competition that you are talking about, as the competition for the ear and the eye of the physician. Senator NELSON. To give you a hypothetical case, you could very well have an unknown minor company that is not established by repu- tation-that is, by name reputation. It may be a very fine company, but it is not known widely to the medical profession. That small com- pany could discover a new and very good drug and go into the market- place with that drug. Even though they will get some kind of a reputa- tion based upon this one drug, it would not be a reputation that could match the reputation of 50 or 100 years of advertising and standing in the medical community achieved by the big companies. So then theoretically, you could have a case where a brand name company that is established, could move in, once the patent has expired, and get a position in the marketplace because of their acceptance by the medical profession, charge an artificially high price and occupy a large per- centage of the market even though they would be selling exactly the same product as the small firm which developed the drug. Dr. SCHIFRIN. Yes. Senator NELSON. Even though the originating company may put their product out at a much lower price and other generic companies may come in at a still lower price. Is that what you are saying? Dr. SOHIFRIN. Yes, and going beyond that-the example posed by this hypothetical case, is something that could result after the patent had expired. But supposing a smaller company did have this development you are talking about. The larger company would not have to wait until the patent expired. Through molecule manipulation, they could come up with a drug with a different chemical structure which would war- rant a separate patent, but still be very close to the drug marketed by the smaller company. And shortly after the development made by the smaller firms, the larger firm could come out with its product, which is not an improvement now, and through very heavy promotion, take away even those initial gains that the small firm hoped to make. So the PAGENO="0325" COMPETITIVE PROBLEMS IN THE DRUG INDUSTRY 1875 small firm may be jeopardized by the promotion of the large firm during the period when the smaller firm's patent may be valid. So it could be effective around it, but not over it. Senator NELSON. You think that the reason the major companies continue to emphasize very heavily the brand name rather than the name of the company is that they can effectively compete, then, in the retail market, where there is no price competition. They can compete effectively against equivalent quality generic drugs that are available at a much lower price. Dr. SCHIFRIN. I would like to change that a little bit and I will agree with you 100 percent. The large firm can keep the small firm from competing with it, because the large firm, you see, which has popular- ized trade name use, renders the small firm, which can't promote under trade name, an ineffective competitive factor. I think we are saying essentially the same thing. Senator NELSON. You are aware that very frequently, the generic drug producer does compete very effectively against a major firm in competitive bidding situations; the city of New York, for example, purchasing for the public hospitals and the welfare department, asks for competitive bids and is well equipped to test the drugs-and generic firms do compete in that situation. Dr. SCrnEPJN. Yes, sir. Senator NELSON. Or would it be more apt to put it the other way around, that in that area, the brand name companies decide to com- pete with the generic firms by reducing their prices substantially. For example, we have cases where within the same 2- or 3-month period, a brand name is being sold across from city hall in New York to a community pharmacist for $8 a hundred and so forth, and then in the same period, the same company moves in and bids to New York City at $1.10 a hundred because they know they have to meet the competi- tion the other brand names who are willing to compete, pius the generic people who are in that market, too. Dr. SOHIFRIN. Yes, sir. And I come to this point in my statement, but to the extent the large firms can shift the competition from price to promotional com- petition-that is, they can shift the burden from the selection of generic equivalents on a price basis-to have that selection done on a trade name basis, they render the small firm very impotent in competi- tion, you see. Perhaps this chart-~- Senator NELSON. Just one moment before you get to that. Does that explain in part, at least, why the highest price charged by the manufacturer in every example I have seen, is always the price charged to the community pharmacist? In other words, if you look at a listing of all drugs by all companies-I have seen no exceptions-at the start you will see that the price charged the pharmacist is the highest price charged in America. Then you look at bids all over :the country to the Defense Supply Agency, the Veterans' Administration, hospitals all over America, cities and counties-and in almOst every single instance, the price asked by the same company for the same drug is lower, many times one-fifth. one-tenth, one-twentieth as low as the price they charge to the community pharmacist. And in those instances where we find a PAGENO="0326" 1876 COMPETITIVE PROBLEMS IN THE DRUG INDUSTRY purchaser in a nonprofit institution, a municipal hospital, paying the same price as the pharmacist does, if you will look a little further, you will see that they did not purchase on a system of competitive bidding. Dr. SCHIFRIN. Yes, sir. Senator NELSON. So, the brand name identification in the mind of the doctor, in part at least, accounts for the higher price consistently charged `the community pharmacist. The industry is not afraid that an unknown firm's drug would be prescribed. Is that the situation? Dr. SCHIFRIN. Absolutely. Mr. GORDON. May I interrupt at this point? Dr. SCIIIFRIN. Yes, sir. Mr. GORDON. Can I conclude that even if a small company has a patent.. it cannot. really protect itself against the large company? Dr. Scmrnix. That is true, because as I say, the large company can frequently obtain a. patent on a therapeutic equivalent, of slight chem- ical difference, you see. The difference in chemical structure would warrant the patent application and then it would have a therapeutic agent. here to rival that. `of the small firm and through its promotional effort could in fact ge.t the bulk of the market. and the small firm, although having come up with a significant development and having acquired a patent, may not succeed very well financially in obtaining innovational profits for t.he accomplishment. Senator NELSON. If a. new chemical compound is discovered by a small company and patented, can another firm actually take that com- pound, do some molecular manipulation, and end up patenting the slightly modified product? I have some serious doubt about that with- out. knowing anything about it.. Dr. SCHIFRIN. Well, it seems to me that when the small firm acquires the patent for its development, the pat.ent comes at. the price of signifi- cant disclosure. That. disclosure, of course, can serve as `the basis for a structural change that the large firm might accomplish to come up with its rival therapeutic a.gent. Secondly, I think a lot of the development, occurs simultaneously. That. is, there is a lot, of literature in the professional journals about progress, technology, chemical manufacture a.nd so on. It may only be that. the small firm may beat the large firm at certain times just by a very narrow margin in the development of its product. The large firm is almost up to the small firm and when it sees the small firm has a patent, it ca.n do a little sidestep and obtain a. patent. on a slightly dif- ferent chemical substance. Mr. GORDON. Do you have examples of that? Dr. ScmFRIN. The only case I can think-it is a little bit different. For example, meprobromate. We know that. it has been an e.normous financial success for Carter and Wyeth. But me.probromate was actu- ally discovered by Dr.. Berger in England. Rather `than initiate pro- duction as a small manufacturer, he went with his discovery into the employment of Carter and entered the market through a large firm. I think this reflects the very unstable and uncertain posi- tion that `a small firm making a development would face in the market. That was avoided in that case by going directly to the large firm. Senator NELSON. In some of `these developments where a number of researchers and companies around the world are working on the PAGENO="0327" COMPETITIVE PROBLEMS IN THE DRUG INDUSTRY 1877 same problem, seeking `the sathe solution, progressing at a similar rate, is `there evidence tha~t because of the fear of a legal contest of the patent by one versus the other, they reach some agreement th'at one will get the patent and that `he will license the others? Dr. SCHIFRIN. Very much so. When Patent Commissioner Ladd appeared before the Kefauver committee, he poInted out that patent interferences-that is, where companies contest a patent application made `by another company-are more frequent in chemical and chem- ical-related industries than any other industry. That is fact one. *T'he second fact, and this came from, I think, Mr. Federico, who was with Commissioner Ladd, is that a patent interference results in a very careful scrutiny of a patent application. In a high proportion of `these cases, it turns out that the result is unp'atentability. That is, a decision that the patent should not be awarded. Large firms are researching along these similar lines. They attempt to test the patent application, because rewards for getting a patent. are very high in this industry. Hence, the major firms that have pursued parallel lines of research, often permit a patent applicant to go ahead unchallenged and to get his patent. But the other Targe firms, for not challenging that patent, will get licenses to produce it. In other words, they will share the patent among three or four firms rather than have nobody get the patent. So licensing is a result of `these agreements not to test patents by the large finns. Senator NELSON. When you say they share the patent., you do not mean that three companies, more than one company, gets the patent in their name? Dr. SCHIFRIN. No, but as a condition for withdrawing from the in- terference, the companies that withdraw get licensing privileges. Senator NELSON. Is that a violation of the antitrust laws in any way? Dr. Scrnn~IN. Well, I would say that in a considerable number of these cases, t'he licensing agreements have been accompanied by price- fixing agreements. We have seen that in `meprobromate, and `in tetra- cycline. We `have seen it `several times t'hrough the industry, that the cooperation involved in eros's-licensing is a v'ery tempting circumstance to lead to a-if not an overt, at least a tacit price-fixing agreement. The tetracycline is a notable example of these things. Senator NELSON. I notice tha't there is more than one type of case. I have seen instances w'here one company `gets the patent and licenses two others who have been working in the same area, doing `parallel research. T'hen there are cases where one company gets the patent and licenses anybody wh'o wants to b,e licensed for all practical purposes. Dr. SCHIFRIN. T'hose are rare. Senator NELSON. There are some; are there not? Dr. `SCHIFRIN. Yes, reserpine, I `think, is the outstanding example. Mr. GORDON. Dr. Schifrin, if the large companies will not test the validity of `the patents owned `by other large `companies, an'd if small ccmpanies can't do it because of the litigation expenses, then who is going to protect the public against the possibility of invalid patents being used `to reap `monopoly profits? Dr. ScrnriuN. Mr. Gordon, as of now, the answer to `that is nobody does. The FT'C, in its tetracycline case a few years ago, under section 5 of the Federal Trade Commission Act, `d'id consider a fraudulent pa- PAGENO="0328" 1878 COMPETITIVE PROBLEMS IN THE DRUG INDUSTRY tent application to be an iuifair method of competition. So to the extent they are willing to do it, they can. They `have been the only ones who have don~ thi~ and of eour~ fhe~e are infrequent examples. There is no regular mechanism by which the true validity of patents can be watched over and ascertained right now. One of my proposals is related very closely `to that point. Senator NELSON. You mean, then, that if a company has what ap- pears to be a new product and there is doubt in the minds of other com- panies that it `is really patentable, in order to avoid free competition, they just do not challenge its patentability; so that it then becomes patented and some compa.nies become licensed and then, for 17 years, there is an artificially high price paid by the consumer because there a patent has been granted, which, if challenged, would not have been granted in the first place? Dr. SornriuiN. Correct. In other words, market domination by a few firms that exercise a monopolistic type of interest comes to exist, yes. Senator NELSON. And you say there is no effective protection in behalf of the public by any public agency against this kind of occurrence? Dr. Scrn~'Rn~. At present, there is not. Senator NELSON. It rather shocks me that you could have a situation where something that may be shown to be not really patentable if challenged becomes patented and no public agency is prepared to pro- tect the public interest. Dr. SoirFnm-. Well, the most careful scrutiny of a patent's validity occurs in the court, not at the Patent Office. Thus, things for which patents are issued by the Patent Office may, if contested, be found to be invalid. The patent may be found to be invalid if contested. Thus the burden is put on usually some small firm. And the small firm, as Mr. Gordon correctly pointed out, usually can't bear the expense of litigation. Senator NELSON. And you are advocating that some mechanism to protect the public interest in this kind of instance be established? Dr. SOHIFRIN. Yes, sir; I think an advisory panel to the Patent Office, reflecting the highest degree of expertise in the medical field should be established, and should give these patents the very careful scrutiny that would never arise otherwise. Senator NELSON. Are you satisfied that there is no mechanism under the present law by which any Federal agency could, on its own motion, challenge the patentability of a product? I do not mean challenge a conspiracy to monopolize or any such thing, but challenge just the patentability of a product? FDA, Federal Trade Commission, anybody? Dr. SOHIFRIN. I am not positive as to the interpretation that these agencies give their own enabling legislation, but I will say this: If any such organ exists, it certainly is not used rigorously, if at all. Senator NELSON. Do you think that your proposal is feasible as a practical matter? That is, would it really put into the hands of public representatives adequate tools to protect against this specific circumstance? * Dr. SOHIFRIN. Yes, sir; becanse I believe that much of this exper- tise, and from time to time, several related important functions, have PAGENO="0329" COMPETITIVE PROBLEMS IN THE DRUG INDUSTRY 1879 been filled by people who are in private industry-I should say mainly in medicine. Many of our people who are outstanding physicians, who teach in medical schools, certainly are competent to assist in this matter. Senator NELSON. You certainly would not give this kind of group final authority; would you? Dr. SOHIFRIN. No; but I would hope, sir, that their expertise would be recognized by the Patent Office in determining patentability; that these people serve as a rather influential body of experts. Senator NELSON. And what is the next step in the event that this independent body that you would establish made a recommendation which they felt very strongly about but was not accepted by the Patent Office. Would you provide a mechanism then, which would allow the Government to go into court to challenge the patent? Dr. SOnIFRIN. That is one alternative. I would consider that as feasible; yes. There is another alternative that is coordinated with that one. I know we have wrestled with the problem of patent standards very often in our industry, and since the drug industry occupies a rather unique position in American industry, I would not be opposed to a unique patent law applying to drug patents, with clearly different and higher standards than are now provided by our general patent law. I think the drug industry warrants separate treatment. Senator NELSON. If this is, as I believe, a serious problem, and it may very well be, would you not be reluctant to place the final arbi- trary authority in a commission or~ committee such as you recommend, or even in the Patent Office itself? In other words, would you not think that if you had a commission and they made a recommenda- tion that an item was not patentable and the Patent Office did not agree with that conclusion, that the legal arm of the Government, the Justice Department, ought to be able to move to test it? Or if the Patent Office did agree and the patentholder did not, they still ought to have a right to go to court? In other words, should you not have as the final resort for both the Government and the company or companies involved the right to go to court? All you are trying to suggest here is that the public interest be protected by establishing some agency with the authority to challenge the patentability of drugs and with the ultimate authority, in behalf of the public, to recommend to the Justice Department that the patent issuance be challenged in the courts? Dr. S0nIFRIN. Ultimately; yes. Senator NELSON. Go ahead. Dr. SomriuN. I have mentioned about the practices and the factors that have become a part of the industry's operation. Now let me de- scribe how this design came into being, this product and market per- formance that I have reviewed now. In the late 1940's and early 1950's the major drug firms became alarmed by vigorous price competition in the sale of drugs, most notably penicillin and streptomycin. They sought to insulate them selves from further price competition for these and other products through the device known as the "specialty" item-that is, finished PAGENO="0330" 1880 COMPETITIVE PROBLEMS IN THE DRUG INDUSTRY products sold under unique trade names and, to the maximum extent possible. made from drugs for which patents had been acquired. These two devices-the trade name and the patent-have., to an astounding degree, eliminated competitive conditions from the markets in which drug products vie for sales. The vast majority of all new drugs have been covered by patents, sometimes of questionable validity. Patents monopolize the sale of products for a single seller or, in some cases, for it and a limited number of licensees. Even for products sold by many firms, the popularization of the use of trade names in prescription writing has led to the dominance in those markets by the few major companies in that product line who can promote through massive advertising outlays the trade names of their specialities. Monopoly or oligopoly has thus become the usual situation-in the antibiotics, the hormones, the mental drugs-and down the list of areas of drug therapy. The competition among firms in the market is no longer one in which producers of comparable items seek customers through more attractive prices, but one in which a single seller often exists alone, or, if lie shares the market with a few rivals, that rivalry is in advertising, claims, and trade-name repetition-certainly not in price. This is a large, costly, and wasteful competition. It is excessive, con- fusing, and largely ignored, but it does serve its purpose. It. popularizes particular trade names and strengthens the use of trade names in gen- eral, thereby rendering price competition ineffectual. Such advertising a.dds little if anything to drug therapy, yet is a large cost, easily shift- able to the consumer. To some extent the peculiar dependency of the physician on the drug firm for information places some of this activity above ordinary adver- tising-~but even this "promotional" literature often has contained mis- leading, insufficient, exaggerated, or otherwise inadequate information. The same features that e.ncourage such wastes and enable drug firms to charge their costs fully to consumers also condition the profit margins contained in drug prices. These margins provide for rates of return to the industry as a whole and to all the large drug pro- ducers far higher than for the economy as a whole, year-after-year. Furthermore, these extraordinary profit levels are remarkably stable. Year after year, the ethical drug industry is right up at the top of the industry profit ranking. Year after year the after-tax rates of return on stockholders' equity for large drug firms exceeds those of other firms; year after year the ranking in "Fortune" show a. dispro- portionate frequency of drug firms among the most profitable large firms in the economy; year after year the FTC-SEC profit data in manufacturing reveals the outstanding performance of drugs. The high profitability reflects the absence of competition. The stability of profits demonstrates the absence of risk to investors. If risk were to~ exist, one would expect to see the high gains of some firms accom- panied by occasional losses-to themselves or to others-but such evi- dence of risk is virtually nonexistent. Unable to exploit the "risk" justification for its profit levels, spokesmen for the industry have tried to rationalize the great profits with the "research" argument- that high profits are necessary to finance the industry's research effort. PAGENO="0331" COMPETITIVE PROBLEMS IN THE DRUG INDUSTRY 1881 This is an illogical theme, however, since profits exist only after all costs, including research, have been covered. A firm that breaks even, or earns only a normal profit, is financing its research just as fully and capably as is the firm earning profits far aibove normal. As that argument folds, the industry turns to the "growth industry" argument-that is, that the drug industry is a so-called growth indus- try and that high profits are needed to finance that growth. In response to that theme I contend that reinvesting your earnings instead of taking them as dividends is one thing; but exacting from consumers a double profit and more, to sustain and even increase dividends, on the one hand, and simultaneously to finance capital expansion and stockholders' equity on the other hand, is uneconomic, unjustifiable exploitation. The consumer who purchases his prescription is thus pay- ing in that price for these things-for activities necessary to bring that product to him; for activities that provided no benefit to him and hence were unnecessary; an attractive dividend to the stockholders, and some part. of a new machine, a new plant, a piece of property that will belong to the stockholder. The dramatic expansion of the industry in the past 15 to 20 years has been almost entirely, if not totally, financed out of profits-yet the industry throughout this period has maintained a payout rate., dividends per invested dollar or per share of stock, that compares favorably with other industries. That, to me, is something like having your cake, eating it, and seeing it grow bigger all at the same time. Thus the market performance of the ethical drug industry, meas- ured against the criterion that "prices paid by consumers reasonably reflect. the costs of efficiently providing useful activities" is seriously deficient. The prices paid by most consumers of drug products are excessive for two major reasons-they are inflated by wasteful cost elements, and they are further inflated by the excessive profits they provide. The question that now arises is this: What features of drug markets render consumers so exploitable? First, there is the peculiar importance of the product, more so than almost any other commodity; then there is the "prescription relation- ship," in which someone other than the consumer actually decides what will be bought-someone who may be unaware of the availability of alternative products, unaware of their relative prices, or indoctri- nated in the practices of prescribing high-priced trade name specialties. Second, given this vulnerability of consumers, to exploitation, is monopolization-the basis of the power to exploit the consumer. This monopolization is both result and. cause of the wasteful competition in development and especially in advertising, and is a prime determinant of excessive profits. all of which the consumer bears. The large drug firms, as I have indicated, strive to create monopoly through patents and trade names. But what is it that permits them to succeed so impressively in that endeavor? To answer that question we must distinguish two separate facets of the industry's operation-on the one hand there is the development and manufacture of drugs, i.e.,,: the active chemical substances that go into pharmaceutical preparations. On the other hand there is the PAGENO="0332" 1882 COMPETITIVE PROBLEMS IN THE DRUG INDUSTRY manufacture of the finished dosage-form products or preparations containing those drugs. As breakthrough discoveries have opened up the fields of antibiotics, cortical steroids, mental drugs, antihistamines, vitamins and nutrients, antidiabetic drugs, and others, a pattern of specialization has emerged. Because of differences in their back- grounds, in their interests, in their early activities, and also because of their inability to pursue many diverse research lines simultaneously, the larger firms have focused their efforts and resources on one, or perhaps a few, areas of exploration. In most of these areas, perhaps with the exception only of mental drugs, developments after the breakthrough discoveries have been of a "stepladder" nature. Those firms first on the ladder have kept ahead of others; entry on the ladder is impossible for smaller firms and extremely difficult even for large firms once the early entrants have gained momentum and ac- quired patents, thereby protecting all the previous rungs as well as the ones they are now perched on, from being used by potential rivals. Thus, scarcity of opportunities for entering into the manufacturing of drugs-the chemical substances-has closed entry into the man- ufacture of ~nished products, except in those few cases where the bulk drugs are available to all fabricators of preparations containing them. In most drug product markets the only sellers are those large firms who hold the patents on the drug ingredients. In a few cases other large firms are licensed to use the patents; and the many small firms in the industry are thus limited to the manufacture and sale of products con- taining unpatentable drugs, drugs whose patents have expired, or drugs which are available because licensing by the patent holder is required as a consequence of antitrust guilt. But even in these few sit- uations, the small firms are not equal competitors of the large ones- because the popularization of trade names (possible only for large firms) in prescriptions renders the generically designated items of the small firms as inconsequential competitive threats. As a result of this specialization, patent acquisition, and the use of trade names, there have developed two different groups of drug- product markets, which I have depicted in this diagram. Across the top, that says, I believe, "Products Made from Drugs that are," and then the lefthand column "Patented," in the righthand column, "Tlnpatentecl." Since most drugs, most commercial items, a.re patented, I have that column to the left larger than the one to the right. Now, reading down the column, the columns then show the products made from drugs that are patented on the left and on the right, the products made from drugs that are unpatented. Reading across, first we have the private prescription market. This is the market in which, you know, the family physician prescribes a medicine and the prescrip- tion is filled at the drug store. Down below, I have the institutional market. Since approximately 70 percent of drugs-maybe 60 percent now-are sold in the private prescription market, I have made the first row across larger than the row below it. PAGENO="0333" PRO]XJCTS PATENTED CONTAINING DRUGS URPATENTED DRUGS A A where trade names . A - A are used where generic names are used . B Dr. SOrnFRIN. Now, what does this show? Those areas on that dia- gram that are marked with "A" shows where monopoly power exists. If a product-that is, a preparation-is made from a patented drug, there will be very ~few sellers, perhaps only one, of that drug. Regard- less of whether that drug then is purchased in the private prescription market or in the institutional market, there cannot be price competi- tion because of an unavailability of competitive alternatives. Thus, all the items in the first column will be "A," the absence of competition. In the second column, we have drug products containing unpatented drugs. There will be many sellers here. But in the private prescription market, where selection is done by trade name, only those few large firms that can promote trade names are effective sellers. That is, they dominate the market. So they can continue to charge the high prices that accompany trade name products and thus, even though these products contain unpatented drugs, there is an effective monopoly or oligopoly nonetheless. Now, for products containing unpatented drugs sold to institutional purchasers, I have both "A" and "B" there. If the institutional pur- chaser for some reason or other continues to do his prescribing, his purchasing, by trade name terms, then the small firms are still cut out of the market. Thus, only in the institutional m~trket, only for drug products made from unpatented drugs, and only where the purchaser-that is, the institution-considers these items to be generic equivalents and goes on a price basis-only in that small area, "B," is there a price competi- tion in the market for ethical drug:products. That represents approxi- mately 5 percent of all drug sales in this country. In other words, all the areas in "A" represent 95 percent of dollar volume of drug sales. The competitive sphere, "B" is 5 percent, approximately. The largest two dozen firms account for the approximately 95 per- cent of total industry sales represented by markets labeled "A"; the remaining hundreds of smaller firms share the 5 percent of industry sales accounted for in market "B." The past history of the industry can be described as largely one of creating and expanding "A" relative to "B." The continuing goal COMPETITIVE PROBLEMS IN THE DRUG INDUSTRY 1883 (The diagram referred to follows:) PRIVATE PRESCRIPTION MARKET 1INSTITIJTIONAL ~4ARKET" PAGENO="0334" 1884 COMPETITIVE PROBLEMS IN THE DRUG INDUSTRY of the large flrms has been the enhancement of their individual and collective profitability by preserving and expanding those sectors oi tue market where effective competition does not prevail. This o~oa.l is sought in these ways-(1) by mnakin~ sure that the new drugs ~at replace older ones, particularly unpatented ones, are protected by patents-even though their patentability might not stand close examination. Mr. GORDON. You stated before, Dr. Schifrin, that there is a sort of gentleman's agreement among the large firms in the drug industry that., "you won't question my patent and I will not quesHon your patent." Is that a fair summary of your position? Dr. SCHIFRIN. That does not summarize my entire position. Mr. GORDON. I mean in the existing situation. Dr. SCHIFRIN. That is part of the existing situation, Mr. Gordon. The second way in which the large firms try to enhance situation "A" is by avoiding patent interference suits, which are fairly frequent in the chemical technology industries because of parallel research by the specializing firms, through often elaborate cross-licensing agree- mnents. These agreements, over and above the price fixing that not infrequently has accompanied them, are enacted to avoid the pos- sibility that a party losing an interference suit may challenge the patentability of the discovery. Mr. GRoss~IAN. Dr. Schifrin, could I ask you to he specific about this? You stay that these agreements were made "over and above price fixing that not infrequently has accompanied them." That is a strong allegation. Could you be specific about this fre- quent price fixing? Dr. SCIIIruIN. Well, in recent years, two of the most successful drugs that have been developed, conmiercially successful and quite important with therapeutic significance, have been meprobrornate, the tran- quilizer, and tetracycline, the broad spectrum antibiotic. In the case of meprobromate, Carter had the patent. Wyeth had the detail men, et cetera. Carter entered into a licensing agreement with `Wyeth for them to also be able to sell meprobromate and to promote it very ex- tensively. Thus for many years the only two meprobrornates on the market were Equanil and Miltown. Price fixing in these two com- panies was found by the courts in recent years and compulsory licens- ing of meprobromate resulted from that case. The second of these two examples I use is tetracycline. Just recently, of course, the Federal court found that there was a Sherman Act violation as a result of price fixing. The Federal Trade Coimnission case that I cited earlier tied the price-fixing agreement very closely to the cross-licensing agreement that brought the five sellers into the market. Mr. GRoss~rAN. I am aware of those cases. You used the words "not infrequently." That is why I wondered if you had other examples and whether the Justice Department has prosecuted either civilly or criminally on price fixing frequently. Dr. SCHIFRIN. I have a list-at one time, I did compile a list of such cases. They go back to 1040 and I would be glad to provide that in- formation at a later time. It is just a matter of finding the appropriate footnotes. PAGENO="0335" COMPETITIVE PROBLEMS IN THE DRUG INDUSTRY 1885 But I think- Mr. GROSSMAN. I-low do you define frequently? Dr. SOHIFRIN. I would define frequently in that-in this way: have there been found to be enough price-fixing agreements in the industry so that the reasonable observer would examine carefully licensing agreements for instances of price fixing? Have there been enough violations of the Sherman Act in order to warrant the assumption that there is certainly a definite possibility that price fixing accompanies cross licensing. Mr. GROSSMAN. You will submit that for the record at a later date? 1)r. SCHIFRIN. Yes.' Mr. GROSSMAN. One other question at this time: I think Mr. Squibb, when he testified, talked about the industry's desire to get into the teaching institutions and therefore to cut their prices to go into these hospitals, to make sure that the young doctors see their product or become familiar with them. Do you think that the consumer, who I think more frequently buys the "A" products there, is paying for this activity by the industry? In other words, he is paying a lot more because the industry is cut- ting these prices in the institutionalareas? Dr. SOHIFRIN. No. I do not believe that the prices to the consumers would be lower than they are if the drug firms did not engage in that activity; no. I believe that the consumer pricing is based, really, on- in fact, I have definite information on what consumer pricing is based on. I have correspondence from people in drug firms going back quite a number of years and this you mention was never determinant as a factor in setting prices. Mr. GROSSMAN. Do you think the drug companies make high profits on those sales to institutions at very low levels? Dr. SOHIFRIN. Yes; I do. You are familiar with marginal cost pricing. One reason for my belief that the consumer bears the expense of the research and devel- opment, bears the expense of the promotion, bears the expense of the large profit. The actual costs of manufacture of most drugs is very, very small. I am sure the price charged to virtually any person covers at least the direct cost of production. The firms are not losing money on those sales to hospitals. Mr. GROSSMAN. Thank you. Dr. SOHIFRIN. Careful scrutiny of patentability is a threat to drug firms because there is a sizeable chance that no patent may actually be deserved; without patent protection, many firms can manufacture or obtain the drug, produce preparations containing it, and sell it with subsequent price competition a possibility. In any case, the mar- * ket will be shared by more sellers than otherwise. (3) Even where patent protection has not been garnered, trade names accomplish a nominal differentiation largely accepted by phy- 1 Dr. Schifrin subsequently stated, "In addition to the meprobromate and t~ r; cy~line cases, other price fixing cases have been: U.S. v. Eli Lilly and Co., USDC (lie) 1941, (insulin) ; U.S. v. Schering Corp., at al., IJSDC (NJ) Civil Action No. 1919, 1)141 (hor- mone products) ; U.S. v. Alba Pharmaceutical Co., et al., USD0 (SDNY) 1941 (imports) U.S. v. Eli Lilly and Co., et al., USD0 (NJ) Cr. 173-58, 1959 (Salk polio vaccine)." PAGENO="0336" 1886 COMPETITIVE PROBLEMS IN THE DRUG INDUSTRY sicians, which effectively eliminates any possible competitive threat posed by the generically labeled output of the small firms. Trade names supplement patent monopoly where it exists and substitute for it where it does not. Thus the large drug flrms have u~ed two devices, patents and trade names, to eliminate virtually all tendencies toward price competition. They have led to a different sort of competition-one in which the consumer comes up a loser. Not forgetting the contribution to better health and longer life made by the industry, there has emerged in it a new competition-one that rewards molecule manipulation, ques- tionable patent tactics, excessive promotional claims, and oftentimes a product inadequately tested or cautioned. Such abuses are part of the industry's record, and have generated an increasing surveillance and regulation by the Food and Drug Administration. Inevitably we come to the matter of public policy in regard to this situation. Specifically, the question as I see it is how to im- prove the market performance of the industry while not impair- ing and hopefully even improving its product performance, as I have used those terms. Stated in perhaps a more meaningful way, the question is: How can public policy restore effective competition to the manufacture and sale of ethical drug preparations and thereby make their prices more reasonable, while preserving sufficient incentives for the discovery and development of new and better products? First, there is the matter of standards of drug patentability. Higher standards than those now prevailing are necessary to halt the routine issuance of patents whose validity is not substantiable in court. Higher standards of patentability will continue to reward true accomplish- ment and even induce more of it by affording it more protection than is now possible; patents for insignificant or substantial coattail developments or modifications would be eliminated. Such a change would greatly limit the financial gains available from molecule manip- ulation, but increase the gains from significant discovery, thus re- directing research and development funds away from imitative into innovational channels. This is the context in which I suggested this commission of experts to assist the Patent Office. I believe they could provide a good deal of influence on this higher standard of patentability for drug patents. Several years ago the Federal Trade Commission found that the tetracycline monopoly was built on patents obtained with "unclean hands and bad faith." Both the ability to acquire patents in such a manner and the economic motivation to do so must give way. My first specific recommendation, then, is that a special group, representing knowledgeable legal and medical expertise, serve as constultants to the Commissioner of Patents in reviewing and determining drug patent applications. My second proposal also deals with patents, but is further reaching in its impact. This recommendation focuses on the duration and scope of drug patents, especially pertinent in view of the monopolizing effect of such patents. It is my view that in duration as well as scope, drug patents provide excessive protection from competition, to the detriment of consumers. Accordingly, I offer two alternative plans for making PAGENO="0337" COMPETITIVE PROBLEMS IN THE DRTJG INDUSTRY 1887 drug patent protection relate more closely to the realities of this indus- try. These two plans are alternative policies for the compulsory licens- ing of drug patents. The first alternative is to require compulsory licensing of all drug patents after some specified period, such as 3 years. Three years of exclusive patent prOtection is a reasonable period in an industry characterized by rapid product turnover and a high rate of obsolescence. Numerous studies have shown that the greatest portion of sales of any product is likely to occur in the first few years after its introduction. Company price policies explicitly take into account these considerations, and most companies, if not all, estimate very con- servatively the anticipated market life of their proctucts, usually tak- ing 3 years as the period to recoup outlays and earn a profit. If licensing were required after the first 3 years of a product's life, that is, after its estimated life expectanc,y for pricing purposes were ended, there could occur entry by other firms into that product market and, hopefully, competition in price among the rivals. Of course, bene- ficial results to consumers would be possible only for those products with a therapeutic or commercial life longer than 3 years, `but despite the swift product turnover in this industry, data on product sales in- dicate that the majority of sales in any recent year represents those of products on the market longer than 3 years. For them, the patent holder would continue earning entrepreneurial profits, though perhaps at a lower rate than before, on his own fin- ished-product sales and those of licensees, and consumers possibly could now purchase their prescriptions at lower prices. `The impact of such a policy on research and development does not seem unfavorable; it provides a time long enough, from the companies own viewpoint, to earn profits justifying the innovational effort. It might even promote greater research and development by inducing even more rapid product turnover. In many cases, a realistic period of exclusive use and compulsory licensing at a fair royalty rate afterward seem unlikely to deter research and innovation. My second alternative patent-licensing policy focuses on scope, as contrasted with the first alternative and its focus on duration. This view of drug patents raises the question as to the justification for any period of exclusive patent use. The contention that none is necessary is based on the fact that a drug patent gives its owner a monopolistic position in either of two markets, that of bulk sales, that is, of drug substance itself-or that of dosage form products. To the extent that he takes his profits in the sale of bulk or in roy- alties from licenses for its manufacture, he must share the market for finished products; if he retains his monopoly in the latter-the fin- ished product market-he cannot reap profits from bulk sales or licens- ing, but of course, can earn substantial profits as the sole seller of finished products. Compulsory sales of the bulk drug or licensing of its manufacture merely specifies that the patent holder must reap his gains in the bulk market rather than the final-product market; it does not take away the opportunity to earn a profit jusifying the effort behind the dis- covery. And entry into the preparations market would occur at the 81-280-68-pt. 5-22 PAGENO="0338" 1888 COMPETITIVE PROBLEMS IN THE DRUG INDUSTRY time when the item is introduced, rather than 3 years later, and corn- petition can possibly come into play that much earlier. Perhaps in this plan the royalty might have to be higher than in the earlier approach, but not by so much as to prevent licensing from being an effective check against exorbitantly high bulk prices. Patentees would earn reasonable and adequate profits by charging attractively high bulk prices or by imposing the most favorable royalty rates permitted. But the cost of bulk ingredients is usually only a small fraction of the total cost of production, and high bulk costs or royalties would be more than offset by the economies in re- source use, particularly reductions in the vast and largely wasteful advertising effort, and the more reasonable profit margins that price competition would bring about. Compulsory licensing, because it permits rival firms to enter into the market, thus is a necessary condition if price competition is to be restored to the industry. But it is not a sufficient condition for that competition to arise, it opens the door to the entry of additional firms into markets closed by patents but it does not make them effective competitors of the dominant one or few. I come then to my last major recommendation-the prohibition of trade name designations. Trade names, as you well know, are those unique company names for its products-simple, catchy, and easily remembered-Syncillin, Achromycin, Tetracyn, Pen Vee, Miltown, Ledercillin, Orinase, and on and on and on throughout the catalogs of the large drug houses. Such names are totally unnecessary in every respect. If differentiation of drug products is necessary, and I am not fully convinced that it is, let it be done not through a proliferation of new names that are intended to displace generic terms for the product, but in the same way as differentiation is made in virtually every in- dustry, by the use of the manufacturer's name. Thus, the names "Car- ter: meprobromate" and ~WTyeth: meprobromate" tell us much more than do the words "Miltown" and "Equanil," while preserving com- pany differentiation. (These statements of mine parallel very closely the testimony of Dr. Garb on June 20.) The use of brand names that combine the company and generic comparability becomes clear and un~bscured, contrary to the purpose and effect of trade names. The elimination of trade names will go far in establishing the facts of generic similarity to physicians. Those doctors who want to select the speciality of a particular firm can continue to do so by using the brand name; but those who feel, as many do, that generic equivalents are therapeutic equivalents can thus prescribe by generic name alone, or by the brand name of a reputable seller whose product bears a com- petitive price tag. These three proposals of mine-a special drug patent board, some form of compulsory licensing, and the elimination of trade names- can, together, go very far in restoring opportunities for competitive entry into markets, in restoring price competition in the place of waste- ful and often harmful promotional competition, and in bringing about reasonable prices to consumers, while preserving the incentives for the research and development effort behind the industry's generally com- mendable product performance. PAGENO="0339" COMPETITIVE PROBLEMS IN THE DRUG INDUSTRY 1889 The opportunities for extraordinary profits still exist, but only for innovation or discovery of therapeutic merit-not through monopoli- zation of markets and exploitation of consumers who have no alterna- tive but to pay prevailing prices. All this seems to me to represent a considerably more acceptable situ- ation than now prevails in the ethical drug industry. Lest the tone of my proposals, or the spirit in which I offer them be misunderstood, I would like to offer a sort of epilog to my prepared statement. I do not view Government as something necessarily separate from the citizenry of this country, nor do I view it as a monolithic force seek- ing to create for itself an increasing power over the individual persons or business firms in our economy. I firmly believe that, as a Nation, we have maintained a prima facie case for individualism, for free enter- prise, and for the free market economy. But our economy, as other facets of our society, requires a system of checks and balances to prevent the generation of private economic power and its subsequent abuse. Our general commitment to competi- tion is based on the importance of the checks and balances that compe- tition imposes on the pursuit of economic self-interest, thus channeling the operation of the economy in a direction that serves the interests of society as a whole. But the absence of effective competition in most sectors of the drug industry must be recognized, and its implications in terms of that industry's operations have been the basis and the main concern of this committee's hearing. The large firms in this industry have acquired an economic power incompatible with our conception of a free market economy; oppor- tunity is severely limited for the many small firms in the industry, and the purchasers and consumers of ethical drug products are denied the benefits of competition in the manufacture and sale of those products. Even if we opposed only the abuse of such power, rather than its existence, the ethical drug industry has not restrained itself in any significant way. The profit record of the industry, unjustifiable by any of the accepted standards Of economic performance, attests to this lack of restraint. Wastes in promotion and other aspects of development, manufacture, and sale add to the costs imposed on society because of the absence of effective competition. Thus, in the absence of restraints on individual power through market competition or self-imposed restraints, it falls to Government to induce an improved total performance from this industry. This, as I see it, is a proper and necessary role for Government in the economic field, and a role not at all incompatible with our commitment to a free enterprise economy. I believe we have come a long way in our understanding of the factors contributing to both the positive and negative aspects of this industry's operation and performance-and that we can devise ap- propriate changes in the context of its operation that will greatly enhance its total performance, while preserving the incentives necessary for its existence and growth. At this point, Mr. Chairman, I would like to request that a recent article of mine, "The Ethical Drug Industry: The Case for Com- PAGENO="0340" 1890 CO~ETITWE PROBLEMS LN TI~ DRUG INDUSTRY pul~ory Liceneing," the Antitrust Bulletin, fall 1967, be included in the record of these hearings in the appropriate place, for its relevance to the proposals I have made. I thank you for honoring me with the opportunity to make these remarks to you and for your kind attention in my presentation of them to you. Senator ~ELSON. The article on the "Case for Compulsory Li- censing" will be printed in the record at this point. (The material referred to follows:) [From the Antitrust Bulletin, fall 1967] THE ETHICAL IJRTJG INDUSTRY: THE CASE FOR COMPULSORY PATENT LICENSING (By Leonard G. Schifrin~) Introduction On December 7, 1959, the Subcommittee on Antitrust and Monopoly of the Senate Committee on the Judiciary, then popularly known as the Kefauver Corn- mittee, shifted the focus of its investigation of administered prices in the Ameri- can economy to `the ethical drug industry.1 In the Spring of 1967, the monopoly subcommittee of the Senate Select Committee on Small Business began hearings on ethical drug prices, particularly on the often large price differential `between finished products sold under company-assigned brand or trade-names and finished products of the same generic designation sold under the chemical or generic name,2 an issue originally raised by the Kefauver Committee. In the seven-and-a- half years between the first and the most recent Senate `hearings, `at least nine other series of hearings dealing with this industry have been conducted, and at least three additional committee reports `or studies have been submitted to Con- gress. `Some legislation has resulted from this extensive examination, but the only action of real substance, the Drug Amendments of 1962, deals mainly with ques- tions of drug safety and perhaps owes its passage as much to the Thalidomide tragedies in Germany and other European countries as to the economic and medi- cal issues raised in these many hearings. Despite the lack of legislative accomplishments, the time and attention spent in scrutinizing this important industry have, for the most part, been productive- The industry grew to maturity without drawing attention to its practices and performance, perhaps because of its close relation to the medical industry which traditionally does not publicize its economic activities or perhaps because of its continuing high profitability. The investigations, however, have revealed flaws in' its `operation, specifically its wasteful use of resources in promotion and research, the dubious contribution of some of its output, and the uneconomic relationship b~tween the costs and prices of its products. In the years in which the industry has been so frequently studied, public concern regarding the health services and products available to consumer-patients has grown: Medicare has become part of our Social Security law; support for the construction of health facilities has- multiplied significantly. These activities are, of course, only part of our growing concern for ever more numerous facets of the quality of human life. To the extent *Head, Department of Economics, College of William and Mary. 1 It Is symbolic of the character of the ethical drug industry that the term "ethical drug" itself has become obsolete. Drugs, technically, are the active ingredients which go into dosage-form products or pharmaceutical preparations, rather than the products or prepa- rations themselves. Now, however, the large majority of all preparations are pre- fabricated, I.e., already In dosage form when sold to pharmacies and doctors. Hence, the- modern ethical drug industry Includes firms primarily engaged In the fabrication, finish- ing, or sale of drug products or preparations in finished dosage forms such as pills, capsules, tablets, etc. Although the industry would be more accurately described as the ethical-drug. products or preparations industry, common usage still retains its now-dated designation. 2 All ethical drugs have generic, I.e., common or chemical, names. In order to distinguish their items from rival products with the same generic designation, some firms (mainly large, prominent ones) employ trade-name or brand-name designations as well. A trade name Is an original, trademarked, name assigned by a firm to its own item, such as Lederle's- use of the trade name Achromycln and Pfizer's use of the trade name Tetracyn for the tetracycline capsules each produces. Brand names, which combine the product generic name and the name of the producer or seller, are used less often than trade names. Examples of brand names are "Cortisone: McKesson and Robbins" and "Armour Thyroid." PAGENO="0341" COMPETITIVE PROBLEMS IN THE DRUG INDUSTRY 1891 that Oongression'al attention has fostered and promoted this central goal of public policy, the examination of the ethical drug industry has been a worthwhile en- deavor on its part. To the economist concerned with the relationships among market structure9 industry behavior, and economic performance, important findings have `been added by these hearings and reports to his catalogue of knowledge. This knowledge serves him well, for it strengthens his understanding of the forces contributing to the desirable `and undesirable aspects of market performance. Conclusions al- ready suggested by other industry studies have gained additional substantiation; other tentative conclusions have been amended, qualified, or rejected in the face of new knowledge gained from these invnstigations. Most important, we have eome closer to an understanding of the determinants of market performance and the manifestations of their influence in this industry. This paper has as its frame of reference `this economic approach. Its goals are two-fold: first, to evaluate the performance of the ethical drug industry on the basis of the material presented in the various hearings and other sources, and, second, to suggest how substantial improvement in this performance can be achieved through amendment of the patent laws applying to drugs.3 The thesis presented is that compulsory patent-licensing is essential for the needed improve- ment in the "market" performance of the industry (where market performance refers to those aspects of the industry's activities that determine the cost and profit elements c'overed by the prices consumers pay for finished products). It will also be argued that such a patent policy will not impair, and may even improve, the industry's product performance (where product performance refers to those aspects of the industry's activities that contribute to better health through the flow of new products to consumers as a result of research and development expenditures). I. THE SALIENT STRUCTURAL AND BEHAVIOR FEATURES OF THE INDUSTRY `The ethical drug industry can trace its lineage back many decades and even centuries to the fields of chemistry, pharmacy, and medicine. Nonetheless, in its present form it is a young industry, arising out of our `break in World War I with Germany, on whose sophisticated and knowledgeable chemicals industry we were then dependent for our drugs. Building on the foundation afforded by wartime successes in drug synthesis and manufacture, medical science and drug therapy began a co-operative effort tha't soon led to the discovery of insulin in 1921 and its commercial manufacture the next year. There followed discoveries and development of sex hormones; vitamins, first from natural sources `and later by synthetic processes; barbiturates; germicides; intravenous anesthetics; im- proved forms of suiphanilamide; the commercial manufac'ture of penicillin; and then advances in the n-hole area of antibiotics. In the post World War II period, successes came more and more quickly: steroid hormones; tranquilizers, anti- depressants, and other mental drugs; oral antidiabetic drugs; polio and measles vaccines; oral contraceptives; and a host of other new types of drugs. The development of new products has been the main source of the industry's growth. The rapid expansion of the indust.ry's output, to a current domestic level of $3 billion per year at the manufacturers' level, is one measure of the increasing success that the industry has had in its research and development activities and the applicability of its discoveries to an, expanding range of illnesses and injuries. But of equal significance to its rate of growth of output are the characteristics which the industry has acquired as it has matured. A. The emphasis on "specialties" In the late 1940's and early 1950's the ethical drug industry faced serious problems. The discoveries of penicillin and streptomycin caught the attention of many firms. Penicillin, a so-called "product of nature," was unpatentable; the streptomycin patents were held by Rutgers University and freely licensed. As a result, markets for these products were easy to enter. The government encour- aged the expansion of facilities, and new production methods greatly increased yields. The combined result of these factors was a large overcapacity in the pro- It is imnortant to note that drugs, which, technically, are the active chemical substances In drug products, are patentable, as are the processes In drug and drug-preparation manu- facture. The finished products are notpatentable per se, although the brand or tradenames under which they are sold are copyrighted. PAGENO="0342" 1892 COMPETITIVE PROBLEMS IN THE DRUG INDUSTRY duction of products containing these drugs, drastic price competition in their sale, and, frequently, sizeable losses. The situation was chaotic from the producers' point of view. With many firms manufacturing items of generic equivalence and selling them under generic labels, the only form which competition could take was in price. The major firms soon realized that shifting the basis of competition to some form other than price and reducing the extent of competitive rivalry were the keys to increased profitability. Thus each of the large firms turned its attention to the production of "special- ties," i.e., differentiated and hopefully, exclusive items. The emphasis on special- ties had two manifestations: the increased reliance on obtaining patents covering new drug discoveries, and a shift to the use of trade-names for the items, both old and new, in company catalogues. Both of these pursuits have been impressively successful. Few of the many drugs discovered and developed since 1950 have not been protected by patents awarded to private firms; and reliance on trade-names in prescribing drug products has become the rule rather than exception, particu- larly in the dominant private-prescription market, in which private physicians prescribe medicines for patients, to be purchased at drug stores. As a consequence of the shift to specialties, the three dozen largest firms in the industry, which represent about five percent of the total number of firms but account for over 95 percent of all research activity and selling outlays, have come to dominate the industry. These few- firms account for virtually all the sig- nificant research advances coming from within the industry; have acquired pat- ents on the large majority of new- drugs; and have successfully induced physicians to make product choices in trade-name terms. The growth of patents and of trade- name use have effectively changed the nature of competition in most of the in~ dustry's markets from frequent and substantial price competition to competition in product development and promotion and advertising. Price competition is quite a rare phenomenon in all but a few limited instances. Presently, only in- stitutional purchasers-mainly hospitals and government agencies-employ ge- neric designations to any substantial extent. "Specialties" have accomplished their primary purpose. B. The industry's research effoft The high rates of new-product introduction and rapid obsolescence that charac- terize most drug markets give testimony to the size and effectiveness of the industry's research effort. Even in those cases u-here discoveries were made in universities (e.g., Salk vaccine) or government-sponsored research (e.g., penicil- lin), it usually has been the developmental work done by drug firms which made the product commercially available. This is not surprising. The drug industry is the most research-conscious of all non-defense industries, with more company- financed research and development in relation to sales than any other industry. In 1964, the all-industry totals for research and development expenditures as a percent of sales was 4.4 percent; for drugs and medicines, 4.7 percent. For company-financed research and development, moreover, the all-industry total was 1.9 percent, as compared to 4.5 percent for drugs and medicine.4 For the pe- riod 1956 to 1964, the average annual increase in research and development cx- penditures for the drugs and medicines industry w-as 13 percent, compared to an economy-wide increase of slightly less than 10 percent in total industry expendi- tures and between six and seven percent in total company-financed expenditures.5 Furthermore, while it is true that most research and development expendi- tures in the drug industry are for applied research and product development, the drug industry devotes a greater proportion of its research budget to basic re- search than does the economy as a whole or the industrial sector. For the drug industry, then, research and development expenditures have shown marked increases in recent years, compare quite favorably with expendi- tures in other industries in relation to sales. and are devoted to basic research. not just product development. The research and development record of the drug industry is commendable. National Science Foundation. Basic R~search. Applied Resenreh, and Development in Industry, 1964, WashIngton, D.C.: U.S. Government Printing Office, 1966, p. 62. There are several industries that have greater research and development expenditures relative to sales than the drug industry. These industries, however. are in scientific. military, and engineering fields and receive the bulk of their funds from the rovirnm~nt. ~ Ibid., p. 62. PAGENO="0343" COMPETITIVE PROBLEMS IN THE DRUG INDUSTRY 1893 C. Product introductwn One prominent result of the extensive research effort of the industry is the market impact of the large volume of new items introduced annually. In any given year, most drug sales are likely to be accounted for by items introduced in the preceding five or six years. However, many products have a longer commer- cial life, and any decline in their share of total sales may be due as much to the acceleration of total sales through the introduction of new products as it is to their actual displacement by newer items. Nonetheless, market success is not an inevitable result of research and discovery or of new product development, and no more than a short life is assured to those products that are well received initially. D. Promotion. and advertising Just as a large part of the industry's research effort reflects its efforts to develop patentable specialties, much of its outlay for promotion and advertising reflects its efforts to establish trade-names as a basis for prescription writing. In the ease of research, no doubt the conquering of illness and legitimate profit considerations are complementary factors. In the case of promotion and adver- tising, the important need for conveying technical information to physicians so that they may employ drug products with maximum effectiveness is a reason which complements profit considerations. A number of factors in medical practice have rendered physicians increasingly dependent on producers as their major: source of information. Among these factors are the proliferation of products containing new and older drugs; increas- ing specialization; increasing demands on doctors' time; the time-lag in the publication of journal articles; and the fact that most of the pre-introduction testing of new products and the continued testing of older products is done by or under the supervision of the drug firms themselves. All of these factors have contributed to the growth of the promotional element of selling costs. The responsiveness of physicians to the use of trade-names is both result and cause of a large-scale advertising effort.: As a consequence, advertising outlays have grown at a pace at least equal to promotion. Expenditures for advertising and promotion now account for roughly 25 percent of the sales dollar and one-third of total costs of production of large firms. Selling outlay ranks second only to the rather inclusive category "cost-of-goods-sold" as a cost component and is about four times greater than the research effort of which the industry is so proud. The advertising element alone is approximately twice that of research and development for the typical large firm. Few other industries compare closely with this in advertising effort relative to sales. E. Market concentration Of the 700 firms in the industry, the twenty largest account for more than 90 percent of total sales, and another dozen or 15 account for half of the remainder. The high costs of research and developiñent and of effective promotion of trade- names have set the large firms off frOm their many smaller rivals. The new products come almost entirely from these few large firms; in the markets for older products, the popularization of trade names has rendered the firms selling generically-designated items ineffective as competitors. Only in the production of products containing freely-available bulk drug ingredients (which are dimin- ishing in relative economic importance) and in sales to those institutional buyers who purchase by generic designation can the many small firms participate. Iii all other market situations, the large firms dominate the picture. This uneven division of shares of total industry sales between large and small firms is but one dimension, of concentration. Concentration is an even more mean- ingful concept in individual product areas than it is for the industry as a whole,. for it is this aspect of structure that conditions price policies and the nature and extent of market competition. Among the large firms there has emerged a pattern of specialization that tends to break them into smaller, rather exclusive groups, each group sharing a product area such as antibiotics, or steroid hormones with little fear of entry even by other large firms. ° Promotion and advertising are interrelated activities, but separable in concept, func- tion, nnd perhaps also In magnitude. Promotion, in essence, Is the conveying of technical information about drug products that makes possible their use in therapy. Advertising is mainly directed at establishing and reminding physicians of the trade-names of company specialties. Estimates within the industry indicate that total selling outlays are divided roughly evenly between the two categories. PAGENO="0344" 1894 COMPETITIVE PROBLEMS IN THE DRUG INDUSTRY Thus, as patented specialties occupied a growing prominence in drug markets, and as the high costs and cumulative nature of research effort induced specializa- tion by drug firms, drug markets increasingly became oligopolized by the few large firms specializing in each of the different areas of research and production. While the exact order of firms in a market may change, the positions of leader- ship are effectively preserved for the large firms specializing in that area. Con- centration thus tends to be both high in degree and stability. Even in the case of those older (and a small number of newer) products not protected by exclusive patents, large firms, with their trade-named specialties and their successful pro- motion campaigns, have come to dominate sales. As an example, in antibiotics-the largest of the various ethical drug classes- there are but ten prominent sellers and even they are unevenly participative within the area. Only six of them are important manufacturers of broad-spectrum antibiotics, and only five (including three of the broad-spectrum oligopolists) engage in the manufacture of the leading medium-spectrum products. There are many firmS producing products w-hich contain the older, unpatented penicillins, but an estimated 70 percent of all such sales are accounted for by Lilly, Wyeth, Abbott, and Squibb, four of the ten major firms. Comparable concentration is to be found for other aiitibiotics, including streptomycin and dihydrostrepto- myciri, penicillin-dihydrostreptomycin combinations, and penicillin-sulfa combi- nations. A similar pattern of high concentration can be found in the manufacture of hormone-drug products, mainly the cortical steroids, where seven firms' domi- nate the market. The same high degree of concentration is to be found in virtually every other product area in the `industry. In the most recent study of market concentration in the drug industry, the four largest firms in each of 13 major product areas accounted for between 60 and 80 percent of sales, sometimes substantially more and rarely less. The main characteristics of the ethical drug industry-its emphasis on spe- cialties, its large research and selling effort, the growing patent protection afforded its developments-have changed the structure and nature of rivalry very much since the penicillin era. But high concentration and a changing nature of competition are one thing, and undesirable market performance and the need for corrective policy changes are another. Judgments based only on industry structure and the over-all nature of rivalry may be unwise; thus, it is to an examination of the various dimensions of the performance that is conditioned by these factors that we must now turn. II. INDUSTRY PERFORMANCE There are many dimensions and facets to the performance of industries and the markets in w-hich they function. Perhaps it is best to consider those of the ethical drug industry and its markets as being of two main types. These group- ings, as noted earlier, may be called product and market performance criteria, respectively. A. Prodvct performance On the product side, the main concerns are the efforts to develop new and better drugs, the increasing availability of these drugs, and their impact on life and health. The large outlays by the industry for research and development have already been discussed. In both absolute and relative terms the industry is justly proud of its emphasis on basic and applied research. While certain qualifications con- cerning both the magnitude and quality of this effort may be in order, on the whole the industry's research input is impressive. The results of this activity have been substantial. In the sixteen year period 1948-03, the total new products introduced amounted to 5,386.~ Most of these new products were duplicate items of products already on the market, new dosage forms of previously known drugs, or new compounds. 618, however, contained chemical entities not previously known (as did an indeterminate number of the new compounds) .~ All of these, including the duplicate single products, may ~ Study of Drug Purchase Problems and Policies, U.S. Department of Health, Education, and Welfare; Washington, D.C., U.S. Government Printing Office, 1966, p. 11. 8 Ibid. PAGENO="0345" COMPETITIVE PROBLEMS IN THE DRUG INDUSTRY 1895 represent contributions to improved therapy. All increase the range of choice for physicians; many represent new alternatives and some represent break- through discoveries into areas where no effective products had been available previously. The benefits of the new drugs cannot easily be measured, since most drugs provide symptomatic relief rather than cures. While the reduced hospitalization and shortened durations of incapacity may be quantifiable, within limits, the great reductions in suffering can only be recognized in general. We do know the preventive and curative drugs have greatly reduced the incidence of certain illnesses. Influenza, tuberculosis, pneumonia, and syphilis no longer are the severe killers they once were. The childhoOd diseases of measles, meningococcal infections, whooping cough, gastritis, duodenitics, colitis, and enteritis have been brought under control. Poliomyelitis is no longer the great crippler it had been. Steroid hormones have greatly reduced the pain and crippling effects of in- flammatory diseases, particularly arthritis. Tranquilizers and other drugs have done mu~h to reduce the seriousness and the hopelessness of mental illness. Indeed success over illness has become an expectation by the public, and this expectation has been considerably met by the industry. These contributions by the drug firms in the form of new products warrant the industry's pride in its accomplishments and its recognition by society. Added to this record is the rapid expansion in the industry's productive capacity and output, which has made both new and older products available in increasing quantity. Yet, while the achievements in research and development effort, new product introduction, and expansion in output have been impressive, certain criticisms of both the magnitude and quality of these activities must be noted. It is apparent that the industry exaggerates its research and development effort, perhaps to convey an impression of extreme risk, or competitiveness, or exceptional enterprise, with the purpose of justifying high profits. The industry, when speaking of its research and development activities to stockholders or to the public, defines them very broadly, being far more inclusive than the Internal Revenue Service or National Science Foundation. For example, industry spokes- men appearing before the Kefauver Committee put the 1959 research and de- velopment outlay at $198 million; the NSF figure for that year was $154 mil- lion, or 22 percent lower.9 Not only does the industry over-estimate the quantity of this activity, but many critics, including a large number within the medical profession, have questioned its nature and direction. There are serious allegations that much research activity is not related to product improvement but is imitative in nature, so (as to generate specialties that are not really needed, or is directed toward the acquisition of patent protection. Furthermore, it is contended that the great profit potential awaiting new products induces their introduction before there is sufficient knowledge of their limitations and dangers. These allegations have been at least partially substantiated. It has been shown above that duplicates, new dosage forms, and mixtures represent the large ma- jority of the products being introduced, and the therapeutic advances they repre- sent may not be very great. Many of the new single chemicals, it has been claimed, represent no significant progress in drug therapy. It would seem, then, that only a few "new products" represent real progress. Further, if research outlays in certain ways are excessive, in others they may be inadequate. When the Food and Drug Administration raised the standards for testing of new drugs as a response to the Thalidomide experience, the number of new drug applications and of approved new products declined sharply. Certainly the behavioral patterns and motives that are alleged to lie behind these criticisms of the industry are compatible with the structural and other features of the industry noted earlier. These are important criticisms of the industry's product performance and must be taken seriously. It thus seems likely that the industry's product performance is not as great as claimed, nor is it as great as it might be with a re-direction of its research and development effort. Despite these qualifications, the available data support the conclusion that the industry has contributed greatly to people's lives and indirectly to the economy. For these contributions the industry deserves great credit, whatever the flaws that exist in its structure and operation. g Research and Development in American Industry, 1962, Washington, D.C.: U.S. Gov- ernment Printing Office, 1063, p. 9. PAGENO="0346" 1896 COMPETITIVE PROBLEMS IN THE DRUG INDUSTRY B. 1[arket performance The other side of performance is what we have called market performance. Though it is related in varying degrees and manner to product performance, such a distinction is a useful one for this study. Market performance concerns pri- manly the efficiency with which the industry uses the economic resources avail- able to it In the development, production, and marketing of its output, and the relationship between prices charged and costs incurred in making goods avail- able to consumers. The criteria of good market performance are quite traditional in economic analysis :-that (1) the costs incurred by sellers be closely related to activities from which consumers derive benefit, be they expenditures needed to develop new and better products, direct and indirect cost of manufacture, or marketing outlays that lead to better product selection or expansions in demand that justify larger-scale, but lower-cost methods of production; and that (2) the prices consumers pay bear a reasonable relation to these costs; that is, prices provide profit margins that are sufficient to reward enterprising firms for ac- tivities that serve consumers, but not so large as to represent gains from behavior not in the consumers' or society's interest. These criteria are now used to evaluate the drug industry's performance. First, in regard to the cost elements it has been indicated above that research and de- velopment and selling outlays are large, although quite unequal. Further, criti- cism that the research outlay is, at least in part, wasteful has been noted. While the extent of such waste from misdirection of outlay may not be closely determin- able, there appears to be logical support for the contention that the goal of ob- taming patentable specialties may conflict with that of achieving maximum thera- peutic gain from research effort, particularly since the commercial rewards for `new" products are sizeable. Protection of an oligopolistic position by imposing patent barriers around existing key products and processes also seems to influence the direction of research effort to some extent. Again, the question as to fre- quency and extent of wasteful research effort is incapable of precise answer; yet the large financial rewards awaiting new products, coupled w-ith the large pro- motional outlays preceding, accompanying, and following their appearance, means that significant therapeutic progress need not be a necessary condition for large profitability. Second, as noted above, selling expenses in the industry are about four times as large a cost factor as is R & D. The promotion part of selling and the advertis- ing element are roughly eq)ial and each is twice the R & P component. Promotion and advertising, both of which may be important marketing functions, nonethe- less include very costly outlays for activities that may provide the consumer no benefit, directly or indirectly, or may actually do him a disservice. There is considerable opinion that promotional excesses are closely related to the appear- ance of new products that are of limited merit but which might become profitable through large-scale promotions. The costs of promotion are, perhaps to a large extent, necessary under present institutional arrangements. Physicians have, for reasons already noted, be- come dependent on drug firms for their information on new products. Thus much promotional effort, parti~ularly that providing information on new and old prod- ucts. on the incidence and nature of side effects, and on other technical matters, is necessary. But advertising, whose function is to popularize and remind physi- cians of trade-named specialties, serves merely to raise the costs that ultimately become part of prices while not improving the drug selection process. Also, the quality of information conveyed has been questioned, largely by the physicians at whom it is directed. Incomplete information, excessive claims, and non-report- ing of side-effects are frequent among their criticisms. Thus, advertising may even be detrimental to sound drug therapy. That there is excessive, wasteful, and mis-directed expenditure for research and promotion in the industry, and that such waste is of sizeable proportions has been reported numerous times in the hearings involving the industry. Prices con- ditioned in large part by these costs are, therefore, necessarily higher than they otherwise would be. Drug prices are uneconomically high for another reason. They provide drug firms with profits that are consistently well above those of most other industries. There is little justification of the profit levels that prevail. With rapid turn- ~over of products and successful research results uncertain, the industry has been described as one of high risk, in which fortunate firms can be expected to earn PAGENO="0347" COMPETITIVE PROBLEMS IN THE DRUG INDUSTRY 1897 substantial returns. Yet the evidence that high risk exists would require low profits for the unfortunate firms, and suchevidence is not apparent. Table 1 shows data for profits in the drug or pharmaceutical `° com- pared to broader aggregative groupings for the past decade. The data show quite clearly the contithiing and substantial above-average profitability of drug manu- facture and sale. TABLE 1.-ANNUAL RATES OF RETURN AFTER TAXES ON STOCKHOLDERS' EQUITY, DRUG INDUSTRY AND ALL MAN- UFACTURING CORPORATIONS (EXCEPT NEWSPAPERS), 1957-61; PHARMACEUTICALS AND ALL INDUSTRY, 1962-66 (In percenti ~: ~ Drug industry All manufacturing corporations (except newspapers) 1957 18.6 10.9 1958 17.7 8.6 1959 `17. 8 10.4 1960 16.8 9.2 1961 16.7 8.8 Pharmaceuticals All industry 1962 14.4 8.9 1963 1964 1965 14.7 16.3 18.0 9.1 10.5 11.8 1966 20.3 (1) 1 Not available. Sources: 1957-61 Federal Trade Commission-Securities Exchange Commission Quarterly Financial Report; 1962-65 `Fortune Annual Directory of 500 Largest Industrial Corporations; 1966, Federal Trade Commission, cf. New York Times, June 4. 1967, p. 14F. Within the drug industry, company profits for the recent past snot only nre high on the whole, but fairly stable. For example, in the decade 1952-61, for 18 of the `largest drug firms for which data have `been obtained, the average after-tax rate of return to net worth was 18.5 percent: Only four firms earned below 10 percent, a reasonable estimate of the all-industry average, as based on Table 1, at any tiiaie in this period. Of these four, Mead-Johnson fared below 10 percent (9.3 per- cent) only in 1952; Merck had annual profits of 9.6, 8.9, and 9.6 percent in the first three years and averaged 15 percent in ,the last seven; Bristol-Myers had annual profits of 7.1, 6.8, and 9.3 percent in each of the first three years and averaged 16 percent in the last seven. Olin-Mathieson is the fourth. It entered the industry `with its acquisition of Squibb in 1955, and, in the seven years within this period in which it participated in drug manufacture, had profits below 10 percent in four. But Olin-Mathieson is a large conglomerate firm, and its drug operations are a small part of its total activities. Merck,'Mead-Johnson, and Bristol-Myers are also conglomerates, though in lesser degree, owing to their sizeable activities in bulk chemicals, dietary products, and proprietary drugs, respectively, and in each case the drug-operations of these conglomerates `have been substantially more profitable than other operations.11 These and other data show that profits in the drug industry as a whole, and for the large majority of its leading firms, rank substantially and consistently above those of industry in general. Examples of unstable or occasionally `below- average company profits are difficult,' to find. Thus evidence in support of the existence of a high risk element facing the large firm is lacking. Another defense of the industry's high profits is that these profits are needed to finance the industry's large research effort. But, as we have seen, the research in The data for the "drug industry" and "pharmaceuticals" Include both ethical and proprietary (non-prescription) products. Since ethicals account for 70 percent or more of total pharmaceutical sales, however, the Inclusion of proprietaries does not seriously distort the accuracy of the data. 15 Subcommittee on Antitrust and Monopoly, Senate Committee on the Judiciary, Admin- istered Prices, Drugs, Report, Washington, D.C.: U.S. Government Printing Office, 1961, p. tLI. PAGENO="0348" 1898 COMPETITIVE PROBLEMS IN THE DRUG INDUSTRY effort is not as large as is often suggested. Nor is the argument a logically valid one. Profits represent the difference between revenues and full costs, including research. To add to costs an additional element for future research (which, of course, will also be paid for by future sales) is to charge consumers twice for the research component. Another point suggested in regard to profits is that the profit record of the industry may reflect the continuing success in new product development by all major firms. But many of these new products, including a few that have had the greatest impact on profits, are produced under license and do not reflect the firm's own research. American Home Products and Smith Kline and French consistently have been among the most profitable firms not only in the drug industry but in all industry as a whole. For each, one or two ethical items dominates the product mix. American Home Products' (Wyeth) leading ethical specialty is Equanil, manufactured and sold on license from Carter. SKF's leading specialties are Thorazine and Compazine, made and sold on license from Rhone-Poulenc of France. Most major firms sell a number of specialty items containing bulk drugs purchased from patent holders or manufactured on license from them. Furthermore, superior profitability is not limited to new product success. A number of firms, e.g., Abbott, Parke, Davis, and Richardson-Merrell, have contributed few significant developments in recent years, yet their profit rates have consistently been well above that of most firms in the economy. It is noteworthy that Parke, Davis' profit rate was substantially above the all- industry average in those years when its one major development and leading sales item, Chioromycetin, was temporarily but almost virtually off the market because of the appearance of drastic side effects. Certainly an alternative thesis that is equally plausible is that high profits are due to meager competition in most of the industry's markets, including the entire private-prescription market and most product areas of the institutional market. The w-eakiiess of competition is largely attributable to the fewness of competitors which results from the patents on most new discoveries, from limited licensing, and from financial and technical obstacles to entry. Patents and limited licensing are, of course, in accord with public policy. And many of the criticisms regarding wasteful cost elements, excessive and inaccurate promotion, and unjustified profitability have been directed toward many industries. But the operation of the ethical drug markets warrants special concern and special public policy measures for a number of reasons. First, there is the fact that health is at stake. New products which confuse drug selection, and "information" and claims that mislead, may impair drug therapy and harm the public. Second, the peculiarly exploitable nature of the patient-consumer, which is conditioned by his need for drug products and the fact that physicians rather than consumers make product selection, renders the reasonableness of drug prices particularly important. The greatest harm befalls those who are unable, because of prices higher than necessary, to afford all the medication they need. III. PUBLIC POLICY The essential question is how to reduce the wasteful outlays by firms and the excessive profit margins and pass these reductions to consumers without impair- ing new product development. The answer lies in inducing those changes in the structure and operation of drug markets that will generate price competition, while retaining the availability of rewards sufficient to promote the discovery and development of new and better products. A number of co-ordinate policy changes are necessary. First, free entry into finished-product markets must be encouraged for en- trants who can become effective competitors of the established oligopolists. A substantial increase in sellers is likely to lead to more vigorous competition, particularly in price, and the elimination of wasteful and uneconomic prac- tices. For entrants to become effective competitors, the barriers posed by patents on drug ingredients must be lowered and the use of trade-names in prescription writing must be prohibited. Stricter quality controls by the FDA and its require- ment of more prominent mention of generic names in labelling and advertising have perhaps already laid the groundwork for an increased use of generic ter- minology in prescription-writing; recent publicity regarding the large differen- tials usually found between generically labelled items and trade-named special- PAGENO="0349" COMPETITIVE PROBLEMS IN, THE DRUG INDUSTRY 1899 ties hopefully will spur further efforts in that direction. Better and more accessible information from professional rather than industry sources may reduce the impact of advertising on physicians, thus reducing the magnitude of its use. Higher FDA standards for new drugs will reduce the appearance of new items of little need, thus eliminating mis-directed research and development expenditures and the promotional outiays and abuses that sometimes accom- pany them. These and other changes need to be effected for an improved industry performance. The net result of these changes can be the selection of items by physicians on the basis of price rather than company identification. The pertinent question remaining is how to reduce the patent barriers to entry without dangerously weakening the incentives for research and develop- ment. The economic philosophy behind Our patent system is that monopoly grants of a temporary nature and of limited scope, while perhaps creating imperfections that temporarily mitigate price competition, serve as induce- ments for rivalry in research and development which provide long run benefits to consumers that more than offset the temporary mitigation of market price competition. Antitrust decisions have defined the limits of patent protection by con- demning tying restrictions and other devices which owners of patents valid in one market have used to reduce competition in other markets. Flagrant abuses of patent rights, such as practiced by the United Shoe Machinery Company, have led to imposition of compulsory licensing on the offending firm. While compulsory licensing has been imposed to date only as a punitive measure on flagrant violators of the antitrust laws, the special importance of the drug industry to society's well-being and the critical flaws in its structure and behavior warrant the adoption of special drug-patent policies that include compulsory licensing as a general condition. This policy is a necessary condi- tion for improved industry performance in terms of costs, profits, and also in the critical matter of prices. Together with other policy changes already adopted or proposed, that induce physicians to prescribe generically, compisory licensing may also be a sufficient condition for improving substantially the market performance of the industry. `It is doubtful that the continuation and advancement of drug research would be impaired by such modification of drug patents. The industry is characterized by rapid product turnover and obsolescence; studies have shown that the greatest portion of sales of any product is likely to occur in the first few years after its introduction. Company price policies explicitly include these considerations, and most, if not all companies, estimate quite conservatively the market life of their products, taking three years as the average period to recoup outlays and earn a profit. If licensing were required after the first three years of a product's life, the period taken as the estimated life expectancy, entry in'to the market and the price competition it would create could contribute to lower prices and profit- ability after the patent holder has earned a profit justifying his innovation. Of course, the beneficial results would be confined to those products with a com- mercial life greater than three years. Although product turnover is swift, data on product sales indicate that the majority of sales in any recent year represents those of products that had been on the market longer than three years. For them, the patent holder would continue earning profits, though at a lower rate, on his own finished-product sales and those of licensees, and consumers would be able to buy at lower prices. In regard to the impact of the proposed policy on research and development activity, it might generate even greater effort by inducing swifter turnover. In any case, a realistic period of exclusive patent use and a fair royalty rate afterward seem unlikely to deter research and innovation. Another view of the role of drug patents raises a question as to the need for any exclusive use period. The contention that none is necessary is based on the fact that a drug patent gives its owner a monopolistic position in either of two markets, that of bulk sales or that of dosage-form products. To the extent that he sells a drug in bulk or licenses is manufacture, he shares the market for finished products; if he retains his monopoly in the latter, he cannot reap profits from bulk sales or licensing. Compelling bulk sales or licensing would require the patent holder to rely on the lulk market for his innovational profits, and would promote price-competition in the finished product market at the time when the item is introduced, `rather than after a few years as in the above scheme. In this plan, the royalty rate would have to be higher than in the previous' scheme, PAGENO="0350" 1900 COMPETITIVE PROBLEMS IN THE DRUG INDUSTRY but not by so much as to prevent licensing from being an effective check against exorbitantly high bulk prices. Patentees would earn sufficient profits by charging high bulk prices or imposing the most favorable royalty rates permitted. But the cost of bulk ingredients is only a small fraction in total cost of production, and high bulk costs or royalties would be more than offset by the economies in resource use and by the smaller profit and selling elements in drug prices that price competition would bring about. Either of these proposals would lead to lower prices for consumers while main- taining the profit incentive for drug discovery and development. Of `the two, perhaps the first is more feasible politically, but that is another consideration,. for examination at another time. Senator NELSON. I want to thank you very much for a most thought- fill and valuable contribution to these hearings. We appreciate this very much. Do you have any questions, Mr. Gordon? Mr. GORDON. I have just one. The drug industry very frequently states that high profits are neces- sary to finance its large research expenditures. What do you think of that argument? Dr. SCIIIFRIN. Well, I think it is entirely fallacious, deliberately so. Profits, of course, are the residual between revenues and costs. Thus you have profits only when all your costs are met. Research is part of your cost. Thus, the high profit exists after the research outlay has already been accounted for. A tw~ist of this is to say tht high profits are necessary to finance future research. But, of course, in future prices, there is an element that covers the research going on. To justify a high price because it is necessary for future research is in fact to charge the consumer twice for the research. He is paying for the present research and future re- search. The future consumer will also pay for present and future research, and on and on. It is an argument that does not justify the consumer paying double in the cost. Mr. GROSSMAN. We talked about proliferation of products before. You stated-that at least that brought in competition. But I think the point you discused with Senator Nelson before was very important and I wonder if your solutions really cover it. That is the failure of the small firms really to be able to compete due to the promotional problem. Dr. SCHIFRIN. Sir, do you want me to comment on how my pro- posals- Mr. GRoss~rAN. Yes; can we ever meet this? Dr. SCHIFRIN. Yes. You see what keeps the small firm from being effective in the market now are two barriers. The patent may keep him out of the market. If there is no patent barrier, the emphasis on trade names keeps him out because he can't promote his trade name. As I say, compulsory licensing is necessary, but not sufficient. They cannot be an effective competitor as long as prescribing is done by trade names. Thus I think trade names have no basis for existence. If brand names are used, they will clarify the matter as to generic equivalency. Once that is established, the small firm with a generic product will have a better chance of getting its items selected, cer- tainly once its low prices become recognized. PAGENO="0351" COMPETITIVE PROBLEMS IN IHE DRUG INDUSTRY 1901 Mr. GROSSMAN. Do you think it will also help if the larger firms will cut down on its advertising? Dr. SCHIFRIN. Definitely, because you will not be advertising a trade name trying to make it seem something different from other' products with generic equivalency. Mr. GROSSMAN. So a lot of wasted effort is cut down? Dr. SCrnFRIN. Right. If you could eliminate the waste, it would reflect on the consumers in two ways: They would either pay less for what they get or get more for what they pay. Senator NELSON. I thank you very much. We will take a 5-minute recess and then hear Dr. Steele. (Recess.) Senator NELSON. Our next witness is Dr. Henry Steele, professor of economicS at the University of Houston, Houston, Tex., and an economist with very distinguished credentials. We appreciate very much your coming to testify today, Dr. Steele.. You have presented to the committee a. very detailed and well-prepared analysis of testimony given by the eëonomists represent.ing the Phar- maceutical Manufacturers Association before this committee as well as additional material. You may present this material in anyway you see fit. All of your statement, as well as your crit.ique of the presenta- tions of the economists for the PMA will be printed in full in the record. You may present your material in any way you see fit. STATEMENT OF HENRY B. STEELE, PH. B., ASSOCIATE PROFESSOR OF ECONOMICS, UNIVERSITY OF HOUSTON, HOUSTON, TEX. T)r. STEELE. Thank you, Senator Nelson. I would like to read the first half dozen pages or so from my major statement and then turn to the supplemental statement commenting on the PMA presentations, if I may. I greatly appreciate the privilege of being invited to make this state- mont before this subcommittee, and it is my hope that t.he information which I am able to present will be of some use t.o you in your delibera- tions regarding the vitally important economic problems arising out of the market context within which the drug industry operates. I am an academic economist with major research interests in indus- trial organization and the regulation of industry, and have done much work in the area of medical economics a.nd drug industry regulation. I received my Ph. D. degree in industrial economics from MIT in 1957, and since then have been engaged in teaching and research, as well as in consulting for private firms, U.S. Government agencies, and foreign governments. At present I am an associate professor of eco- nomic at the University of Houston. My research in the drug industry has continued over the last 7 years, and I have written three articles on drug industry economics and regulation and coauthored two arti- cles on the supply and distribution of physicians' services, all of which have appeared in professional economics journals. In March 1965, I presented a liaper on drug industry regulation before the University of Illinois Medical School at the invitation of PAGENO="0352" 1902 COMPETITIVE PROBLEMS IN THE DRUG INDUSTRY Dr. Harry Dowling, the chairman of the ck~uncil of Drugs of the American Medical Association; and in February 1967, I presented a comprehensive program for the reform of Canadian drug laws and regulations before the Special Committee on Drug Costs and Prices of the Canadian House of Commons. Senator NELSON. May I interrupt just a moment? Do you have with you the three articles on drug industry economics and regulation and the two articles on the supply and distribution of physicians' services? Dr. S~riiiIAE. I have the three articles on drug industry regulation with me. Senator NELSON. Would you submit them for the record? Dr. STEELE. Yes; and I can submit the other two articles as well. Senator NELSON. Would you also submit the other two articles, and they will be printed at the conclusion of your testimony.1 Dr. S~u~I~E. Thaiik you. I continue. In making the Canadian presentation, I represented the government of the Province of Alberta, and of the 14 recommendations which I submitted, 11 were incorporated in the final report of the committee to the Canadian House of Commons. All of my research on the drug industry in the United States has been entirely self-financed, and in presenting this statement to the Subcommittee on Monopoly I wish to make it clear that I represent no one but myself. I. INTRODUCTION What can an economist contribute to hearings on drug industry problems? It is curious that in all the hearings held in the United States, as well as those in England and in Canada, the original demand for the hearings has come about because of the conviction that prices are "too high," but very much of the hearings have been occupied by investi- gations into the safety and efficacy of drugs, and medical and phar- macological considerations have quite generally been predominant over economic issues. Yet, for every person wtho is moved to voice a complaint over poor drug quality, there must be a hundred who complain about high prices. But the emphasis on drug safety and efficacy inevitably redounds to the advantage of the major firms, who would much rather fight on the battleground of relative quality than of relative prices. Their argu- ment `has two parts: First, drug prices are related to costs, particularly quality control costs. Second, drug quality and hence safety and efficacy is related to these same costs. Hence it is asserted that one caunot divorce questions of cost and price from questions of safety and efficacy; therefore, since problems of drug quality, being matters of life or death, are obviously more crucial than drug price prdblems of mere dollars and cents, steps to assure high quality should take precedence over economic reforms. Economic reforms are correspondingly delayed. (This was the net effect of the otherwise admirable Kefauver-Harris Act `of 1962.) 1 The five articles submitted by Dr. Steele begin at p. 1950, Infra. PAGENO="0353" COMPETITIVE PROBLEMS IN THE DRUG INDUSTRY 1903 This argument is faulty in several respects: First, drug prices are not related to drug costs, but instead to de- mand and ability to pay. Second, while drug quality obviously depends upon care exercised in manufanture, the cost of quality control has been shown to be a very small part of the total costs, and the difference in cost between a mini- mal satisfactory program and a deluxe program would not begin to account for the difference in prices `between the generic drug and its brand name equivalent. But to `return to the question: What can an economist contribute to drug law reform hearings? If the data were made available, he could analyze cost-price condi- tions within the individual drug firms, and the pattern of mterfirm price and product competition, and arrive at an informed judgment regarding the status of competition in the industry. But such data have not been made available, even to economists retained to defend the industry.' In view of the absence of the data in the analysis of which the econ- omist has a comparative advantage, what constructive role can he play? Primarily that of coordinating and synthesizing the economic aspects of the data which is in the record, and evaluating the economic relevance or credibility of certain of the arguments advanced by the drug interests. It is noteworthy that drug spokesmen produce arguments in their defense which either stress or ignore similarities or differences between drugs and other industries to suit their convenience. Thus, in the PMA studies presented last month, one study treated the drug industry just like any other industry in relating the variance of the earnings of member firms in an industry (rather arbitrarily called "riskiness") to the average rate of earnings in that industry, while the other anaJyzed product competition in drugs in a vacuum as it were, without introduc- ing comparative data from any other industry. But both the similarities and the differences of the drug industry should be analyzed and allowed for before making any comparative study of drug prices, costs, and profits in relation to those of other industries. 1 As Professor Markham stated before this Subcommittee on December 19, 1967, in response to just such a question, "you are just not going to get those data, and I do not think-I would be less than honest if I said I would try to get them, implying that I could get them for you." (Transcript, volume 23, p. 2805.) Markham apparently referred not only to the confidential status given the information, but also questioned whether or not drug firms bothered to make all the cost allocations Involved. Although it Is to be admitted that many of the calculations can be made only on the basis of arbitrary assump- tions, one would expect that well-managed firms would find it prudent to undertake such analyses for their own information. In fact, Dr. M. A. Phillips, in his Sainsbury Committee memorandum to the British Ministry of Health stated that the drug Industry was no different from other organic chemicals industries in observing the customary precautions of making detailed cost studies prior to engaging in producing projects. These studies include the costs of research and development and of promotion. Dr. Phillips' statement Is unusually authoritative in that he is a drug industry conslutant who has made many economic evaluation studies for drug firms. Phillips complains that "It has been found very difficult to obtain figures for the cost of research and development and of promotion and advertising, although this must be known to those who have to spend this money in these ways . . ." and explains that even with the approximations his organization has to use in estimating these costs, he is satisfied that the accuracy of the estimates for these items Is within 25 percent. See Competitive Problems in the Drug Industry, Part I, pp. 54-55, of the Hearings before this Subcommittee on the present matter. (It might be observed that only if there is a very large gap between cash flow and expenditures is a company actually likely to indulge in som4 carelessness or negligence in the relating of total costs to individual items sold.) 81-280--68-pt. 5-23 PAGENO="0354" 1904 COMPETITIVE PROBLEMS IN THE DRTJG INDUSTRY The major similarity between drugs and other industries is that the firms are privately owned and are managed in the interests of maximiz- ing profits and the value of the shareholder's investment. Provided that there is effective competition among firms, the profit- maximization goal is an excellent coordinating mechanism since it motivates managements to produce a. given output at minimum total cost and hence stimulates efficiency. At the same time, price competition among firms will keep price levels from exceeding for any great period of time the equilibrium levels necessary to elicit from producers that level of production which consumers demand at that price. This is basic to the classical economic doctrine of Adam Smith (who was mentioned more than once with approval by PMA witnesses last month) that the force of price corn- petition taking effect through the market, acts as the famous "invisible hand" which by allocating resources efficiently makes each economic agent serve the general welfare even though he is only interested in furthering his own private fortunes. But unless effective competition prevails, private and public welfare in the market are not consonant. In the drug industry, the invisible hand is invisible chiefly because it is so deeply buried in the consumer's pocket. And it is the extreme vulnerability of the drug buyer to eco- nomic exploitation which makes the drug industry (both in economic and public policy terms) a unique market which cannot be compared directly with any other. This peculiar vulnerability of the drug buyer to exploitation is related to several major characteristics of the drug market which prevent price competition from acting as a safeguard and which also tends to make an economically unregulated drug in- dustry productive of much misallocation of resources in its attempt to maximize the profits of the individual firms. I do not intend to advocate punitive regulation of the drug indus- try. But the industry at present enjoys the benefits of what amounts to public regulation in its favor, through the availability of the patent privilege, trademark and copyright protection, and the laws support- ing prescribing by brand name, to name only the major advantages. This stacks the cards heavily in favor of the industry and against the drug buyer. I am in favor of corrective legislation to redress the bal- ance and increase the chances of the patient's getting fair value for his prescription dollar. At the same time, it must be stressed that to be critical of the drug in- dustry is by no means to be critical of private enterprise as such. Most industries are routinely accorded exemption from special economic regulation because they naturally tend to function tolerably competi- tively in a free market environment. But if the drug industry is permitted to retain its present special position of institutionalized protection in the economy, it will continue to display elements of both monopoly and rivalry. Spokesmen for the industry habitually refer to the intense degree of competition among firms. Unfortunately, however. the "competition" referred to is of the type which raises costs instead of reducing prices. This category of activity is generally referred to by economists as "rivalry" rather than "competition" since the latter term is usually reserved for the eco- nomically beneficial activity specifically of price competition. Ob- PAGENO="0355" COMPETITIVE PROBLEMS IN THE DRUG INDUSTRY 1905 viously, the industry which is subject to intense price cornpeti~ion at all times is the exception rather than the rule in today's economy. Nevertheless, the perennial threat, and occasional outbreak, of price competition does much to keep the price policies of the typical industry on the side of moderation. My research and consulting experience in the field of industrial organization during the last 10 or 12 years has been such as to convince me that the great majority of product inarket~ in the United States are more or less workably competitive, but the specific legal and marketing arrangements which the drug industry enjoys are such as to make it virtually a foreign body in an otherwise workably competitive economy. Then if I may I would like to turn to my supplemental statement, which focuses on the presentations of the PMA witnesses last month. Senator NELSON. We will print in full the statement that you just read from in the record at this point and we will then start your sup- pleinentary statement. Dr. Smi~. Thank you. (The complete prepared statement of Dr. Steele follows:) STATEMENT OF HENRY B. STEELE, Pu. D., ASSOCIATE PROFESSOR OF ECONOMICS, UNIVERSITY OF HOUSTON I greatly appreciate the privilege of being invited to make this statement before this subcommittee, and it is my hope that the information which I am able to present will be of some use to you in your deliberations regarding the vitally im- portant economic problems arising out of the market context within which the drug industry operates. I am an academic economist with major research interests in industrial organization and the regulation of industry, and have done much work in the area of medical economics and drug industry regulation. I received my Ph. D. degree in industrial economics from MIT in 1957, and since then have been engaged in teaching and research, as well as in consulting for private firms, United States government agencies, and foreign governments. At present I am an associate professor of economics at the University of Houston. My research in the drug industry has continued over the last seven years, and I have written three articles on drug industry economics and regulation and two articles on the supply and distribution of physicians' services, all of which have appeared in professional economics journals. In March 1965 I presented a paper on drug industry regulation before the University of Illinois Medical School at the invitation of Dr.0 Harry Dowling, then chairman of the Council of Drugs of the American Medical Association, and in February 1967, I presented a comprehensive program for the reform of Canadian drug laws and regulations before the Special Committee on Drug Costs and Prices of the Canadian House of Commons. In making the Canadian presentation. I repre- sented the government of the Province of Alberta, and of the fourteen recom- mendations which I submitted, eleven were incorporated in the Final Report of the Committee to the Canadian House of Commons. All of my research of the drug industry in the United States has been entirely self-financed, and in pre- senting this statement to the Subcommittee on Monopoly, I wish to make it clear that I represent no one but myself. 0 I. INTRODUCTION What can an economist contribute tO hearings on drug industry problems? It is curious that in all the hearings held in the United States, as well as those in England and in Canada, the original demand for the hearings has come about because of the conviction that prices are "too high", but very much of the hearings have been occupied by investigations into the safety and efficacy of drugs, and medical and pharmacological considerations have quite generally been predominant over economic issues. Yet, for every person who is moved to voice a complaint over poor drug quality, there must be a hundred who PAGENO="0356" t906 COMPETITIVE PROBLEMS IN THE DRUG INDUSTRY complain about high prices. But emphasis on drug safety and efficacy inevitably redounds to the advantage of the major firms, who would much rather fight on the battleground of relative quality than of relative prices. Their argument has two parts: First, drug prices are related to costs, particularly quality control costs. Second, drug quality and hence safety and efficacy is related to these same costs. Hence it is asserted that one cannot divorce questions of cost and price from questions of safety and efficacy; therefore, since problems of drug quality, being matters of life or death, are obviously more crucial than drug price prob- lems of mere dollars and cents, steps to assure high quality should take prec- edence over economic reforms. Economic reforms are correspondingly delayed. (This was the net effect of the otherwise admirable Kefauver-Harris Act of 1962). The argument is faulty in several respects. First, drug prices are not re- lated to drug costs, but instead to demand and ability to pay. Second, while drug quality obviously depends upon care exercised in manufacture, the cost of quality control has been shown to be a very small part of total costs, and the difference in cost between a minimal satisfactory program and a "de luxe" pro- gram would not begin to account for the difference in prices between the generic drug and its brand name equivalent. But to return to the question: what can an economist contribute to drug law reform hearings? If the data were made available, he could analyze cost-price conditions within the individual drug firms, and the pattern of inter-firm price and product competition, and arrive at an informed judgment regarding the status of competition in the industry. But such data have not been made avail- able, even to economists retained to defend the industry.1 In view of the absence of the data in the analysis of which the economist has a comparative advantage, what constructive role can he play? Primarily that of coordinating and synthesizing the economic aspects of the data which is in the record, and evaluating the economic relevance or credibility of certain of the arguments advanced by the drug interests. But since much of the evi- dence and many of the arguments transcend the realm of economic analysis as such, an economist is vulnerable to objections that he is exceeding the limits of his professional competence. Certainly the economist is not alone in this. During the drug industry investi- gations in the English-speaking countries, testifying physicians have been criti- cized for not being economists, economists have been challenged for not being physicians or pharmacologists, medical educators have been chided for not being doctors in full-time private practice, etc. But until the ideal witness appears,2 someone who is less than fully qualified has to stick his neck out and attempt to put the entire picture together. There are reasons why an economist who special- izes in the area of industrial organization and regulation is not the least qualified of all specialists to make such a presumptuous attempt. First and foremost, 1 As Professor Markham stated before this Subcommittee on December 19, 1967, In response to just such a question, "you are just not going to get those data, and I do not think-I would be less than honest if I said I would try to get them, implying that I could get them for you." (transcript, volume 23, p. 2805). Markham apparently referred not only to the confidential status given the information, but also questioned whether or not drug firms bothered to make all the cost allocations involved. Although it is to be admitted that many of the calculations can be made only on the basis of arbitrary assumptions, one would expect that well-managed firms would find it prudent to undertake such analyses for their own information. In fact, Dr. M. A. Phillips, in his Sainsbury Committee memorandum to the British Ministry of Health stated that the drug industry was no different from other organic chemicals industries in observing the customary precautions of making detailed cost studies prior to engaging in producing projects. These studies Include the costs of research and development and of promotion. Dr. Phillips' statement is unusually authoritative in that he is a drug industry consultant who has made many economic evaluation studies for drug firms. Phillips complains that "It has been found very dif- ficult to obtain figures for the cost of research and development and of promotion and advertising, although this must be known to those who have to spend this money in these ways . . ." and explains that even with the approximations his organization has to use in estimating these costs, he is satisfied that the accuracy of the estimates for these items is within 25 per cent. See Competitive Problems in the Drug Industry, Part I pp. 54-55, of the Hearings before this Subcommittee on the present matter. (It might be observed that only if there is a very large gap between cash flow and expenditures is a company actually likely to indulge in some carelessness or negligence in the relating of total costs to individual items sold.) 2 The ideal witness would probably be a practicing family physician who has also in the past been a pharmacist, a pharmacologist, a medical educator, a hospital administrator a drug industry executive, a compiler of official therapeutic compendia, a medical public welfare program administrator, a high official variously with the AMA, the PMA and the FDA, a patent attorney, a research scientist, an economist, and a member of Congress. PAGENO="0357" COMPETITIVE PROBLEMS IN THE DRUG INDUSTRY 1907 firms in the drug industry are economic enterprises. The economist George Stigler has defined business as a collection of devices for surmounting barriers to profits, and the splendid profitability records of the drug firms indicate that their business-oriented managements have very efficiently achieved this goal. Secondly, the anomalous price structures and price-cost relationships in the drug industry are so striking as to be obvious to the layman,3 while other evi- dences of delinquent performance in the, discharge of the drug industry's responsibilities are not exactly so subtle or equivocable as to be apparent only to those with highly specialized backgrounds in professional disciplines.4 It is noteworthy that drug spokesmen produce arguments in their defense which either stress or ignore similarities or differences between drugs and other industries to suit their convenience. Thus, in the PMA studies presented last month, one study treated the drug industry just like any other industry in relating the variance of the earnings of member firms in an industry (rather arbitrarily called "riskiness") to the average rate of earnings in that industry, while the other analyzed product competition in drugs in a vacuum as it were, without introducing comparative data from any other industry. But both the similarities and the differences of the drug industry should be analyzed and allowed for before making any comparative study of drug prices, costs, and profits in relation to those of other industries. The major similarity between drugs and other industries is that the firms are privately owned and are managed in the interests of maximizing profits and the value of the shareholder's investment. Provided that there is effective competi- tion among firms, the profit-maximization goal is an excellent coordinating mechanism since it motivates managements to produce a given output at mini- mum total cost and hence stimulates efficiency. At the same time, price com- petition among firms will keep price levels from exceeding for any great period of time the equilibrium levels necessary to elicit from producers that level of production which consumers demand at that price. This is basic to the classical economic doctrine of Adam Smith (who was mentioned more than once with approval by PMA witnesses last month) that the force of price competition taking effect through the market, acts as the famous "invisible hand" which by allocating resources efficiently makes each economic agent serve the general welfare even though he is only interested in furthering his own private fortunes. But unless effective competition prevails, private and: public welfare in the market are not consonant. In the drug industry, the invisible hand is invisible chiefly because it is so deeply buried in the consumer's pocket. And it is the extreme vulner- ability of the drug buyer to economic exploitation which makes the drug industry (both in economic and public policy terms) a unique market which cannot be compared directly with any other. This peculiar vulnerability of the drug buyer to exploitation is related to several major characteristics of the drug market which prevent price competition from acting as a safeguard and which also tends to make an economically unregulated drug industry productive of much misallocation of resources in its attempt to maximize the profits of individual firms. I do not intend to advocate punitive regulation of the drug industry. But the industry at present enjoys the benefits of what amounts to public regulation in its favor, through the availability of the patent privilege, trademark and copyright protection, and the laws supporting prescribing by brand name, to name only the major advantages. This stacks the cards heavily in favor of the industry and against the drug buyer. I am in favor of corrective legislation to redress the balance and increase the chances of the patient's getting fair value for his prescription dollar. At the same time, it must be stressed that to be critical of the drug industry is `by no means to be critical of private enterprise as such. Most industries are routinely accorded exemption from special economic regulation because they naturally tend to function tolerably comni~etitive1y in a free market environment. As Senator Nelson observed, In Inquiring of Professor Cootner bow a price differential of $32.62 versus $2.60 might be justified between the United States and Canada, "I didn't think I was asking a question that required expertise." (Transcript of these hearings, rolume 23, p. 2706.) Drug Industry spokesmen have patronized Morton Mintz, author of The Therapeutic Wightmare, referring slightingly to his having become an expert on the drug Industry during the Kefauver hearings. He did not become an expert on the Industry, but upon the abuses practiced by the industry-as did anyone else who read the entirety of the Kefauver Hearings carefully. PAGENO="0358" 1908 COMPETITIVE PROBLEMS IN THE DRUG INDUSTRY `But if the drug industry is permitted to retain its present special position of institutionalized protection in the economy, it will continue to display elements of `both monopoly and rivalry. Spokesmen for the industry `habitually refer to the intense degree of competition among firms. Unfortunately, however, the "com- petition" referred to is of the type which raises costs instead of reducing prices. This category of activity is generally referred to by economists as "rivalry" rather than "competition" since the latter term is usually reserved for the eco- nomically beneficial activity specifically of price competition. Obviously, the industry which is subject to intense price competition at all times is the exception rather than the rule in today's economy. Nevertheless, the perennial threat, and occasional outbreak, of price competition does much to keep the price policies of the typical industry on the side of moderation. My research and consulting experience in the field of industrial organization during the last ten or twelve years has been such as to convince me that the great majority of product markets in the United States are more or less workably competitive, but the special legal and marketing arrangements which the drug industry enjoys are such as to make it virtually a foreign `body in an otherwise workable competitive economy. iA,s I have elsewhere stated, the market characteristics of the drug industry bias it in the direction of inefficient and non-competitive performance in- five major respects.5 (1) Essential to the effective operation of a free market is the ability of the buyer to choose among suppliers on the `basis of an adequate knowledge of the price and quality of the alternative products which they may provide him. But in ethical drugs, the buyer `has no practical means of gaining access to knowledge of the range of price and quality alternatives in the market; indeed, his purchas- ing agent, the prescribing physician, is constantly oversupplied with biased infor- mation and even misinformation which facilitates confusion and ignorance of prices. -(2) The price-conscious -buyer should be able to identify the lowest-priced seller and purchase from him without `artificial impediments. Instead, the possessor of a newly-written prescription is unable to buy any -but the specified drug, regard- less of price. The willingness of the price-conscious physician to prescribe lower- priced -drugs may be compromised if he `h-as been exposed to repeated attempts to disparage low priced drugs on the part of representatives of brand na-me drugs who contend that -low price means low quality. And even if a generic prescription is written, the `buyer -has no power to compel the dispenser to sell him a reason- ably priced generic drug instead of substituting a less reasonably priced brand name equivalent. (3) There must be freedom of entry into the industry by new firms, such that high profits being made by existing firms will attract new competitors who will, by engaging in price competition, drive profits down to competitive levels. But freedom of entry in drugs is greatly lessened by the existence of the patent privilege, the trademark device, and the necessity for newcomers to match the enormous advertising outlays of existing rivals. (4) There should be an adequately large number of competitive sellers offer- ing buyers genuine alternatives in terms of product price and quality; none of the sellers should be so large that he overshadows the magnitude of his com- petitors and poses a potential threat should they incur his displeasure. In drugs, restricted entry limits the number of sellers, and while there are few if any genuine product monopolies, the size of the major firms is certainly appreciably greater than that of their smaller generic-name competitors. (5) A market is not workably competitive unless all firms act independently- there must be no overt or tacit collusion, no passive acquiescence in prior deci- sions arrived at by others and established by mutual consent . . . (in the drug industry there are) two circumstances which act to hamper independence of action. First, there is the practice of price leadership and the pricing of new medications at exactly the same levels charged for existing substitute drugs. Second, there is the fertile field of patents. While an individual patent confers a inonpoly, the scope of the monopoly privilege is limited. But in an industry with complex technology, the efficient production of a drug may require the use of processes controlled by rival patent-holders. The negotiation of the resulting Submission of the Government of the Province of Alberta to the Special Committee on Drug Costs, and Prices of the Canadian House of Commons, Minutes of Proceedings and Evidence, No. 33, pp. 2427-2429, Feb. 14, 1967. PAGENO="0359" COMPETITIVE PROBLEMS IN THE DRTJG INDUSTRY 1909 cross-licensing agreements requires the mutual compromise of patent monopoly positions, and may well stimulate such meetings of the minds as will lead to the development of a greater sense of community of interest in policies regarding prices, production, and participation in world markets. In connection with this last point, it should be emphasized that the United States drug industry is not unacquainted with outright collusion, in December 1967, Pfizer, Lederle, and Bristol Myers were convicted by the Federal District Court of New York Oity of conspiracy in restraint of trade, conspiracy to monop- olize, and monopolization in the tetracycline market. (Squibb and Upjohn were named co-conspirators but not defendants.) A further respect in which the drug market can be distinguished from other markets is in the involuntary nature of the purchase. While an individual ordi- narily ought to pay no more than the competitively determined full supply price for a product, he should also pay no less than this price since in order to supply his wants, the economy has had to allocate scarce resources which could have been used alternatively in the production of other products. But there is a differ- euce between paying the full cost of financing an activity deliberately engaged in, in contrast to one forced by accident or misfortune upon the buyer through no fault of his own. This is to suggest that the financing of drug purchases, like other aspects of health care, has an element of insurance against risk in it. Such insurance arrangements could conceivably be either public or private. A prudent man of sufficient income might participate in a voluntary health insurance pro- gram including drug costs. But the required income to make participation attrac- tive does not depend upon the "full competitive cost" of drug supply, but instead upon actual prices charged in highly non-competitive markets. Unless drug prices can be made reasonable, the possible expenses of drug therapy under a comprehen- sive private health insurance program might be so great that enormous premiums would be required. Under these circumstances, the expenses of drug therapy w-ould not constitute an insurable risk for practical purposes. This is all the more applicable to public health insurance and welfare programs, of the medicare vari- ety or otherwise. Truly comprehensive drug coverage under such plans might allow sellers of patented high-price drugs to levy a publicly-underwritten tribute on the sick and afflicted and divert a not-negligible portion of tax revenues and the national income into the hands of the pharmaceuticals industry. The only difference, fortunately, is that public authorities are in a position to exert more effective efforts to discipline high drug prices than are private insurance corn- panies. The above consideraions should be kept in mind when assessing the effects of the great variety of drug industry activities and expenditures on the price of drugs, when such costs are generally borne in full by persons involuntarily afflicted, whose earning power and ability to pay may be greatly reduced by the very circumstances which make medication imperative. The absence of workable competition among sellers is compounded by the barriers which consumers face in obtaining information regarding drug prices and quality. If neither a seller's customers nor his rivals can force him to compete, what limits are placed on his ability to, exploit his customers? Essentially only two-self-restraint and public constraint. Self-restraint is ordinarily an impediment in the management of a business enterprise, and under competitive market conditions would detract from efficiency. During the Kefauver hearings several witnesses referred to their impressions that there bad been within memory some decrease in the degree of self-restraint in marketing among drug firms. Actually, it is to be doubted that self-restraint in itself ever posed much of a barrier to high profits. Upjohn, for example, made over 30 per cent after taxes on its net worth in each of the deep depression wears 1930-1935.° But this does not necessarily contradict the observations that self- restraint was still more prevalent among drug firms prior to the second world war. In the post-war era, however, it became obvious to all that the profit possi- bilities inherent in the "Miracle Drugs" era of the industry's history were simply too vast not to be fully and intensively exploited. Although it did not prove pos- sible for new small firms to enter the market and become genuine factors to be reckoned with in the industry, larger firms found it possible to diversify by merging with existing drug houses, and producers of bulk chemicals and fine chemicals found it profitable to integrate forward into drug making and market- 6 on Administered Prices, Part 20, p. 11082. PAGENO="0360" 1910 COMPETITIVE PROBLEMS IN THE DRUG INDUSTRY ing. Some firms which had previously specialized in patent medicines began to produce prescription or "ethical" drugs. For this reason and others, the change in the composition of the industry was accompanied by an increased emphasis on sales promotion. Several physicans and medical educators testified during the Kefauver hearings that as the quantity of promotional matter increased greatly, there was a concomitant decrease in its average quality. It was argued by some that the industry should be saved from itself, it being alleged that if only one or two indifferently scrupulous firms saturated the physicians with marketing appeals of dubious merit, all the others had to imitate their tactics in self-defense. These observers tempered criticism with courtesy by omitting to single out the firms involved, but I have been amused by a phrase which I keep running across in my own contacts with people in the medical community who are interested in drug reform: "Where there's smoke, there's Pfizer." It seems safe to conclude, in any event, that the intensified rivalry which transformed drug industry marketing practices in the 1950's in all likelihood eliminated the factor of industry self-restraint from the category of feasible solutions to drug industry problems. Public constraint has so far been somewhat asymmetrical with regard to safe- guarding the drug buyer. While the acts of 1900, 1938, and 1962 have progressive- ly improved the quality of drugs in terms of increased assurance of their safety and efficacy, legislative concern for the economic health of the patient has not been nearly as marked. This is unfortunate since a sick man's economic health may often be more delicate than his physical condition. Economic convalescence can be much more prolonged than physical recuperation. And tw-o or three acute illnesses in succession, requiring drugs and hospitalization, may put a person of limited means into the ranks of the chronically medically indigent. Even the legislation regarding drug quality was obtained only with great difficulty. There seems to be two reasons for this. First, the lobbying and other governmental relations activities of the drug industry, made effective not only through the individual firms and their trade associations, but also through their allies in organized medicine and some areas of pharmacy, and probably through the "army" of some 15,000 detailmen. Second, public apathy tempered with the w-idespread uncritical attitude that the makers of miracle drugs must be miracle workers who are without fault. Even so, these obstacles to legislation have not rendered the industry completely reform-proof. The Elixir Sulfanilamide tragedy of 1937 prompted the passage of the reforms embodied in the Food, Drug, and Cosmetic Act of 1938. And although the industry has not been backward in its efforts to cultivate the public image of the private drug firm laboratory as having a complete monopoly over the creation of "miracle drugs," it has been more neglectful in publicizing the fact that, by the same token, miracle drugs can produce "miracle diseases." Dr. Walter Modell of Cornell University Medical School has observed that some 40 or more new diseases have been recognized and attributed to the em- ployment of modern drug therapy.7 But in 1961 the point was made more dra- n1atically through the creation of a sizeable "miracle generation" under the auspices of thalidomide. Again, the reaction of the public to another tragic blunder by the drug industry was to pass a reform law, the drug amendments act of 1962, to reduce the possibility of the recurrence of such a calamity and to further safeguard the public by insuring more adequately the safety and efficacy of drugs. But the danger is that public interest rapidly subsides after each calamity, while the industry itself is continuously in search of expedients which may be employed to minimize the impact of regulatory reforrns.s The net result is that while the drug buyer who obtains a drug of substandard quality (such that his health is impaired for this reason alone) is probably a very rare individual, the drug buyer who obtains his purchase for a genuinely reasonable price is probably a still rarer individual. Reform laws are now needed to institute price competition in the drug industry. One certainly hopes that such a law can be passed without the necessity for a fresh tragedy to mobilize public opinion. Still, it is hard to imagine a sufficiently striking tragedy in purely economic terms, short of some revelation that a large number of low-income indi- viduals with no access to welfare case status had died of malnutrition during Drug Industry Antitrust Act Hearings, Part 1, p. 318. 8 See, for example, Morton Mintz, The Therapeutic Nightmare, pp. 222-220. PAGENO="0361" COMPETITIVE PROBLEMS IN THE DRUG INDUSTRY 1911 some period of time because purchase of medications at inflated prices had usurped the entirety of their budgets.° But one must first inquire into what sort of reforms might most appropriately be embodied in legislation to erhance the likelihood of price competition and lower prices in the drug industry. To do this requires a study of supply and de- mand in the drug industry-of drug costs in relationship to prices. What func- tions does the industry perform? Of what benefit are these functions? How ef- ficiently does it perform them? Could any of these functions be more effectively performed under Other auspices? An attempt is made to answer these and other questions in the next section. II. AN ECONOMIC ANALYSIS OF SUPPLY AND DEMAND FOR PRE5CRIPTION DRUG5 No economic analysis of the operations of a privately-owned industry is properly oriented unless primary attention is given to the operation of the profit motive in influencing the allocation of economic resources within the market. On this point, drug firm managements, investors, trade associations, and industry critics should find themselves in agreement. But while price competition presup- poses mutually opposed interests on the part of separate competing firms striving to increase their share of the market, it also requires a pervading fundamental understanding that the institutons of the market-the price mechanism, for example, and the contractual arrangements of purchase and sale-shall be ad- hered to in a cooperative manner, such that the "rules of the game" exclude so-called "unfair" competitive tactics like the disparagement of a competitor's goods, outright dishonesty in financial dealings, and the like. One has not said much when he simply states that the relationship of supply and demand in a particular market determines the price charged, and the quan- tity bought and sold. One must also investigate the factors determining the sup- ply and demand functions. Unfortunately, many of the determinants of these relationahips are rather subtle and tend to elude quantification. Generally, speak- ing, supply is influenced by the costs of production, promotion, and distribution; by the arrangements in the terms of which production costs are allowed to influ- ence pricing and production; by the channels of distribution employed; and by the legal provisions affecting the cost and availability of imports, the techniques of promotion, the ability of new products and new firms to enter the market, and the strategies available to sellers to temper competitive pressures. Demand is influenced by the severity of the patient's need for the drug, his ability to pay (either privately or through eligibility for public aid or reimbursement), and the degree to which sales promotion efforts succeed in capturing the attention of the physician in such a way as to influence him to prescribe a certain brand of drug. A. An analysis of the factors influencing the supply of prescription drugs 1. Costs incurred by drug manufacturers The major costs incurred by drug manufacturers may be categorized as follows: (1) basic research; (2) applied research; (3) product development; (4) manu- facturing the active ingredient; (5) preparation of finished dosage forms; (6) sales promotion outlays. Each cost categOry may be best discussed by contrasting activities actually undertaken with those which would be appropriately in an effectively competitive drug industry. Basic research-The drug industry has a distinct weakness for arguments to the effect that high drug prices are due chiefly to high research costs, that most of the research is basic in nature, and that high profits are necessary to finance these large research budgets. The argument is beginning to seem a bit anachronistic in view of the repeated demonstrations that the research budget is only a very small part of the sales dollar, being lil the vicinity of 6 per cent. But there are further weaknesses in the argument. It is obvious that while high prices might be necessary to finance large research outlays, high profits are not. As has often been pointed out, profits are a residual after deducting all costs, including re- search costs. High profits indicate simply that a large part of the sales dollar w-as not accounted for by research costs-or any other costs. Furthermore, the majority of what is included under the category of "research" is something ° B. D. Bransome of the Arthritis and Rheumatism Foundation, while fortunately hav- ing no mortalities to report, did state that some individuals with severe inflammatory diseases would go without food in order to be able to afford to buy drugs. Hearings on Administered Prices, Part 14, pp. 7992-7993. PAGENO="0362" 1912 COMPETITIVE PROBLEMS IN THE DRUG INDUSTRY quite different than fundamental research. (It appears that within the last few years, some drug industry spokesmen are becoming more willing to admit that the share of their research budget devoted to truly basic research is a minor one; but there is no sign of any retreat from the position that the type of research which they do engage in is of immense benefit.) It is not surprising that drug firms devote relatively little time to strictly fundamental research since this sort of activity is inherently risky in that the firm incurring the costs may not be in a position to capture all the benefits. Basic research, in the sense of activities which extend the frontier of knowledge, is potentially the most productive form of research, but also the riskiest. Risky to whom? Here in a narrow sense the answer is clear-to the firm which invests in this activity, since failure to obtain results will mean the loss of investment. But to whom is basic research productive? It depends upon the ultimate implica- tions of what you find, which are quite unpredictable. It may help the firm doing the research, but it may help its competitors even more. Or it may help firms in different industries. Or it may simply accrue to the benefit of humanity generally. To engage in basic research, a drug firm must satisfy itself that this use of the company's funds is more productive than an alternative use in applied re- search, marketing, or the like. Such a decision is contingent upon obtaining satis- factory answers to the following questions: 10 (1) What is the risk of failure of a given project? (Fundamental research is always highly risky; this is its very nature.) (2) If the project is successful, will the findings ever be commercially applicable? (Many findings are too generalized to be immediately applicable, although they may lead to subsidiary findings which will have commercial applications.) (3) Will the resulting findings ever lead to patentable discoveries? (Often the findings are not patentable in the form in which they are ebtained, and unless they can be embodied in some sort of marketable item, may never be profitable even though patentable.) (4) Will the time horizon between initiating the research project and its fruition in the sales of commercially marketable products be sufficiently short that the discounted rate of return on the investment will justify the outlay? (If initial risk is high and the research and development period is long, the net return over time may be lower than can be earned through such alternative uses of the funds as marketing, even though the project is eventualiy successful.) (5) Will the "gestation period" of product development for patentable discoveries `be short enough that patent protection will be commercially profitable? (6) Will the discoveries prove to be of equal or greater benefit to the rivals of the firm? (With the well-developed state of the art of molecular manipulation, this is a great risk and may discourage basic research to some extent.) (7) Will the discoveries prove to be of greater application in industries outside the pharmaceuticals field? (This is a distinct possibility since phar- maceuticals is a very specialized field. It might be more likely that a major firm in the chemical industry could undertake basic research that might lead to either a chemical or pharmaceutical discovery, and in the latter event it might be possible to diversify into drugs. The reverse is less likely.) (8) Will the discoveries pose the threat of obsolescene to presently profit- able products? These ~onstitute a large number of difficult hurdles to clear, and so it is not surprising that drug firms should do little truly basic research. Actually, funda- mental research is ultimately a philanthropic activity in the sense that it always has some potential for benefiting society generally. Hence it is appropriate that most basic research be carried out by endowed foundations and universities as w-ell as publicly financed agencies specialized for this purpose. It may be concluded that under efficiently competitive conditions, private drug firn1s would not do much basic research because greater profits would be more likely if the money w-ere invested in other functions. Instead, the industry would rely upon the results of basic research made publicly available by more broadly- 10 See Submission of the Government of the Province of Alberta, op. cit., p. 2435. PAGENO="0363" COMPETITIVE PROBLEMS IN THE DRUG INDUSTRY 1913 financed agencies. And it appears that in fact the drug firms do not engage exten- sively in truly basic research. But this does not necessarily imply that drug in- dustry research activities give rise to no distortion in the spectrum of basic re- search efforts. It has been said that too little basic research in areas relating to health and therapy is done by non-industry organizations, at least partly be- cause the ability of the industry to pay high salaries (in turn dependent upon the high profitability of drugs under present market conditions) diverts too large a portion of the very small pool of qualified investigators and technicians away from public employment in basic research and toward private employment in applied research and product development, testing, and application-all of which are lower-priority uses' for their very scarce skills. The remedy would seem to lie not in increasing the amount of basic research done by private firms, but in taking steps `to reduce their ability to drain off the best scientific personnel for work in less productive employments than they are capable of pursuing. Applied Research--The economic rationale of applied Tesearc'h is quite straightforward in any industry: to serve as one means of implementing a profita- ble marke'ting operation. The direction and emphasis of this research in the field of drug therapy is influenced by (1) the' nature of the patent system: (2) the impact of the patent system on the organization of the industry; and (3) the effect of industry activities on research outside the industry. The mere existence of the patent privilege for drugs biases research toward patentable inventions and away from areas where no patents can be obtained. This discriminates against basic research and stimulates applied research. It also discriminates among different channels of applied research. From the medical point of view, research is unbalanced due to an unduly intense emphasis on chemotherapy, while the complementary fields of nutrition, public health, biochemistry, and preventive medicine are underemphasized. Antibiotics provide the most sharply focussed example. Concern has been widely expressed that antibiotic therapy may ultimately prove to be a blind alley due to overuse and the development of resistant strains of micro-organisms. It would seem wiser to spend less effort on activities which tend to make micro-organisms increasingly resistant to control, and more effort on attempts to make man naturally more resistant to micro- organisms. Furthermore, by biasing efforts toward applied research, the patent system will reduce the scope of basic research findings which can be applied, and ulti- mately will depress the productivity of' applied research. There has been much discussion in recent years of the "increasing cost" of drug research per new discovery. (This was true even before the FDA began to implement more strin- gent controls over new drugs.) But to speak of increasing costs is simply to refer indirectly to the decreasing productivity of efforts. Again, antibiotics offers a good case in point. Applied research here was productive for a good many years, in large measure because fundamental research in this field had already elu- cidated much of the mechanism of infection by micro-organisms. Bacteriology was already an established field of study. But the same is not true in the other major areas of drug research, such as tranquilizers and oral antidiabetic agents. Here there has been a less prolific output of various useful drugs and less enthu- siasm among independent authorities regarding the extent to which the later drugs are advances over the earlier drugs, and perhaps even over related drugs which antedated the "miracle drug" era. Another bias of some interest is related to the fact by protecting new prod- ucts, either as such or through exclusive process privileges, the patent law biases applied research in the direction of concocting new products rather than fully investigating the properties of known compounds. As Prof. George Wright of the University of Toronto has contended, it seems to indicate rather overly one-sided emphasis that new drugs coming from drug houses are almost invari- ably novel concoctions and therefore patentable, while the reservoir of some two million already known compounds hasP only been pharmacologically investiga'ted to a very modest degree.1' To the extent that patent reforms can reduce biases of this type, resources will be allocated more efficiently throughout the drug research sector of the ~` Professor Wright advocates more screening of known compounds rather than an exclusive emphasis on the concocting of new ones, on the assumption that "much is yet unknown about the association between chemical structure and pharmacological action," observing that the screening approach originally brought the sulfa drugs into existence. Canadian Henrings, op. cit., no. 8, p. 540. PAGENO="0364" 1914 COMPETITIVE PROBLEMS IN THE DRUG INDUSTRY economy. Basic research will increase relative to applied research, so that there will be a greater variety of fundamental research findings on the basis of which upplied research can be conducted. The potential which resides in the areas of nutrition, biochemistry, public health, preventive medicine, and other areas will be more nearly capable of full achievement. The bias in favor of contriving new compounds instead of systematically conducting an empirical study of existing compounds will be reduced. And the amount of exclusively imitative research, and of other types of research of secondary importance aimed at finding a patentable vehicle for a "blitz" sales promotion campaign, will be reduced or even eliminated. This latter phenomenon of duplicative activities is a major element in drug research today, as conditioned by its patent orientation. Basically, there is an over-intensive exploitation of those approaches known in the past to have yielded profitable drugs. Since the number of known approaches is limited, it is within the capacities of major firms to explore several of them, and since all firms are conscious of the commercial advantage of being able rapidly to duplicate the successful new drugs which their rivals may find, the research programs of large drug firms tend to duplicate, at least in part, the programs of their major rivals. (This was attested to not only by many witnesses at drug industry hearings, but also by the near-simultaneous discovery of several drugs by two or more firms.) This constitutes a compounded misallocation of resources: not only are scarce talents diverted from basic to applied research, but wasteful duplication of effort on precisely the same applied research projects seems to be common. Much of the criticism of the "molecular manipulation" approach can be most appropriately directed at this phase of the industry's operations. The ideal manipulated molecule is one which is pharmacologically identical with the profit- able product of a rival, but is legally distinct in the sense that a patent may be obtained. However, it is the latter criterion which is crucial, not the former, and the typical me-too version of an existing drug is of dubious superiority, if not absolutely inferior, to the original drug which it is intended to supplant. The most impressive testimony regarding the prevalence of misdirected research in the major drug houses came during the Kefauver hearings from two physicians who formerly served as medical directors for major firms. Dr. A. Dale Console, formerly with Squibb, when asked whether there was much drug research which produces nothing worthwhile and is not intended to, replied: "I think the majority of it is in that category . . . and I should point out that with many of these products, it is clear while they are on the drawing board that they promise no utility; they promise sales. It is not a question of pursuing them because something may come of it . . . it is pursued simply because there is profit in it." 12 He also reported that imitative research could crowd out productive work: "When a `crash program' comes along in which some product is being pushed in order to get it out before a competitor gets it out, it is not unusual for a worthwhile research program to be postponed so that the people can be taken off it to be put on the `crash program'. Very frequenty some of these programs are never picked up again. So I think that good research is actually hampered by this type of thing." 12 Dr. Haskell J. Weinstein, formerly w-ith the Roeririg division of Pfizer, tie- nounced industry managements for wasting the time of their research personnel: "Their talents should not be expended on patent-bypassing chemical manipula- tions, on ridiculous mixtures of drugs, or inconsequential additives to established drugs. Since the number of well-trained capable scientists is severely limited, their potential should not be wasted. The long-term benefits of the appropriate utilization of the abilities of these skilled individuals would be immeasurably greater." 13 This illustrates some of the subsidiary distortions in applied research resulting from the patent incentive: not only modified molecules, but the development of often irrational combinations of existing drugs which lack flexibility and com- pound the problems of dosage and toxicity, and the devising of additives which 12 Hearings on Administered Prices, op. cit., part 18, p. 10379. 12 Ibid., part iS, p. 10254. PAGENO="0365" COMPETITIVE PROBLEMS IN THE DRUG INDUSTRY 1915 may constitute questionable and/or necessary flourishes in the interest of in- creasing a drug's absorption rate, guarding against side effects, etc.'4 Product development-The category of product development includes many activities, such as experimental and clinical testing, determination of appropriate dosages and dosage forms, obtaining FDA approval for marketing new drugs, constructing pilot plant facilities, etc. Subsequent to initial marketing there would be product application work relating to long-run evaluation of the total effects of a drug, improvements in dosages, revisions of brochures, and related activities. In an efficiently competitive drug industry, profit prospects from marketing new drugs would be moderate, and the temptation for extreme haste in product development would be correspondingly slight. It is very important that this temptation be minor, since proper testing and evaluation of new drugs is an important and time-consuming task. And, as Dr. C. D. Leake of Ohio State Univeriity has observed, "There is no shortcut from chemical laboratory to clinic, except one that passes too close to the Morgue." 15 But before certain needed reforms were legislated in 1962, many firms yielded to the temptation to rush new drugs thru the development phase and on to the market as soon as possible, limiting experimental and clinical work to the minimum acceptable levels under the old legislation, harassing FDA staff members into approving inadequate applications,'6 and even skipping such seemingly essential product development stages as pilot plant operatio~." Furthermore, with an inflated num- ber of drugs being clinically investigated in the expectation of reasonably rapid FDA approval, the available time of the most highly qualified investigators was soon completely employed, and recourse to less trained, less capable, and in some instances less scrupulous individuals was necessary. But drug evaluation by unqualified investigators can be worse than useless.'8 Since the passage of the 1962 legislation, there has been much improvement in this area. More stringent requirements for approval of new drug applications have been imposed; the number of new drugs being evaluated has apparently declined, making possible an increase in the average quality of evaluations, and the morale and effectiveness of the FDA has greatly improved. This is apparently one area in which drug safety reforms may have been successful in eliminating certain economic wastes as well as improving drug quality. But it should be noted that any reduction in total drug development outlays would be likely to result from a reduction in the number of new drugs under investigation; the average cost of investigation per drug is likely to increase, and this increase is certainly in the best interest of public health. Manufacturing of the active ingredient-In an efficiently competitive drug industry, each stage in the production process would be carried out at minimum 14 Dr. Harry F. Dowling of the University of Illinois Medical School cited an excellent example involving both molecular manipulation and the use of Inconsequential additives. Lilly discovered erythromycin in 1952, and in 1953 Pfizer retaliated with a molecular shadow, carbomycin, which proved less effective in human disease than in the test tube, and was finally withdrawn from the market in 1960. Pfizer tried again in 1956 with an- other chemical echo of the erytliromycin, eleandomycin, and in 1957 modified its own modification, called it triacetyloleandomycin, and advertised it widely as a major break- through in that the same oral dose as eleandomycin produced somewhat higher concentra- tions of the drug in the bloodstream. Lilly responded in 1958 by modifying its original erythroinycin and marketing it in the form of its propronyl salt, claiming a higher blood concentration rate than could be achieved with triacetyloleandomycin. None of the four later drugs had any real advantage over the original discovery, since slightly higher doses of the original drug would have been as effective as the later variants. Ibid., part 24 pp. 14167-14168. 15 Ibid., part 15, p. 10418. 1~ See testimony of Dr. Barbara Moulton, Ibid., part 22, Pp. 12025-12032. "Lederle bypassed the Pilot-Plant Stage with its Triamcinolone. See Fortune May 1960. p. 276. 18 Dr. Maxwell Finland, Harvard University Medical School, cited an instance u-here a clinical investigator had reported successful treatment of 100 cases of staphylococcal pneumonia u-ithout a mortality. Since the usual mortality rate among the patients con- cerned is 50 percent, the drug would appear to be miraculous. But upon investigation Dr. Finland concluded that not a single case of staphylococcal pneumonia had been present, and inferred that the investigator was incompetent to diagnose the presence of the true disease from the laboratory cultures. He concluded pointedly: "This is the sort of thing that I say is dangerous because another doctor who knows how to make a diagnosis of staphylococcal pneumonia will use that drug to the peril of his patient." Hearings on Administered Prices, op. cit., part 24, pp. 450-451. PAGENO="0366" 1916 COMPETITIVE PROBLEMS IN THE DRUG INDUSTRY cost, and without incurring the cost `burdens of excess capacity, which one suspects are characteristic of industries like drugs which shun price competi- tion and rely on product differentiation.19 Although very large investments in plant and equipment are common in the heavy chemicals industry, in the pharmaceuticals field this is not the rule. The investment in facilities required for the production of the active ingredients varies considerably from drug to drug. For those active ingredients which can most efficiently be produced by truly mass-production techniques, production by makers of fine chemicals or even bulk chemicals would be indicated. But for many drugs, the required investment is relatively modest in comparison with the supply of funds avalilable in capital markets. Relatively small firms can efficiently produce the active ingredi- ents for such drugs. Mass production methods are not appropriate for many drugs since the physically minute quantities used in dosage forms require only a small total annual volume of output. But it still might be more efficient for a small firm to buy the basic ingredient under contract from a larger firm. If the market were competitive, the relative economies of scale in produnction versus those in dosage form preparation and in distribution should govern the extent of the functions assumed by different producers at different stages in the industry. If patents posed no real barrier to entry into drug selling, chemical raw mate- rials and intermediates could be made by bulk chemical companies, the active ingredient could be made by fine chemicals producers, and the finished dosage forms could be tableted and packaged by drug makers to be distributed through various channels. Without barriers to entry, the comparative costs of each stage in the industry would determine the allocation of functions among differ- ent firms. As far as financial requirements are concerned, there is no reason why a large number of relatively small firms might not compete effectively in the drug market. But under present market conditions there are a number of factors which dis- tort the division of labor among firms and introduce other criteria than compara- tive costs as determinants of The Degree of Specialization within and Among Firms. A relatively small drug maker might find a new drug, patent it, and under- take production of the active ingredient on his own premises, despite the circum- stance that his costs might be high due to his inexperience, inappropriate facili- ties, and overall lack of adaption of his operation to the requirements of fine chemicals manufacture. Production would still be undertaken at higher costs, however, if the patent holder wished to prevent the "know-how" which is ordi- narily not disclosed in the patent, from being acquired by another firm. The inef- ficiencies inherent in this arrangement could be partially overcome by the merger of the small firm with a larger producer of bulk or fine chemicals, but this would increase the market power of the formerly small firm relative to its rivals. Forward integration by merger is also stimulated by marketing practices. The intense sales promotion of drugs under brand names particularly thru nationwide advertising and detailing, is a practice which creates economies of large scale marketing even though none may exist in production. This is unusual; ordinarily it is economics of large-scale production which prevent the successful operation of a large number of small firms, but in drugs it appears that while efficient pro- duction might occur at a very small level of output, the exploitation of modern marketing techniques in the drug market context can be taken full advantage of only by a very large firm. Where the factor of "know-how" is not important, drug firms may contract out the production of the active ingredient to specialists. In such cases, the ratio of the price of the bulk drug to the market value of the substance when embodied in final dosage forms and sold to distributors is well worth noting. Ratios of the order of magnitude of one hundred to one are not unknown.2° This is very simply explained. There is price competition among the firms which make the active in- gredient, but none in the sale of the finished product. If competition were to be 19 By way of example, when Bristol was producing about one-third of national tetracyc- line output, it still had 80% excess capacity in this drug. Drug Industry Antitrust Act Hearings. purt 4, p. 2056. 20 During the recent Canadian Hearings, Empire, a small generic firm, estimated that it could manufacture the drug diazepam for $68 per kilogram and a kilogram of diazepam embodied in dosage forms is worth about $20,000-a ratio of 125 to 1. (This makes diaze- pam worth about 16 times as much as gold.) See Alberta Government Submission, op. cit., p. 34. PAGENO="0367" COMPETITIVE PROBLEMS IN THE DRUG INDUSTRY 1917 introduced at the finished product level, drug prices would decline until they were more in line with production costs for the active ingredient. Preparation of jinished dosage forms.-The simple technology of the prepara- tion of most finished dosage forms, the low operating costs of these processes, and the modest capital requirement for such facilities, renders this stage of the industry ideally suited for workably competitive market performance. The processes involved for most dosage forms are technologically routine and ele- mentary, tabletting and bottling being particularly trivial operations technically. (After all, every pharmacist is taught-and taught well-to do such compound- ing operations on his own premises. It is both amusing and dismaying to ob- serve industry attempts to convince the general public that there is some magic in the preparation of even the simplest dosage forms, which is by implication a secret known only to the major brand name firms.) It has been shown by evidence presented at drug hearings both in the United States and Canada that for the typical drug, "factory costs" (producing the active ingredient and making and packaging the dosage forms) are a minor part of the wholesale price. There is no purely economic reason why numerous small firms could not contract out the manufacture of the active ingredient and then tablet and package the finished dosage forms on the basis of a quite moderate total invest- ment. Brisk price competition between many small sellers of drugs might develop if production costs were the only barrier to entry. And this in fact is the prevailing mode of market behavior for those small firms which produce generic name drugs for which tight patent control could not be achieved over their manufacture and/or the sale of their bulk powder. These firms can either produce or buy the bulk powder at the lOw prices which result from competition among bulk suppliers. These firms then tablet, package and sell the drugs at low prices representing their low costs of production. But in most markets these generic drugs compete with their presumed brand name equivalents, and it is likely that the true production costs of the brand name sellers are even lower than those of the generic firms. But does this mean that the large firms choose to undersell the small generic houses? By no means, they charge prices up to ten or more times as high. But does this not mean that they are not able to make any sales at these high prices? Again, by no means. They outsell the lower-priced drugs ten or twenty to one. To an economist who has been trained to expect that quantity sold is inversely related to prices charged, this is a dumbfounding situation. The answer of course, as is well known, is related to sales promotion tactics. $ales promotion ontlays.-Any spokesman for the domestic drug industry will tell you that its outstanding accomplishments have been in the area of research. But any well-informed expert on marketing is much more likely to tell you that the drug industry's real expertise lies in the area of sales promotion. And indeed the relationship of the marketing budget to the research budget suggests as much. But the myth seems to persist in the general mind that research budgets exceed advertising budgets, despite repeated demonstrations that the latter is several times as high, as can be verified by even a superficial examination of the financial statements of any large drug firm. It seems to me that public education can sometimes be furthered more effectively, therefore, by witty ancedotes or epigrams than by mere statistics. For example, in my experience, Dr. A. D. Console seems to have done more than anyone else to expose to salutary ridicule the dubious nature of much drug industry "research" in his memorable statement during the Kefauver hearings; "They stress that there are many failures for each successful drug. This is true since it is the very essence of research. The problem arises out of the fact that they market so many of their failures." ~` The same device can be used to put drug prices research costs, and sales promotion outlays in perspective. During the recent Canadian drug hearings, one participant observed: "for every dollar the dru° industry spends on research, they spend four dollars telling you about it, and charge you ten dollars more for listening." If anything, this witty statement errs on the side of moderation. Research costs apparently amount to something less than one-fifteenth of sales price, and somewhat less than one-fourth of sales promotion outlays. ~` Hearings on Administered Prices, op. cit., part 18, pp. 10372-10373. PAGENO="0368" 1918 COMPETITIVE PROBLEMS IN THE DRuG INDTJSTRY In an efficiently competitive drug industry, sales promotion activities would be adapted so as to reflect the structure of information needs and to take ad- vantage of the ability of the relevant individuals to absorb information, in order to minimize the cost of providing the necessary data. Since ethical drugs are ethical in that they cannot be bought over the counter, advertising to final con- sumers would seem unnecessary. Instead, advertising need be directed only to prescribing physicians. But doctors would seem to constitute a specialized market for several reasons. First, they are highly trained professional men. Second, they are extremely busy. Third, they are unquestionably prosperous. A rational ad- vertiser would adapt his sales strategies accordingly. Since the physician is quite intelligent and well-trained, he can respond adequately to modicum of purely informative advertising material and need not be bombarded with masses of in- sistently persuasive promotional appeals. (If a word to the w-ise is sufficient, the drug firms are guilty of a colossal insult to the physician's intelligence.) Since doctors are busy, the time they can devote to the study of drug information of all sorts is limited. And since doctors are men of some financial means, and profit from the availability of good drugs, they should be expected to pay the costs of being supplied with adequate drug information. The sales promotion outlays of the drug industry, like all their other expenditures, are paid for entirely by the patient. Even if doctors find that detail men are convenient sources of informa- tion, the patient is still subsidizing the doctor by providing him with a number of tutors, and biased tutors at that. It is debatable whether this sort of subsidy is justified. However, the amount of the subsidy, rather than the nature of it, is the chief point. If the doctor were to be required to bear the costs of obtaining his own drug information, he could still pass the costs on to his patients through higher fees. But the costs of alternative means of being supplied with informa- tion-subscriptions to official compendia and their periodical supplements, or to independent newsletters such as The Medical Letter, would be a very small frac- tion of the amount spent at present on sales promotion by drug firms. Obviously, the necessary information on drugs must somehow be supplied. It may be supplied by the companies, by independent evaluating organizations, or by government. If supplied by government it can either be made available upon subscription or can be distributed free of charge. By whatever medium it is com- municated, it shOuld achieve four objectives: (1) insure adequate flow of accu- rate and unbiased information; (2) minimize the volume of redundant communi- cation; (3) make informative communications more concise; (4) eliminate all misinformation. The provision of this information by the companies themselves has been sharply criticized, chiefly by the recipients of these promotional attentions, on numberless occasions. This is not the place to deal in detail with these criti- cisms, but the chief complaints may be recapitulated in terms of failure to achieve the goals of a satisfactory communication system: (1) constant interference of commercial bias; (2) excessive communication, such that the volume of indiffer- ent information and just plain "noise" minimizes the likelihood of the detection of the occasional communication of genuine value; (3) emphasis on persuasion and suggestion rather than upon providing genuine information; (4) redun- dancy of communication as a result of the mutually offsetting nature of sales pro- motion rivalry among firms not competing on a price basis; (5) the presence of a certain amount of outright misinformation, chiefly in regard to inadequate dis- closure of side-effects or contraindications. Evaluation by independent sources, or by public bodies, should eliminate com- mercial bias and minimize the temptations to indulge in excessive or overly per- suasive communication. Intentional misinformation should also be eliminated, although the fallibility of human agency will render any drug information sys- tern less than perfect. It is, however, doubtful that the first goal of an informa- tion system-insuring an adeqvate flow of information-will be achieved by any system which leaves the aquisition of the information source up to the discre- tion of the physician. Proof of this is supplied by the small fraction of the medical profession subscribing to the Medical Letter: only about 15 percent, which is to be deplored. (It is of course possible that if the institution of price competition in the drug industry lowered profit margins and eliminated the de- tail-man, more physicians would subscribe.) In England the Prescriber's Journal is distributed free to all doctors, and a similar step is being considered in Canada. We in the United States should be hesitant to extend such a subsidy PAGENO="0369" COMPETITIVE PROBLEMS IN THE DRUG INDUSTRY 1919 to the members of a well-paid profession, but this may still be the least costly way of insuring at least the availability to all doctors of adequate and un- biased drug information. And while an element of subsidy to the doctor would remain, the subsidy would now be financed by the public as a whole, rather than that smaller and to some extent disadvantaged segment of it which in a given period of time finds it necessary to purchase drugs in varying amounts. If adequate drug information were to be made available from unbiased sources, by how much might drug industry sales promotion costs decline? Unless steps were also taken to institute price competition in drugs, expenditures would probably not drop at all, and might even increase, each firm trying not only to offset advertising efforts of other firms, but also the now more widely-dis- tributed data from unbiased sources. To explain why this outcome is likely, the crucial role of sales promotion in the drug industry must be further explored at this point. The key role of sales promotion lies in its ability to substitute for productive research on the one hand, and for genuine price competition on the other. If a firm's research department discovers a highly effective new drug of un- mistakable value, the drug tends to advertise itself. Physicians themselves rapidly spread the news. On the other hand, as Dr. Walter Modell of Cornell University Medical School has stated, "The more a drug has to be peddled, the more one begins to wonder why." 22 Dr. Console, formerly of Squibb, was more explicit-about the relationship between unproductive research and the advertising budget: "Advertising is called upon to make successes of research failures." 23 Since it appears that almost any drug will sell, at least for a while, if promoted intensely enough, it does appear that the good offices of the sales promotion department may compensate greatly for the indifferent fortunes of the researchers-but at quite a cost, in terms of premature and usually unneces- sary obsolescence of existing products. Even these costs, however, are con- verted by drug spokesmen into rhetorical~ capital: they are construed as meas- uring the risks of product obsolescence, allegedly due to the rapid development of superior medications, and these "risks" are supposed to constitute a justifica- tion for the high profit rates characteristic of the industry. The "risks" involved are not wholly illusory, but are not, as claimed, inherent in the research process. Instead, they arise from the way in which drugs are developed and promoted. The high profits cannot be justified by the high risks because the height of the profits induces expensive and disruptive product competition which manifests itself in such "risks". But if one were tO create price competition and reduce profits, this mode of product competition would be too expensive to support, and the "risks" of product obsolescence would be diminished in proportion to the decline in profits. Even more important is the degree to, which sales promotion can substitute for price competition. This is true in most industries, but drugs are unique in that such efforts can almost completely suppress price competition and further- more can seriously discredit the price competitor. Again, the sheltered institu- tional circumstances of drug marketing should take the credit. If a company can monopolize the eye and ear of the prescribing physician, it monopolizes the most important drug market, given the existence of laws supporting brand-name pre- scribing. But monopolizing the physician's attention becomes increasingly ex- pensive when more than one firm tries this strategy. It soon becomes so expensive that only a limited number of firms have the resources to continue-hence, as mentioned before, economies of large scale marketing act as barriers to entry where production costs in themselves would not be prohibitive. Smaller firms cannot ordinarily bring their products to the attention of the physician, even though they may be selling at perhaps 90% less than the brand name drugs. And there is yet another way in which the sales promotion techniques of the industry put the small firm at a prohibitive disadvantage. Although the small low-price firm cannot ordinarily make its presence in the market directly known to the physician, the doctor may eventually become aware of its existence and perhaps wonder how the midget firm can undersell the giant by a ten-to-one ratio. The detailman is ready to supply the answers. It appears likely, on the basis of the testimony of many doctors at public hearings, that a major reason 22 Drug Industry Antitrust Act Hearings, Part I, p. 32P. 28 Hearings on Administered Prices, op. cit., part 18, p. 10372. 81-280-68-pt. 5-24 PAGENO="0370" 1920 COMPETITIVE PROBLEMS IN THE DRUG INDUSTRY why drug firms spend so much on detailmen is because of their peculiar ability to disparage, with relative impunity, quality of the products of the low-priced generic firms. The conscientious physician is naturally vulnerable to having his confidence in low-priced drugs undermined by such disparagement since he is personally not in a position to evaluate the relative quality of drugs. Whenever a buyer lacks full information on the nature of the product at the time of pur- chase selling efforts take on some of the coloration of a "confidence game" where the buyer is induced to take the seller's word for it that his product is better than his rivals', it is suggested that to be unknown is to be suspect in the drug industry, and that it is common knowledge that most generic drugs are sub- potent, adulterated, etc., and may even be made by counterfeiters. Thus the i;hysician may be induced to equate low price with low quality and hence shun generic prescribing. Hence drugs priced 90% below the market may account for less than 10% of all sales. For these reasons there is no other industry in exist- ence where the disparagement of the quality of lower priced products can so completely substitute for price competition. Is there any substance to this disparagement? There should be very little reason to suppose that low price need be associated with low quality. If any- thing, doctors should be more hesitant to prescribe by brand name, since counter- feiters (who make more on $100 bills than on nickels and dimes) naturally specialize in the high-priced brand name drugs. As far as economic motivations to save costs by cutting corners are concerned, these should be minimal in drugs. The strength of the positive motivation-cost reductions-is relatively small, since neither the costs of quality control nor of the active ingredient itself are particularly large components of total cost. For a major firm, quality control costs seem to range from about one to three percent of the sales dollar, and al- though the figure would probably be higher for a small firm, it should still not be a controlling factor in costs. And while official compendia specify a certain range of allowable variation for drug potency, the typical range is only about 90 to 110 percent. But since the cost of the active ingredient in a given drug is usually only a minor part of total cost, the cost saved by orienting the production process to produce an average content of 90 percent of stated label potency would save at most only 10 percent of this cost. Furthermore, it would inevitably mean the production of a number of substandard drugs and would expose the firm to punitive actions by the FDA. This brings up the negative aspect the deterrents to substandard performance. Both generic drugs and their brand-name equivalents must meet official standards specified in drug compendia. Experts have testified that there is no therapy gain to be achieved by producing to purity standards "exceeding official" "minimum" standards. The products of all producers are held to the same inspection standards, and a small firm will be even more strongly motivated than a large firm to conform to the regulations since the impact of a given fine will be much more disruptive to its finances. Brand name firms have alleged that there is no therapeutic equivalency even among drugs which sati~fy the requirements of official compendia. I sympathize with such witnesses as Dr. Solomon Garb, who professes to find this sort of argument both elusive and baffling.24 I respect Dr. Garb's opinion and share his suspicion that the differences are trivial and that they cannot be meaningfully specified. It is very hard to follow drug industry arguments which suggest that because no two drugs, or capsules, are absolutely identical, that one should buy brand names and shun generic names. The emphasis on the unique nature of each pill is reminiscent of the philosophical doctrine of nomanilism, which im- plies that no generalizations are possible since everything is in a unique category by itself. I submit that drug firms are more pragmatic than nominalistic in their serious moments. The most authoritative testimony of this point would appear to have been given by Dr. Lloyd Miller, Director of Revision of the United States Pharma- copeial Convention, before this Subcommittee, in stating that "there are not 24 Dr. Garb expressed himself as follows: "It seems to me that if any group of drug manufacturers wish to use the argument that their brand name drugs are better because of certain differences, and that the doctor knows what these differences are, they should show how the doctor finds out these differences . . . I think the differences are trivial, but my point is I do not know that they are trivial, because I cannot find out why they are. I have never been able to find out what the difference is between one brand of the drug and another brand of the drug." Competitive Problems in the Drug Industry, op. cit, Part 2, p. 545. PAGENO="0371" COMPETITIVE PROBLEMS IN THE DRUG INDUSTRY 1921 more than a dozen examples where the difficulty (questionable clinical equiva- lency of drugs meeting DSP standards) has been discovered, and it is not gen- erally true even for all of them." My own evaluation of the "generic equiva- lency" or more precisely therapeutic equivalency dispute is that the brand name firms have greatly overstated the significance of any valid or partially valid elements in their position, and that the argument on the whole is with- out merit. The verdict of lack of merit seems inescapable since there should be no reason to assume that the therapeutic equivalency of two brand name prepara- tions is any greater than that of a brand and a generic name preparation. As Dr. Martin Cherkasky recently testified before this Subcommittee, "I must tell you that I feel quite insecure at the same time about the performance of some of the major drug companies in this country." 26 The intensity of the dispute is all the more puzzling since it appears to take for granted that pharmacology is much more of an exact science in practice than in fact it isP It would seem that the best argument which the major drug firms have in support of their charges that unsafe drugs are on the market is the contention that FDA inspection is not adequate. This is a perennial charge, and one which in recent years has become much less plausible because of long-needed increases in FDA staffs and budgets. But if any doubt remains as to the adequacy of FDA inspection, it should be resolved by providing FDA with still more funds to the full extent which may prove necessary. Perhaps. Still, it is the direct respon- sibility of the FDA to safeguard the public from poor drugs, in a much more immediate sense than this responsibility is shared by USP. Batch certification of all drugs might be one solution. Dr. Solomon Garb's proposal that all drug manufacture be subject to continuous federal inspection is another solution. Dr. Miller charges that the former proposal would be unnecessarily costly; Dr. Garb reports that drug makers contend that the costs of his plan would be astronomical (although, as he points out, there is private inspection today at less than astronomical cost, and it would seem quite feasible simply to substitute pub- lic for private inspection with avoidance of excessive duplication of efforts.) An economist would wish to make one point. Regardless of the methods adopted, the physician must be made confident that all drugs on the market are safe to use. Relative costs of the methods employed to bring about this confidence are im- portant, but it is very unlikely that the costs involved would fail to justify the resulting benefits: complete eliniination of substandard drugs and the opening up of much of the prescription market to price competition between low-priced generic drugs and currently high-priced brand-name equivalents. The cost of the industry's current system of "insuring" drug quality by advertising intensively to promote brand-name drugs and discourage generic prescribing does tend to in- crease the sales of the former drugs and reduce the sales of the latter. But beyond that, it perpetuates a great barrier to price competition and places an ~ Ibid., Part 2, p. 514. ~ Ibid., Part 2, p. 668. ~ A layman can perhaps be forgiven for Introducing Into the record the Idea that with few exceptions (but among them some admittedly very important ones) in the present state of our knowledge, drugs tend to be relatively blunt Instruments when employed In human therapy, so that the question of "therapeutic equivalence" when generalized to Include all drugs frequently becomes subsidiary to the question as to whether or not the drug can be depended upon, In a particular circumstance, to do its job at all. Placing some emphasis upon this may put the issue of therapeutic equivalence In a more adequate per- spective. The significance of this Issue would become much greater If one could always assume that If each drug Is optimally manufactured and administered, It will accomplish a predictable, effective, and exclusively beneficial therapeutic result upon each adminis- tration. But where the effect of the drug Is unpredictable, Imperfect, and tempered by side-effects, therapeutic equivalence as such is a much less paramount consideration, and the Importance of other factors, such as biological variations among human recipients, incidence of side effects, and the inherent uncertainty regarding the mechanism of action of the active Ingredient Itself, become more significant. The notion that many drugs are blunt instruments is not propagated by the Industry, but It Is hard to avoid drawing such a conclusion, If for no other reason than the Inability, to date, of molecular engineers to eliminate side effects. Candid doctors have submitted that these effects should not' be rerarded as incidental disadvantages, but are an integral part of the total action of the drug, and should rather be referred to as "concomitant effects." Just as burning down the house would be a "broad spectrum" recipe for roasting a pig, so also must one regard antibiotics which Indiscriminately kill both harmful and beneficial organisms within their range of activity as being relatively blunt Instruments. One is reminded of Buck- lngham's characterization of his parody of the dramatic hero, Drawcansir, in the play The Rehearsal. Drawcansir's battle cry was: Let petty kings the names of factions know; When e'er I fight, I slay both friend and foe. PAGENO="0372" 1922 COMPETITIVE PROBLEMS IN THE DRUG INDUSTRY unjustifiable price premium on the brand name drug. This system amounts to insuring drug quality by what is tantamount to private taxation, and levied through the charging of prices far above production costs. But this does not eliminate the production of substandard drugs, since FDA studies show that a disturbingly high percentage of both brand and generic name drugs do not meet requirements. It should be cheaper to institute universal FDA inspection, of plants and products. While the cost to the government might increase (unless as in the case of meat inspection, the companies were to pay the salaries of the inspectors) the resulting increase in price competition which should be expected, once arguments that low price means low quality become manifestly specious to all, should reduce the expenditures of drug buyers greatly.~8 2. Influence of Economic Aspects of Drug Distribution on the S'uppiy of Drugs There is one complaint voiced by the drug manufacturers with which I have some sympathy. While about 50 per cent of the retail price of drugs is accounted for by distribution costs and markups, rather more than 50 per cent of the criticism of high drug prices has been centered upon the manufacturers. The proximate cause of the seeming disproportion in attention given the drug makers would appear to be the great visibility of their profit achievements-outstand- ingly high average profit levels over the last 15 or 20 years-in contrast not only with the absence of any data to show that the druggists have been similarly successful in feathering their nests, but even a general impression to the con- trary. The number of retail pharmacies seems to have declined during the period of greatest drug firm success, and many pharmacists who have failed to make what seemed to them a reasonable living in their own profession have reportedly been hired by the drug industry, where they can obtain more remunerative-if perhaps less productive-employment. But low profit margins in themselves do not necessarily attest to effective coin- petition, since not even a complete monopolist is assured of high profits unless he can operate efficiently. On the other hand, not all inefficiencies are necessarily traceable to the mismanagement of the individual firm. Some inefficiencies may be thrust upon the distribution level from without; others may be unintended consequences of policies fostered within group itself. But before looking into these questions, it is first useful to distinguish between wholesale and retail distribution. The wholesaling function seems to be the most efficiently performed stage in the industry, chiefly because the wholesaler operates in the most competitive market. Drug manufacturers have their markets protected by patents, trade- marks, sales promotion outlays, and the relatively small number and large aver- age size of the major firm. Druggists also enjoy rather protected markets be- cause of the practice of brand-name prescribing, antisubstitution laws, and other regulations which put the consumer at a disadvantage, plus the advantages associated with being a closed profession regulated by semi-autonomous profes- sional associations which have at least the potential for limiting the number of qualified practitioners and hence influencing the rate of entry into pharmacy. The wholesaler, however, has no comparably strong bargaining position. There are many w-holesalers, mostly very small, and no appreciable barriers to new entry. Furthermore, if drug makers can perform their own wholesaling func- tions more efficiently than the independent distributor, they will integrate for- ward and sell directly to retailers. And if retailers can do better for themselves ~ I am aware of no recent estimates of the cost of making FDA inspection fully adequate. However, an order of magnitude approximation of the relation of probable costs to re- quired cost reductions can be made from data presented at the Kefauver Hearings. In 1959, FDA Commissioner Larrick stated that it would take a budget increase of $3,418,000 to permit adequate drug inspection. Profits of the 22 major drug firms were $5G2 million in 1958. A decline in drug prices sufficient to cut drug firm revenues by $7,121,000 would of course cut before-tax profits by the same $7,121,000. With a tax rate of 52% in effect at that time, this reduction in pre-tax profits would have cut tax receipts by $3,705,000. The net gain can be roughly measured as the $7,121 000 saved by drug buyers, minus the $3,705,000 in reduced tax receipts, or the required ~3,418,000. Thus If adequate drug in- spection could create confidence in lower-priced drugs to the extent that the resulting competition would lower major drug firm prices by enough to cut total profits by as little as 1.27 percent before taxes, the savings realized would pay for the expanded enforce- ment budgets. If total profits are about 20% of gross revenues, the necessary percentage price cut would be as little as ii of the 1.27 percent, or i4 %. See Hearings of Adminis- tered Price, op. cit., Part 22, p. 12132. PAGENO="0373" COMPETITIVE PROBLEMS IN THE DRUG INDUSTRY 1923 by forming their own wholesale supply agency, on a cooperative or other basis, they caa integrate backward and buy directly from the factory. Hence the in- dependent wholesaler must meet the test of a competitive market, and provide efficient, reasonably-priced services, or find himself by-passed. If all stages of the drug industry were similarly competitive, there would probably be considerably fewer complaints about drug prices. But at the retail level the pressures of competition do not work as beneficially, and more inefficiencies and impediments to the proper allocation of resources are present. Some inefficiencies are probably traceable, at least in part, to certain of the tactics of drug makers designed to maximize profits at the manufacturer's level: the proliferation of branded products, combinations, dosage forms, etc. which increase the druggists inventory costs; the liberal distribution of free samples to physicians, which probably reduces average retailer drug turnover rates; the economically more complex question of the "discriminatorily" low prices made to large-volume non-profit buyers, which again probably reduces drug turnover rates; and possibly certain aspects of policy on returns of unsold or outdated drugs. But certain of these marketing policies are not without costs to the drug companies as well as the druggists, and it is probably unwarranted to impute any primary hostility on the part of the drug makers toward their re- tailers. But certain inefficiencies have also been forced upon druggists by the ac- tions of their own spokesmen and trade associations. The National Association of Retail Druggists, for example, was certainly the prime mover in facilitating adoption of the so-called "fair-trade" laws by state after state in the l93O~s.29 And yet although these laws prevented or greatly limited price competition for trademarked drugs, by enhancing the unit profit margins on these items, more dealers were induced to sell them, and the resulting increase in the number of sellers reduced turnover and earnings on these products. And druggists supported the passage of the Robinson-Patman Act, which prevents them from taking ad- vantage of any possible cost savings available through obtaining supplies on com- petitive bid, and in other ways prevents~ the lower-cost distributor from benefit- tiug commensurately from the potential economies in his operations. From this, one might conclude that not all drug price problems originate at the manufacturer's level. The druggist's markup on the average prescription item is no doubt higher than it might be, but~ then his unit costs are also higher than they might be for many reasons, including those outlined above. Clearly, the druggist has his problems. But the drug buyer has his problems, too. These include: (1) inability to purchase a low-price generic drug if he has been given a prescription for its brand name equivalent; (2) inability to shop around for the lowest available price on a prescription, regardless of its manner of specification, if the medication is needed quickly for treatment of an acute condition, or if the prescription holder is otherwise suffering marked distress pending the securing of his medication; (3) ignorance of the content of the prescription, in many cases, which can simply mean inability to decipher the pre- scriber's jargon, or lack of knowledge of the brand and/or generic name of the drug-either of which may give rise to collateral inabilities, such as (a) inability to determine whether or not a generic prescription was actually filled with a brand name equivalent, and (b) inability to determine whether a generic pre- scription actually filled as written was dispensed at the lowest generic price; (4) buyer ignorance or docility such that he does not even realize that the prescrip- tion form is his own property and does not have to be surrendered to the first pharmacist to whom it is presented-who may be the one whose name is on the prescription pad; (5) the frequently poor prospects for reasonable prices present even for the unusual buyer who does shop around for a low price, due to the tradition of hostility among most druggists toward price competition, and the way in which this tradition is fostered and buttressed by the inhospitable atti- tudes of pharmacy agencies toward price competition and the advertising of prescription prices, by the state "fair trade" laws, by the Robinson-Patman Act, and by still other influences. How can greater efficiency be obtained in the retail distribution of drugs? The characteristics of an efficiently competitive retail drug market can be broadly outlined in a few sentences. All sellers should act quite independently with re- spect to pricing policies; no formal or informal arrangements which would facil- 20 See, the example, ClaIr Wilcox, Public Policies Toward Business, Third EditIon, 1066. pp. 707-710. PAGENO="0374" 1924 COMPETITIVE PROBLEMS IN THE DRUG INDUSTRY itate uniformity of action on prices should exist. Prescriptions should be written so as to facilitate the ability of drug buyers to stimulate price competition among pharmacists. There should be no barriers to the dissemination among buyers of information on the prescription drug prices of individual pharmacies. Buyers should be free to seek out the lowest-price seller, both for the original dispensing of a prescription and for refills. Entry into the retailing of drugs should be free from any artificial barriers, legal or otherwise. Under present market arrangements, there is no real incentive for the druggist to stock low-priced generic drugs. First, there is little demand for them, physi- cians' prescribing habits having been influenced as they have by industry efforts. Second, if the usual two-third markup is added to invoice cost, the unit profit to the pharmacist is proportionately smaller for the lower priced drug. Third, the same logic applies to the wholesaler, so that even if a druggist wishes to stock generic drugs, he may find it hard to obtain them. It is obvious that the substitution of the "professional fee" approach in the place of the uniform percentage markup would make the dispensing of generic drugs relatively more attractive to druggists. But the application of compensa- torily higher percentage markups to the lower-invoice-cost drugs would accom- plish the same purpose. Optimal economic efficiency in the dispensing of drugs would require that relative markups on individual items be determined by price competition among sellers. The markup should be at the minimum rate above cost which is consistent with the retailer's cost of distribution, including a com- petitively-determined rate of return on an appropriate level of investment. If genuine competition exists, the method by which the markup is arrived at will be less important than the amount of the markup, since competition will re- quire that this amount be substantially equal among competing sellers. The notion of adopting a uniform professional fee for any and all prescriptions has drawbacks. It lacks the necessary flexibility in the pricing of services which must exist if price competition is to prevail. And the level of the fee is very important. While I doubt if the size of the fee will be set at too low a level, setting it too high will not insure druggist prosperity. Instead, the high unit profit margin on each prescription will induce new entry into the industry. Many pharmacists now among the ranks of the detailmen w-ill be encouraged to return to pharmacy. As the number of sellers increases, average turnover declines to the point where a balance is achieved between high unit profits and low turnover, and further entry is finally discouraged because of low total profits. In comparing this situa- tion with the low prices and high turnover which would prevail under price competition, it is apparent that competition is to be preferred since prices are lower and excess capacity and investment in underutilized resources is mini- mized, while the profits on investment should be about the same in either case. A few words should be devoted to contrasting druggist retailing of drugs with other drug dispensing media. One can readily understand the unhappiness of retail druggists who pay the full dealer list price when they read about the much lower prices obtained by hospitals and government agencies in response to com- petitive bids. Drug firms have tried to account for such price differences by ref- erences to economies of large scale selling, and to promotionally low prices for the sake of introducing their products to hospital physicians. But the price dif- ferences are clearly too great to be accounted for merely as the equivalent of quantity discounts. And the "promotionally low prices" argument can he dis- missed as a rationalization since it is not characteristic of major drug firms to be so negligent of sales promotion possibilities that the doctor would be likely to overlook a drug if he did not have it on hand in a hospital. The basic reason for the price differences is simply the fact that price competition can often he kindled between brand and generic name drugs and even among major pro- ducers of brand name drugs by means of the competitive bid approach. It has been contended that sales to druggists at high prices "subsidize" the lower price sales to hospitals and public agencies. If this is construed to imply that the latter sales are actually made at a loss, it is no doubt an error. From all evi- dences, drug production costs are very low. And a firm can always add to its total profits by sel1in~ goods at special low prices, provided these prices are above the out-of-pocket costs incurred on the sale, and further provided that these transactions do not affect the prices received on other sales. To the extent that firms have excess capacity, they will be more intensely motivated thus to increase their rate of output and spread the overhead costs of total productive PAGENO="0375" COMPETITIVE PROBLEMS IN THE DRUG INDUSTRY 1925 capacity over a larger volume of production. Differently considered, the seller is simply charging as high a price as he can in two markets which can be sep- arated from each other. Separation exists since hospitals do not resell to drug- gists at higher prices. By doing so, the seller is engaging in price discrimination (in an economic if not legal sense) since he can get much higher prices in one market than in the other because of the different type of demand and supply situations in the two markets. The supply situation differs not because the cost of production differs, but because sellers can be induced to compete in price in one market and not in the other. The difference in the demand situations de- serves more attention. The druggists' demand for drugs is derived from the demand of the individual patient, which is of course very insensitive with regard to price because of the reasons outlined above. Retail druggists can afford to pay high prices for drugs since they can charge even higher prices to their customers without discourag- ing sales. The demand on the part of hospitals and other public agencies is dif- ferently constituted. These are non-profit agencies which operate within a gen- eral budget, and while they do not concern themselves directly with selling in- dividual items in accordance with a market-oriented schedule of charges, they are concerned with lowering total operating costs and staying within budgets. Their purchasing agents may also take pride in the successful exercise of pro- fessional skills through economical buying. But although an interest in eco- nomical purchasing is a necessary condition for obtaining low prices, it is not sufficient unless rival sellers can actually be compelled to compete in terms of price. Other things being equal, retail druggists might have less interest in economical purchasing because they could more readily pass on their high costs to their paying customers. But hospitals with charity patients and a whole host of other exotic financial problems may be less able to forego possible savings from economy in purchasing. Furthermore, hospitals may manufacture some of their own drug needs if they cannot obtain reasonable bids, and this is another factor which makes their demand for drugs less insensitive to price. With demand more responsive to price, hospitals and public agencies can obtain lower prices for many of their requirements, particularly where absence of patent monopoly makes it possible for generic firms to compete. But where patent mono- poly obtrudes, there may be no avenue through which prices may be lowered for any buyer. B. Factors influencing the demand for prescription drugs It is chiefly the nature of demand for prescription drugs which makes the drug industry an inappropriate vehicle for the unregulated exercise of market power by sellers. Demand is so insensitive to prices charged that there is little ex- aggeration in stating that prices have no relation whatsoever to costs. This is contrary to the economics of almost all other industries, where price is broadly determined by the relationship of demand and supply, and where supply is at least directly conditioned or influenced by costs of production. In a purely com- petitive industry the relationship is conceptually precise: market price is deter- mined by the relationship between the supply price of a good, defined as the cost of production of a given rate of output (including in costs the competitive rate of return on necessary investment) and the demand price which the market is willing to pay in order to buy a given volume of output. While supply price depends upon production costs, demand price depends upon the consumer's need for the product and on his income level. Wealthy and needy buyers will con- stitute the highest-price or most price-insensitive segment of demand, while buyers with low incomes and/or slight interest in the product will be potential buyers only at very low prices. But since those who are willing to pay a very high price for the product will also buy it at all lower prices, a reduction in price will increase sales, and while it attracts new buyers into the market, it also benefits buyers by cutting their cOsts of purchasing so that in a very real sense they obtain a "free" increment of "use value" over and above the price which they pay. Since price discrimination is impossible in pure competition, all buyers pay the same price, regardless of their incomes or their relative need for the good. Under monopolistic market circumstances, even assuming that the structure of demand and the costs of production are the same as in competition, the sellers are in a better position to restrict output, limit sales only to those who are most willing and most able to pay high prices, and in this sense charge "what the PAGENO="0376" 1926 COMPETITIVE PROBLEMS IN THE DRUG INDUSTRY traffic will bear." Now as long as the act of purchase is voluntary, no one will literally pay more than what the product is worth to him. In both monopolistic and competitive markets, the market price is equal to the evaluation placed on the good by the least interested buyer who actually does make a purchase. But in pure competition, with no way of mutually restricting output, total produc- tion will generally be large and prices low, while under monopolistic circum- stances, output will be reduced in the interest of enhancing prices. This means that the chances of the typical buyer obtaining a substantial free increment of "use value" above price paid is much less under monopoly. But the major difference between monopoly and competition is in the relationship between price and cost of production. In pure competition the two are identical. But if monop- oly power is present, price is also a function of the elasticity of demand and will exceed production cost in direct proportion to the insensitivity of demand to price. Therefore when spokesmen for the drug interests argue that the consumer is paying no more than the drug is worth to him, the obvious answer is: of course not!! Unless force is used, no one can be induced to pay more for anything than it is worth to him, no matter what sort of fleecing or price-gouging scheme, may be employed. It is not surprising that a drug may be worth more to a sick person than its cost of production, but this does not justify charging more, and in a reasonably competitive market, prices would be much closer to cost than to need-value.50 The above is a general treatment of the contrast between the relationship of demand to prices in monopoly and in competition. In drugs, the argument applies with even more force because of the extreme insensitivity of demand to price, and the way in which this final consumer demand is mediated through the physician. How should the needs of the sick be reflected in the market demand for drugs? Ideally, the total potential market for a drug or group of related drugs is measured by the total need for medication on the part of the individuals afflicted by all the various disorders which are capable of being treated best by the drug or group of drugs. Economically, the total effective amounts demanded at the level of market price may fall short of total physical need in the case of those with low incomes and no access to public care. But effective market demand may also exceed ideal tOtal physical need to the extent that individuals not suffering from those conditions for which the drug or drugs are of use may nevertheless be treated with them. For any given drug it then follows that the actual market is comprised by the total effective demand for medication on the part of all individuals who can be induced to consult physicians, and who are afflicted, or can make it seem convincing that they are afflicted, by those disorders for which doctors may be inclined or persuaded to prescribe the particular drug. The challenge to drug marketers then consists primarily in nersuading physicians, hut also to some extent in spreading the good word to the general public that Brand X can cure symptom Y. And drug marketers doubtless earn their salaries. Ch~nres in the effective demand (ie., prescriptions written and purchased as w-ritten) for individual drugs are brought about by the familiar techniques of direct mail advertising, journal advertising, the dividend of free samples, the financing of symposia, the rental of exhibit space at conventions, and above all the insistence of the ever-present detailman. Although advertising cannot yet manipulate the total incidence of genuine disease, it can readily shift effective demand from one drug to another. And advertising can in a sense actually create demand, even for drugs~ Articles planted in newspapers or magazines may mention the name of a drug alleged to be useful in treating certain conditions, and may thus bring to the attention of more people who suffer (or imagine they suffer) from such ~ Examples where representatives of drug firm interests have defended high drug prices as being no more than what the drug is "worth" to the buyer are legion. The most recent exposure of this Subcommittee to this arrument is embodied in page 7 of Professor Simon Whitney's written statement in behalf of PMA, where he states: "If a $5 prescription, or 6 of them, will keep a patient from losing a couple of days pay or spending a night in a hospital. the price is reasonable." This can scarcely be taken seriously by an econo- mist unless the price is also commensurate with the competitive supply cost of the drugs. But no one who has presented this argument has so far been able to outdo Austin Smith. who mused publicly during the Kefauver Hearings: "I wonder If any member of this Subcommittee knows how much it costs to die? . . . death costs about $900 . . Hence Smith has proved conclusively that any price less than $900 for a handful of pills (a2 in this case) is a bargain since it must be worth at least $900 to the patient to avoid the expense of demise. Hearings on Administered Prices, Part 19, p. 10615. PAGENO="0377" COMPETITIVE PROBLEMS IN THE DRUG INDUSTRY 1927 conditions the fact that specific drug therapy might be purchased. Two of the greatest drawbacks of sales promotion in drugs stem from these characteristics of demand. First, doctors may be oversold on a drug which is then overpre- scribed, often for minor conditions where it can do no good and may cause mischief. Although antibiotics are usually cited in this context, other drugs may also be overused and abuse is compounded when it is administered for chronic, rather than acute, conditions.31 A second drawback, associated with the first, is that patients themselves often insist on unnecessary drug medication. III. APPROPRIATE DIRECTIONS FOR PUBLIC POLICY The difficulties involved in reducing prescription drug prices to reasonable levels can scarcely be exaggerated. There are four major parties to the typical prescription drug transaction: the drug maker, the physician, the druggist, and the patient. The only party with any direct interest in reducing drug prices is the patient, and he has by far the least bargaining power. In fact, at the time the transaction is made, his interest in low prices may usually be quite sub- ordinate to his concern over his health. The major interests of the other parties in drug prices lies in different directions. The drug corporation, whether large or small, has to maximize profits to keep the stockholders happy. The doctor's chief professional interest is in healing, and if he can be made to believe that quality of one drug is superior to that of another, he will be inclined to prescribe it regardless of price. His insensitivity to price is naturally increased by the fact that he does not pay for the medica- tion he prescribes. But the doctor himself is also a business man and may not be entirely unconcerned with maximizing his own net income. It may hell) his reputation if he is always among the first to prescribe all the new drugs, and it may increase his prestige if he prescribes the higher priced drugs. And as an independent businessman in an age of "organization men," he may even admire the buccaneering tactics of the more flamboyant drug firms. Beyond that, he may well own stocks in one or more drug firms.32 But whether he likes this or not, he is vitally affected as a practicing physician by the policies of the AMA, which since has in recent years received over half its revenues from the drug com- panies. Hence under present institutional circumstances, the average doctor has little direct interest in prescribin gthe lower price drugs and is contained within a professional environment which may discourage such tendencies as he may develop in that direction. At the drug retailing level, druggists, like other dealers, resent price coinpeti- tion because although it is a good servant to the consumer, it is a harsh master to the producer. Druggists, however, are somewhat unusual among retailers in that they have been more active and more successful than the others in securing the passage of laws aimed at limiting price competition and protecting the inter- ests of the existing group of competitors at the expense of the vigor of competi- tion. This fact, in conjunction with the closed profession aspect of pharmacy suggests a relatively poor prognosis for the rapid development of price compe- tition at the druggist level. Even so, the awareness that it is desirable to restrict competition, and even the presence of institutional arrangements which might be used to implement this awareness, do not necessarily combine to produce the hoped-for prosperity of the profession. As long as entry is reasonably possible, and unit profit margins high, low turnover and excess capacity are likely to de- velop and cancel out the advantages of high price levels. A satisfactory solution to the problem of high drug prices must await the adop- tion of a series of related reforms which will alter marketing and prescribing ~ Professor Mark Nickerson of the University of Manitoba Medical School reported that the sales of adrenal steroids in the United States and Canada in 1960 was about $250,000,000, and commented: ". . . personally I feel that I am being very liberal when I say that fifty million of that was needed." Report Concerning the Manufacture, Distribu- tion, and Sale of Drugs, Restrictive Trade Practices Commission, Department of Justice, Ottawa, 1963. ~ Dr. Calvin Kunin of the University of Virginia School of Medicine submitted to this Subcommittee for inclusion in the record an article reporting on a survey among medical students, interns, and residents at the University of Virginia Medical Center. Six out of 73 owned stock in one or more drug firms, and 58 of the 67 non-stockholders stated their belief that such stocks were good investments. As these students and residents continue to pursue their careers and increase their affluence, it is likely that many of the favorably disposed non-stockholders will buy stocks. Competitive Problems in the Drug Industry Part 2, p. i34. PAGENO="0378" 1928 COMPETITIVE PROBLEMS IN THE DRUG INDUSTRY practices and bring about price competition at both the manufacturing and retail levels. The elements of such a solution would include: (1) Abolition of drug product patents and compulsory licensing of drug process patents at reasonable royalty rates. (2) Outlawing of brand names for drugs, with the requirement that drugs be identified and hence advertised and sold only by use of the generic name(s) of the active ingredient(s) in conjunction with the company name of the seller. (3)Provision of FDA with sufficient authority, staff, and funds to permit it to carry out a drug inspection program adequate not only to prevent the sale of substandard drugs but also the plausible insinuation of the possibility that substandard drugs might be on the market. (4) Elimination of unnecessary barriers to entry of new drug firms into the industry. If a drug has been cleared for marketing as the result of adequate data compiled by one applicant, the same drug should be approved for marketing by any firm capable of producing the identical drug. Unneces- sarily burdensome requirements by way of conducting studies which merely duplicate existing studies should not be imposed. In this regard, the sug- gestion of FDA Commissioner Goddard before this Subcommittee that such drug data submitted to FDA be made a part of the public record is an excellent one. (5) Provision of the medical profession with more accurate, systematic, and objective drug data. If the price competition injected into the industry as a result of reforms succeeds in reducing profit margins and eliminating the detailman and if the medical profession does not then respond by subscribing adequately to independent newsletters, the provision of a publicly sponsored newsletter, similar to the Prescribers Journal in England may become necessary. (6) Exertion of every feasible effort to infuse more price competition into drug retailing. Serious consideration should be given, at all relevant levels of government, to the liberalizing of the requirements for operating drug- stores and dispensing prescriptions, so that the further development of lower- priced outlets such as discount pharmacies and drug mail order houses can be stimulated. (7) Recognition of the possibility that even the above reforms may not be sufficient to reduce drug prices. If after a reasonable period of time, prices have not declined sufficiently, consideration should be given to such addi- tional reforms as (a) compulsory licensing of imports of patented drugs: (b) complete abolition of drug patents; and (c) price control or public utility regulation. The interrelationships of these recommendations may briefly be summarized. As has been ephasized by other witnesses, the absolute nature of the drug patent privilege in this country is paralleled only in Panama and Belgium. All other countries with drug patent laws provide either for the denial of drug product patents, for compulsory licensing under certain circumstances, or for both. The abolition of product patents and the making available of compulsory licenses on patented drug production processes will increase the number of firms, both large and small, making and selling each type of drug. This will stimulate price competition, particularly since the small firms will naturally be selling at low prices in order to counter the initial advantage of the highly advertised brands. But the limitation of promotion to generic name plus drug company name w-ill reduce the relative appeal of the major firm's drugs in the market, and this. coupled with the cancellation of disparagment efforts by adequate FDA inspec- tion. will make sales promotion efforts less differentially profitable. And since production costs are low, price competition between large and small firms will greatly reduce unit profit margins and in time will reduce the ability of major firms to engage in sales promotion contests among themselves. By such means. drug prices might in time be very substantially reduced. Drug firm spokesmen claim that even if all profits were eliminated, prices would not be cut by more than 15 or 20 per cent.33 But this overlooks the amounts spent on ~ Dr. Harold Burrows, for exn male. testified before this Committee to this effect: "If Parke. Davis. for our 1966 year. had reduced our prices by 20.5 percent. we would not have made any money . . . This is the maximum marrin that we are talking about Compel it~re Problems in the Drug Industry, op. cit., Part 2, p. 812. PAGENO="0379" COMPETITIVE PROBLEMS IN THE DRUG INDUSTRY 1929 1)1omotion, the costs of excess capacity, and the like. In an efficiently competitive drug industry, sales outlays could probably be cut by something like 90 per cent, and excess capacity costs would be substantially eliminated. It is premature to speculate in regard to the possible magnitude of price reductions, but bringing prices more in line with production costs would reduce them by more than half. The major firms would have a certain amount of time in which to adjust to changed circumstances, owing chiefly to the backlog of "goodwill" built up in their behalf by intensive selling efforts. Although it has been argued many times that the physician must be free to write brand name prescriptions because he only trusts one particular company's product, the use of the brand name rather obscures the identification of both the company and the drug. The use of generic name plus the company name will still allow the doctor to specify the maker of the drug. It is likely that much of the appeal of brand name prescribing lies in its convenience; if it is made less convenient to the doctor by requiring that he specify the company as well as the drug, it is probable that more purely generic prescribing will result. If these factors are operative, plus an increased reliance on purely generic drugs due to confidence in the adequacy of FDA inspection, then in time the great initial advantage of the major firms in terms of their "ethical" image will be dissipated and they will tend to lose their favored position. It is not certain that the major firms' physical sales volunie will decline, although the unit profit on such sales will certainly fall. But the impact should be sufficiently gradual to allow major firms to diversify out of drugs and into other areas, such as luxury goods, where the marketing methods which the managements have per- fected at such cost can be applied in ways less mischievous to society. As to the impact on research, there will always be a place in the industry for the firm which engaged in basic, and in the long: run, truly productive research. Unfortu- nately, there seem to be fewer of these, firms in the industry than one has been led to believe. It is hard to attribute credit properly because to do so requires the judicious deflation of the barrage of outrageous assertion surrounding each firm's own public estimation of its research accomplishments, rather like trying to find out which Hollywood spectacular really is the most Super-Colossal. One firm, Merck, does rather stand out, if only because of the credentials of the Nobel Prize winning "character witnesses" (the phrase is Dr. Louis Lasagna's) it has been able to summon in its defense. While a firm which is interested in truly fundamental research may not earn extremely high returns on the funds it invests in such research, in the long run it is probably more likely to survive. Without doubt, the greatest single obstacle to the securing of drug patent reforms has been the argument that drug research would suffer. The issue should be faced head on. Would a reduction in the expendi- tures which the firms classify under the heading of research necessarily be detri- mental to public health? This depends upon the types of research outlays which are reduced, and whether or not any possible decline in productive private firm drug research might be offset by increases in productive drug research undertaken under other auspices. After drug law reforms, the level of drug firm research expenditures may be reduced except in those firms where research is permitted to be pursued in large part for its ownsake. But it is in the environment created within such firms that research is likely to be most beneficial in the long run. On the other hand, research of the "copyshop" type is likely to dwindle, but this is a gain to the extent that such research typically produced less of genuine social value than it consumes in terms of the alternative uses of the human resources employed, even-or perhaps a particularly-in such an operation as Pfizer, where molecular manipulation reportedly attained the status of a true art.34 Even if total drug industry research spending does decline, professional person- nel resources will probably be shifted into non-profit channels. It can be argued that a ma3or diversion of pharmaceutical research endeavor from private firms to public, university and foundation channels will in due time result in equally major gains. Private firms appear to carry on relatively little fundamental re- search, and more of this is needed at present to make applied research eventually more productive. Non-profit research will also mean less waste of very scarce human resources in imitative and duplicative programs, and in marketing- oriented activities masguerading as research. Silberman, "Drugs: The Pace~ Is Getting Gurious," Fortune, May 1960, pp. PAGENO="0380" 1930 COMPETITIVE PROBLEMS IN THE DRUG INDUSTRY The possibility remains, however, that drug manufacturers' prices might be reduced by competition and yet these reductions might not be passed on very efficiently by retailers. To increase the likelihood of competitive performance on the part of druggists, the appropriate levels of government should review the requirements for entry into drug retailing and compare them with the re- quirements for satisfactory performance of drug merchandising services under current conditions. Clearly, the pharmacist's function has changed from that of actively compounding prescriptions to those of passive merchandising; many of his skills have become obsolete as the result of the revolution in the drug industry. Those interested in drug industry reform should join in urging the maximum practicable liberalization of the traditional restrictions limiting entry into drug retailing. This liberalization should be such as to constitute recogni- tion that the traditional pharmacists' distinctive functions are being altered away from professional competence in compounding and toward skillfulness in merchandising. Such recognition would be likely to stimulate new entry by those not traditionally opposed to price competition. In many lines of retail trade, dealers were inertia-bound and distribution methods unprogressive until price competition developed from such sources as chain stores, supermarkets, and mail order houses. In my own state of Texas, where we have never been burdened by resale price maintenance laws, discount pharmacies and the drug departments of large dis- count houses are doing a thriving business, and have not only given the price- conscious drug buyer an alternative source of supply of both brand and generic drugs, but have exerted downward pressure on the margins of traditional drug- gists. Similar competition is rendered much more difficult in those states in which fair trade laws are enforceable, but one of the advantages which drug nomenclature reform should possess is in eliminating trademark names for drugs and thus making them ineligible for the protection against price com- petition which the fair-trade laws now allow. Some modifications of the other laws limiting price competition, such as certain provisions of the Robinson- Patman Act, should be also accomplished so as to allow druggists at least the possibility of soliciting competitive bids from suppliers. All of these reforms represent movements in the direction of creating a market framework within which a freely competitive privately-owned industry can efficiently operate. If these reforms do not prove sufficient to bring about the desired result, two further measures would then become relevant. First, the patent privilege for drugs might be further modified by allowing the importation of patented drugs from abroad if the dealer could more cheaply purchase them abroad than produce them domestically. This would fall short of patent abolition in the sense that the U.S. patent holder would still be able to collect a reason- able ad valorem royalty on sales by the importer. (Naturally, the quality of the imported drugs would have to be acceptable, and the importer might perhaps be required to pay for the costs of FDA inspection at the port of entry.) If this did not prove sufficient drug patents could be completely abolished. And only if drug price levels still prove impervious to reduction after all these reforms should such measures as price control or comprehensive public utility regulation of the drug industry be imposed. These latter remedies are likely to be less effi- cient in operation because of the absence of a competitive market criterion for prices of drugs under price control, and the general unsuitability of the drug industry as the subject of regulation of the conventional public utility variety. Hence these expedients should be regarded as last resorts, to be used only after every effort to inject price competition has been exerted. Dr. STEELE. Prof. Paul Cootner, professor of finance at MIT, made a statement in which he discussed risk and rate of return in very gen- eral terms and made little reference to the particular economic situa- tion of the drug industry. I. STATEMENT OF PROF. PAUL COOTNER (A) In his presentation, Professor Cootner makes each of these statements: 1. First, he admitted quite candidly, and I quote: Now, I do not appear here as an expert on the drug industry, either with re- gard to its pricing policy or the riskiness of its investments (p. 1). PAGENO="0381" COMPETITIVE PROBLEMS IN THE DRUG INDUSTRY 1931 2. And again I quote: * * * Neither Schumpeter nor I, nor indeed any responsible economist, will argue that industrial abuses should not be corrected, when found (p. 4). 3. Finally, what I think is a very significant statement on page 9: * * If one were to decide, by administrative action, to reduce the average rate of return in a risky, competitive industry without at the same time reducing the risks, we would find an immediate impact on that industry's investment policy (p. 9). (B) In general, I think that Professor Cootner's paper is a sound pedagogical exercise which in commonsense terms conveys some of the major subjective factors which influence an investor's frame of mind in appraising investment prospects. In fact, I find myself in agreement with virtually every statement he makes. But I particularly agree with the three statements quoted above. Statement No. 3 is sound economic theory. Statement No. 2 is also true, in fact a truism if one defines responsibility in an economist in terms of sensitivity to industrial abuses. And if one agrees with statement No. 2, surely one must agree with statement No. 1, since Professor Cootner neither suggests that he is aware of drug industry abuses nor proposes corrections for them. (C) The consequences of Professor Cootner's admitted lack of ex- pertise on drug industry economics is that his paper, although educa- tional in a general sense, is misleading because it suggests that the industry is like any other industry in that high returns are likely to be associated with high risk, that the drug industry's aggressiveness in anticipating demand and "promptness in accepting innovation and change" (p. 10) is socially benefièial, and that if industry risks are reduced it will lead to a reduction in the net social productivity pre- sumably due to these risky investments. But things don't seem to work this way in the drug industry. Con- sequently, the net effect of Cootner's paper is misleading because he says all the favorable things about the productivity of risk-taking in industry generally, without elucidating any of the drawbacks of the policies which result in high profits and hence in allegedly high risks in drugs. I would like to emphasize that: 1. Statement No. 3 does not really refer to drugs since it specifies "a risky, competitive industry" while drugs are a profitable and rivalrous industry, not too much troubled by true price competition. Among price-competitive industries, one can expect the average profit levels of firms showing positive profits in risky industries to exceed those shown by similar firms in safer industries. But once we drop the assumption of price competition, there is no such clear-cut relationship. A pure monopolist of an absolute necessity could make enormous profits in perpetuity and face no risks. But in the drug market there are elements of both monopoly and rivalry. Patents confer monopoly power with respect to a certain product and extremely inelastic demand allows enormous unit profit margins. But these generous margins will attract new entrants who will find it profitable to spend vast sums in imitating the patented product legally. Once a rival com- pound is concocted, how can the new drug take sales from the old? Price competition is one route but a very costly one, and unless there are a large number of rivals, it is not likely to break out. Instead, PAGENO="0382" 1932 COMPETITIVE PROBLEMS IN THE DRUG INDUSTRY sales promotion outlays will be the vehicle of the rivalry, and the origi- nally enormous profit margins are whittled down progressively by the necessity of engaging in increasingly extravagant sales promotion campaigns to counter those of one's rivals. Senator NELSON. Is the ordinary consumer of drugs benefited in any way by this rivalry? Dr. STEELE. No, I would say the benefit is essentially negative in that rivalry functions in the drug market as a substitute for price competi- tion on the one hand and genuine research advances on the other. When Dr. Console, former medical director of the Squibb Co., testified before the Kefauver hearings, he said the drug industry is an un- usually safe industry. Risks are low because if the research department fails to make an advance, the advertising department can substitute its expertise and make the drug appear to be an advance. So in this sense, this rivalry prevents price competition by the ways which Professor Schifrin has just described. And it also confuses the issue regarding drug information and the substantive advantages, if any, made by new drugs. Senator NELSON. Thank you. Dr. STEELE. In order to make the greatest profits per drug it is usually necessary to be first in the market, otherwise the advertising cost of wresting the market away from the first (and also heavily advertised) drug is disproportionately great. Hence the motivation to devise new drugs. But at the same time, the new drugs found by others must be rapidly copied, so that the costs of research, both primary and imitative, come to mount up. And the fact that every- one is trying to copy and/or improve everyone else's drugs leads to an overly rapid rate of product obsolescence and an artificially in- duced "risk" of short commercial life for the average product. Since many doctors have testified that there is generally no net surplus of advantages over disadvantages for the manipulated mole- cules, rapid changes in market shares betray motion but not progress. Thus sales promotion and product competition substitute for price competition, and unit profit margins decline not through price reduc- tions but through cost increases. It might even be contended that this route to profit erosion roughly equates risks and rates of return, such that even though the risks are self-created by the seller's own choice of rivalry tactics, they are real risks and only enough will be invested in sales promotion and patent bypassing research to keep the rate of return on investment from dropping below the minimum satisfactory rate relative to the artificial risks built into the market. One might agree with this analysis and still contend that it would be socially beneficial to alter the market structure so that price com- petition would be forced upon the sellers and lowered profit margins would become insufficient to support constant devising of new drugs and their rapid copying. Hence the "risks" would decline in direct proportion to the decline in rate of return. But it is also possible to dissent from this analysis. First, inherent commercial risks in the drug industry are probably lower than in most industries because of the depression-proof character of the industry. Drugs are needed in fair times and foul, and a sick man will buy as many drugs as lie needs regardless of his income, right up to the limit of his ability to ffnance current purchases. Second, drug firms may have come to realize in recent years that the market is practically satu- PAGENO="0383" COMPETITIVE PROBLEMS IN THE DRUG INDUSTRY 1933 rated with advertising appeals, such that although it is necessary for each firm to maintain current sales promotion levels in order to offset the efforts of others, incremental advertising expenditures will have no real impact; therefore, profit rates remain above risk-determined levels because additional spending on sales promotion, and possibly even on production competition, simply doesn't lead to even the pros- pect of further differential gains. In such a case, the only remaining tactic to bring profit rates down to equilibrium levels consistent with risk (regardless of the nature of the risk, whether inherent or artifi- cial) is price competition, and drug industry repugnance to this force, plus the fact that the number of sellers in each market; that is, effective competitors is still too small to compel price competition, generally re- suits in a sort of high-profits truce between companies which refuse to engage in the price competition which would effectively reduce profit levels to equilibrium relationships in comparison with true risk. (How- ever, Mr. George Squibb's testimony with regard to the frequency of special deals to distributors indicates that although list price com- petition may be unknown, there may be a substantial amount of hidden competition in terms of such special inducements to dealers. And I sus- pect that excess capacity in production processes is probably a major reason for this.) Mr. GORDON. Are you aware that one of the risks attributed to the drug industry is the possibility that the industry may become competitive? Dr. STEELE. I think this is true. I think Mr. Squibb also alluded to this risk in his statement. Senator NELSON. Well, in economic terms, is competition considered a risk? Dr. STEELE. Well, I think it is really a pleasant risk. Competition would not be a risk in itself. In comparing two equally competitive in- dustries, one would expect that the~rate of return would be greater in the inherently more risky industry.~ But if an industry, let us say, had been monopolized, and there is a likelihood that for one reason or another, competition will break out, then investors will see this as a risk in the sense of the institutional frame of risk or something like that, but it won't be something that is inherently in the market struc- ture of ordinarily competing sellers. (2) Professor Cootner does formally recognize the possibility that the firms might react to profit reduction measures by reducing the risk- iness of the ventures in which they engage. But the relevance to the drug industry is limited since the example mentioned assumes that the risk taking activities hypotheticali curtailed are socially productive, and do not consider the possibility that these risk taking activities are also creating. (3) Cootner makes two statements which are not quite compatible. On page 8 he states: "This basic conclusion is that as risk rises so does the required rate of return." But the "required" rate is a subjective expectation; it is basically an ex ante phenomenon. However, on page 4, Cootner states that "one should not be surprised to find large average profits in risky enterprises * * ~ which is not quite the same thing. One would expect to find large average profits for com- panies making profits, but if true risk is significant, also large average PAGENO="0384" 1934 COMPETITIVE PROBLEMS IN THE DRUG INDUSTRY losses for companies failing to break even. I would say petroleum exploration would be a good example of this. If you take the com- panies which only show profits and average their returns, the indus- try looks better than if you take companies which incur losses. In such an industry, if you take their returns and add them in, it is not clear that the average ex post rate of return on all investment in a risky industry should always necessarily be higher than in a safe industry where losses are very rare. This is, of course, one of the maj or weaknesses of statistical measurement of the risk rate of return hypothesis: the data are usually available only for the largest firms in an industry, which because of their very size face relatively less risk than small firms (since riskiness is basically a property of individual investment projects, and not of the firm itself, which is collection of many projects) the small ventures which fail, or do not get included in the blue ribbon listing of firms, may so regularly make losses that total industry rate of return is significantly depressed. II. STATEMENT OF PROF. SIMON WHITNEY (A) While I am in general agreement with almost all of Professor Cootner's statements, I find myself in strong disagreement with the majority of Professor Whitney's. Time does not permit the criticism of more than a limited number of these statements. (1) Whitney's statement as a whole is based on the implicit assump- tion that if private drug firms don't do drug research, it just won't be done. But great sums of public money are now spent on research in health and medicine. Even if drug industry reforms do reduce private drug research, there is reason to believe that the researchers could be more productively employed by universities, foundations, and government agencies in doing much the same type of research. I would like to comment now on the statement of Prof. Simon Whitney and preface my comment by saying that I am in general agreement with almost all of Professor Cootner's statements. The only thing wrong with them is they are rather irrelevant. But I find myself in strong disagreement- Mr. GORDON. You say you are in general agreement. Yesterday, Dr. Mueller of the FTC, stated: Thus Conrad and Plotkin unwittingly have made a case for the inference that a substantial part of the high profits earned by drug companies are really due to advertising-and-promotion-created barriers to entry, rather than risk. This, of course, coincides with the conclusion of nearly every economist who has studied the drug industry. Dr. STEELE. I would agree with that. I would regard myself as one of those who agree with the statement. But the Cootner statement is different from the Conrad and Plotkin statement. The Cootner statement says, really, very little about the drug industry. just discusses the problems of risk and so on. Mr. GoRDoN. I see. He doesn't discuss the drug industry at all. Dr. STEELE. Not really. Whitney's statement as a whole- Senator NELSON. Doctor, has anyone done a study to find out how much research has been done by Government in the health field-by PAGENO="0385" COMPETITIVE PROBLEMS IN THE DRUG INDUSTRY 1935 NIH and other agencies-that is really directly related to the drug field? WTe know what the total research budgets are. We have received some material from time to time indicating expenditures by Federal agencies of hundreds of thousands or millions of dollars, part of which was spent in research on a particular drug or testing of a drug, but the amount spent on the drug alone was not clearly isolated so you could specifically identify it. Are you aware of any studies, reports, by anyone who has attempted to identify the amount of money that goes into research that you would say is related rather directly to the drug industry? Or to put it another way, research of the kind that the drug indus- try does? I am not talking about molecular manipulation or anything like that. Are there any studies of that kind? Dr. STEELE. Unfortunately, I do not know of any. All the ones I know of share the defect you have just described. They refer to pro- ject research. Only a small proportion of the project may be devoted to evaluating the particular drug. Senator NELSON. We intend at some stage to have testimony from the various Government agencies on this but I though perhaps some- body had done a study and made some breakout of the funds spent in this area. You are not aware or any such study? Dr. STEELE. No, I am afraid I am not. On page 3, Professor Whitney states that: "* * * $3 of additional stockholders money per dollar of net worth in 1950 went into drug manufacturing for every $1 in all manufacturing * * Stated in this manner, it sounds as if new stock issues were sold to raise the three additional dollars. Actually, as Whitney explains in a footnote (without appearing to appreciate the significance of this factor), most of the increase came about from reinvestment of earnings. Professor Whitney is trying to have the best of two worlds: the Adam Smith world of atomistic com- petition, a.nd the modern corporate finance world of noncompetitively high prices and profits, and discretionary plowbacks of retained earn- ings by management. On pages 2 and 3 Whitney implies that the first function of profit is the allocation of investment, which is true in the classical sense. But the classical theory of consumer sovereignty ap- plied also to investors; in principle, all profits in a corporation, under this theory, should be paid out to shareholders, and the decision as to whether to reinvest in the same enterprise or in others should be made anew upon the receipt of each dividend. This would insure a more im- partial appraisal of alternative uses of dividends; whether for rein- vestment or for consumption spending, than does the modern system of having managements, rather than owners, decide upon the reinvest- ment of about half of total earnings. (The obvious tax disadvantages of higher dividends to high-income stockholders and underwriting costs for new securities have been instrumental in transferring discre- tion over total profits from stockholders to managers, but this does not alter the basic principle involved.) 81-280-68-pt. 5-25 PAGENO="0386" 1936 COMPETITIVE PROBLEMS IN THE DRUG INDUSTRY And with price competition, relative profit margins in different industries would reflect the scarcity of supply relative to demand and hence indicate the relative need for expansion of investment and pro- ductive capacity. But if monopoly power exists, output may be restricted and prices and profits can be at high levels although the monopolist may actually have excess capacity and no need to expand his facilities. Furthermore, discretion over prices may result in prices which are too high relative to costs and in resource misallocation relative to the outcome which would have prevailed in a purely competitive market. If excessive profits, made by overcharging buyers, are plowed back into the industry by bypassing the investor's discretionary power over all profits, the investor benefits from what is tantamount. to a capital levy on the consumer. The drug buyer thus contributes much of the capital-the great majority of additions to capital investment- on the basis of which the drug stockholders now expect high earnings because of the "risk" to which "their" capital is subjected. Further- more, Whitney takes it for granted that this increase in capital value resulted in at least commensurate social gains; on page 4 he identifies these gains with increases in drug sales and with research and develop- inent expenditures. But in neither instance is it necessarily the case that the true value of drugs or drug research is measured by dollars spent. The individu- als most qualified to judge these matters are physicians and medical educators, and their judgments as recorded in public hearings on drugs have not been such as to encourage those who wish to equate dollars spent and value received in drugs. (3) "Many hundreds of new drugs, as documented by earlier PMA witnesses, resulted from this profit-motivated research" (p. 4). This is misleading if the reader naively interprets this to mean genuinely new chemical entit~es. Non-PMA witnesses supply different "docu- mentation." Dr. Martin Cherkasky has previously stated before this committee that the industry's claim in the early 1960's to have PI~- duceci over 400 new drugs required more than 90 percent deflation. He said: On examination, only 29 of those were really new contributions. The rest of them were qimmicks, new dosages, new combinations that really hadn't much value ("Competitive Problems in the Drug Industry," pt. 2, p. GiG). And during fiscal 1967, FDA Commissioner Goddard stated that of the 83 New Drug Applications approved, 62 were for "what have been called `me too's' or molecular manipulation." (Ibid., pt. 2, pp. 757- 758.) (4) On page 6, Whitney suggests that the lower prices charged by generic name firms reflect the absence of research, quality control, and original distribution costs on their part. This sort of approach is a favorite with pharmaceutical manufacturers associations. The Cana- dmn P~iAC made similar charges against Canadian generic name firms during their drug hearings for 1966-67, and, not unsurprisingly, it developed that small generic firms also incurred costs for quality con- trol and research. The chief cost savings of the generic firm is in the area of sales promotion, which was somehow overlooked by Whitney-unless PAGENO="0387" COMPETITIVE PROBLEMS IN THE DRUG INDUSTRY 1937 original distribution" is taken to be a very inclusive euphemism indeed. In any event, the major firms are simply voicing their indignation about having to be faced with price competition even if it is only effective in the minute interstices of their total market. I would refer again to Professor Schifrin's diagram, the roughly 5 percent of total sales by dollar volume, which would be a relatively higher percentage of the physical volume of drugs sold because of lower average cost. What these firms do not realize is that any argument against price competition is an argument for price, control. It is a long-established principle of public policy that where competition will not work, regulation must be substituted. "Do the lower prices charged by gen- eric name manufacturers reflect willingness to operate on lower profit margins?" (p. 6). By no means. Price competition, such as prevails among generic firms, compels produëers to be efficient. and to accept the low-profit margins determined by competition. But the major companies would ordinarily have little direct experience wit.h this. (5) "What does it mean, after all, to say that. a. price is `too high' ?" (p. 7). Any trained economist would answer that it is too high rela- tive to production cost, since this is the standard of efficient pricing in a competitive market. But Whitney invokes a sort of value-of- service standard in connection with prescription drug prices, implying that if the price does not exceed what it is worth to the buyer, then it is reasonable. But this is just the monopolistic practice of charging what the traffic will bear; the full value to the consumer is the absolute minimum price which can be exacted from him, and the major benefit of price competition to the consumer is that it allows him to obtain the goods at a price related to production cost as well as to demand, such that the typical buyer pays less for the goods than its actual maximum value to him, and thus enjoys some measure of what economists call "consumer surplus." Under monopoly pricing, the average value of this surplus will be reduced, and under systematic price discrimination-the exercise of which requires considerable monopoly power-it may disappear' entirely. (6) "You have heard of large economies made by hospitals through purchases of drugs by generic name. Were all purchasers to do the same, many research-based companies would be put into serious. straits" (p. 8). The implication here is that these companies would cease to do research. But since they spend only about 6 percent of' their sales dollar on research and 25 percent on sales promotion, it would seem that much greater scope for economies lie in the market- ing budget. But even if research outlays were cut, much of this research could probably be more efficiently accomplished under nonprofit auspices,. as mentioned,under point 1. And a more equitable distribution of the cost of drug research might also be accomplished if more of it were publicly financed. Although drug spokesmen have defended the price of drugs by reference to research costs on innumerable occasions, I have never once heard of them raising the question of the propriety PAGENO="0388" 1938 COMPETITIVE PROBLEMS IN THE DRUG INDUSTRY of having these costs financed entirely by the sick. And this is a ques- tion which deserves much consideration. Senator NELSON. You suggest that the more equitable distribution of the cost of drug research might also be accomplished if more were cofinanced. Would you elaborate on that? How and in what way? Dr. STEELE. Well, the argument is that drug research, particularly fundamental research in drugs is ultimately a philanthropic activity which benefits everybody. You might say in a narrow sense, if a person is sick, he benefits and benefits exclusively from taking a drug which he has prescribed for him and which he pays for. However. the narrow view is overly narrow, in that availability of the drug protects the health of everyone. To the extent that individuals who have diseases prevent the disease from continuing and affecting others, the society as a whole will gain. The availability of good medicatioi~ increases the health standards, the health potential, of the economy as a whole whether or not the medication has to be used. It is a potential benefit. If the sick benefit entirely, then the people who by virtue of the sick's having obtained medication do not themselves become affected. So they benefit in a sense almost unjustifiably at the expense of those who have had to pay for medication and because of their sickness, have been in perhaps a worse position than others to pay for the medi- cation and hence for research. Senator NELSON. What you are saying is that since it is a benefit to the whole public, the individual and the public who never becomes ill, it is in the nature of an insurance benefit. You may own insurance, never have an accident, never become ill, but it is a protection if you do. Is that what you are saying? Dr. STEELE. Yes, this is it. Senator NELSON. How would you finance more of this research? In what way? WTould you contract with companies to do it? Dr. STEELE. You could do that provided that the benefits to society were not disproportionately appropriated by the companies involved. I think that primarily, basic research is an activity which is not done too efficiently or effectively by profit-oriented firms. Obviously, you cannot tell what is going to happen when you start on a basic research program. You want to increase your knowledge. You may end up bene- fitting your competitor rather than yourself. Research may be a fruitless activity for years, and the loss a com- pany may incur on this basic research may deter it from doing further research. So I think research either under public foundations, uni- versities and so on, or public financing of private research, with ap- propriate restrictions on patent monopoly could be carried out more effectively. Senator NELSON. Thank you. Mr. GORDON. Is not this the principle of roadbuilding, of building highways? The Government pays for highways throughout the coun- try and in States as well. You do not expect the users of the roads to pay for them. Do we not consider that all of the society benefits by it through the opening of communication and so forth? Is that not the same idea? PAGENO="0389" COMPETITIVE PROBLEMS IN THE DRUG INDUSTRY 1939 Dr. STEELE. There is a very good analogy. But I think that in the case of drugs, the argument is even stronger, because a person who buys a car does so voluntarily. He engages in the use of roads in order to further his own gains. But a person who buys a drug is not doing something voluntarily in the sense that he has told society to use its resources specifically to cure his disease. Instead, he is reacting to a situation which has been forced upon him against his will. What Senator Nelson said, using the insurance in case of a risk, rather than using a voluntary activity. The next point is a rather technical one. Professor Whitney is in error when he multiplies profits per dollar of sales in drugs by the ratio of profits per dollar of equity capital in drugs to profits per dol- lar of equity capital in all manufacturing and claims that this brings profits per dollar of equity capital in drugs down to a level equal to that earned in all manufacturing. The fallacy is that he is assuming a capital turnover ratio of unity for the drug industry. Profits on equity capital are equal to the prod- uct of two terms: Profits per dollor of sales and sales per dollar of capital during some time period suèh as a year. In manufacturing, the product of these two ratios is said by Whitney to be only 61 per- cent as high as for drugs. But if this ratio is applied to one of the two terms, rather than the product of both of them, in the drug industry, the implicit assumption is that the value of the other term is unity. But Whitney cites the FTC-SEC survey as showing that the drug industry earned 19.7 percent on sales but 22.2 percent on investment, hence the implicit capital turnover ratio is about 11/s. Hence prices could be reduced by 8.8 cents in the dollar, instead of 7.7. (8) "Someone will have to make up the corporate income tax pay- ments lost as a result of declining profits of drug manufacturers" (p. 9). In the first place, the Government would save 8.8 percent on all its drug purchases at the manufacturer's level; this would go a good way toward offsetting the lower income tax receipts from drug firms. And to the extent that the Government indirectly finances other drug purchases, the cost savings would be still greater. But the argu- ment is a peculiar one. There is the suggestion that one should not lower excessive profits because the profit maker is liable for Federal income taxes on the full amount of the profits. But the same is true for incomes from fraud, embezzlement, extortion, and other illegal activities, and yet the possible adverse effect on income tax liabilities does not deter policymakers from trying to eliminate such activities. The next point I think is certainly i~nportant. Whitney states: Certain trends are alarming. New chemical entities marketed per $100 million of R. & D. expenditures, for 1959 through 1966, were 32, 22, 21, 11, 6, 6, 7 and 3, respectively, * There may be no real recovery if the profits from research are threatened (pp. 9-10). This is not alarming if it simply means that tighter FDA regula- tions have reduced the total flow of ~mew drugs by eliminating the in- efficacious ones, which FDA did not have authority to do until 1962. But the fact that the decline wa~ steady even before 1962 probably indicates, in addition, that diminishing returns to the molecular engi- neering techniques of applied research are being increasingly felt as that method is more and more intensively and extensively applied to PAGENO="0390" 1940 CO~ETITIVE PROBLEMS IN THE DRUG INDUSTRY a fixed universe of discovery possibilities. What is needed is money spent on fundamental research, to make the major breakthroughs for the applied researchers in the drug firm subsequently to exploit. Since profits rose throughout the period of declining discoveries, there is no reason for Whitney to suggest a linkage between recovery in discovery rate and the lifting of the threatening atmosphere regarding profits from research. Actually, profits have little to do with it since the drug firms do not find it attractive to invest enough funds in the fundamen- tal researches which constitute the only path to eventually reestablish- ing the productivity of applied research. (10) On page 10, Whitney notes that drugs as a percent of total medical care costs have been declining. This is due not to declining drug costs but chiefly to the phenomenal increase in hospital costs. And most of this increase is owing to the inclusion of hospital charges in health insurance programs without adequately disciplining hospital costs and charges. Health insurance programs increase effective demand for hospital services without at the same time taking any steps to increase their supply. The natural result is an increase in charges. And if drug cover- age increases under health insurance plans, including medicare, the same thing is likely to happen in drugs. So the fact that drug costs have increased less rapidly than hospital costs does not mean that drug pricing has been characterized by self-restraint but simply that the coverage of drug costs under health insurance programs has been less comprehensive than their coverage of hospital costs. As pressures build up to increase drug coverage, steps must be taken to discipline drug prices and increase the supplies available at competitive prices. Senator NELSON. Have some of the declines in drug prices been due to the expiration of patents? Dr. STEELE. Undoubtedly, this is true. Senator NELSON. That is, on a particular drug, as the patent expires, there is a tendency for the price to consumers to go down? Dr. STEELE. This is true, yes, especially in Professor Schifrin's mar- ket "B." Some of the major drug firms may sell at lower prices. The patent holder himself may not reduce his own price, or may reduce it only more slowly. Next I would like to make some comments on the Arthur D. Little report.1 III. COMMENTS ON THE A. D. LITTLE REPORT. "RISK AND RETURN IN AMER- ICAN INDUSTRY" A. Criticisms of the basic orientation of the study (1) The study is designed to test the hypothesis that there is a posi- tive relationship between risk and rate of return. But the definition of risk simply as average intercompany variance within an industry is rather arbitrary, even though this has frequently been used as a meas- ure of risk. "Average Intraindustry Variance and Return in American Industry" would have been a more accurate title for this study. I am indebted to my colleague Prof. James Willis for suggestions and comments In regard to this presentation. PAGENO="0391" COMPETITIVE PROBLEMS IN THE DRUG INDUSTRY 1941 (2) It is misleading to include the drug industry in a comparative study where a large number of industries are analyzed on the basis of measured average rates of return and variance, and conclusions are drawn as to the relative risks encountered by the groups of industries as a collection of observations. This is because the markets of the drug industry are so protected from price competition by patents, trade- marks, extremely inelastic demand, brand name prescribing, the mediation of effective demand through a financially irresponsible pur- chasing agent, etc., that there is no other industry in which the seller has a comparably great power to rise superior to price competition. The drug industry is truly in a category by itself in this respect. Ideally, one should hold constant the degree of competition among in- dustries in making a study of this sort. The relationship between risk and rate of return should be most clear cut if all industries studied were purely competitive. The value of the study is reduced to the ex- tent that industries far removed from price competition are included. After all, a pure monopolist could conceivably earn enormous returns with no risk. (3) It is difficult to understand why the investor is conceived of as measuring risk in terms of variance alone, without regard to the aver- age rate of return in the industry. It seems likely that an investor will regard two industries as having different risks, if one has an average rate of return of 25 percent and the other has an average return of 5 percent, even though each has a variance of 50, for example. B. Specific criticisms (1)~ On page 11, the authors state: "It is within the individual cor- poration that the balance between expected returns and expected risks is struck." This is contrary to the approach taken in ecomonic theory and in financial analysis, where the basic unit is the individual invest- n-ient project, not the firm as a whole. The firm is a collection of pro- jects, some of which may be very risky while some are not. (2) On page 14, the authors admit that they were unable to allow for possible biases resulting from the fact that their data source limited them to the larger and more successful firms in each industry. This might be the source of considerable difficulty. It is generally known that larger firms make higher profits than smaller firms, both as a general rule and within the typical industry. Hence if an industry composed of 50 medium sized firms is compared with one made up of 25 large firms and 25 small firms, the variance and hence the riskiness of the latter is likely to be greater, although it is debatable if this is in fact the case. And limitation of data to the larger firms tend to under- estimate the risks faced in a given industry. The fact that the firm is a collection of investment projects provides an additional reason why the larger firm may be able to make a higher rate of return than the small firm. Not only is it able to take advantage of economies of large-scale production, transportation, distribution, advertising, and finance, but it can also adopt a broader scale risk diversification program which makes it less vulnerable to the possibly unfavorable outcomes of in- clividual projects. Hence the industries containing the largest firms may actually face the smallest risks, and yet their rate of return on investment will be relatively higher. PAGENO="0392" 1942 COMPETITIVE PROBLEMS IN THE DRUG INDUSTRY (3) Another reason why it is unfortunate that data could not be obtained on the smaller and less successful firms is that which is given as comment C-3 on Professor Cootner's statement. (4) The industry rate of return is defined simply as the average rate over the entire period. But if there has been a significant upward or downward trend in an industry during the period, investors would react to the trend as well as the average. For example, the average might be 15 percent in two industries, but if one had declined from 20 to 10 percent, while the other had increased from 10 to 20 percent dur- ing the period, investors would be likely to regard the former indus- try as a much riskier investment even if both average rates of return and variances were the same. (5) On page 16 the author states: A company may be receiving monopoly returns (returns higher than justified by risk) on its book assets (i.e., monopoly real returns) while the holder of its equity instruments would receive a "normal" return if the monopoly profits were capitalized when the stock was issued. In model No. I, we note that the drug industry's rate of return is significantly higher than the value given by the regression line for its variance. The computed value is about 14.75 percent, while the actual value is 17.52 percent, or about 20-percent higher. Since this is the book value regression, one might conclude that some degree of monopoly returns are present, since the returns made are higher than those which would be justified by the risk versus rate of return relationship embodied in the regression line. (Here, a statement from page 5 of Professor Cootner's statement comes to mind: "If abuses are found, one must take care to eliminate only excessive, not not necessary returns on investment * * ~." Hence one might conclude tentatively that about a 20-percent reduction in drug industry returns would still leave returns commensurate with the risks as measured by Conrad and Plotkin.) In other words, as far as the investor, the stockholder is concerned, the rate of return he is receiving is just about the average on the basis of Conrad's-that is, the A.D. Little study's-regression line. And at a glance at figure II, shows that the market value basis is slightly lower than the value indicated by the regression line. 1-lence one might interpret the two regressions as showing the presence of sub~ stantial monopoly returns on book value consistent with only competi- tive returns to stockholders due to the competitive nature of the capital market in distinction to the monopolistic nature of the market in which prescription drugs are bought and sold. (6) The statement from page 16 quoted just above is not entirely consistent with the listing on page 11 of "the various phenomenon that might contribute to interindustry differences in basic riskiness." On page 16, monopoly returns are defined to consist of those which are higher than justified by risk. But on page 11, certain of the ele- ments contributing to monopoly power-or conversely, to its absence- are listed as factors contributing to basic riskiness. These include, for example, differences in ease of entry. in elasticity of demand. in i'~~ flexibility, in exposure to foreign competition. and finally, differences in competition among existing, prospective, or potei~tial new products. it would appear that there is serious confusion between the monopoly PAGENO="0393" COMPETITIVE PROBLEMS IN THE DRUG INDUSTRY 1943 elements which contribute to basic risk and those identical elements which, by resulting in monopoly returns, provide earnings greater than those justified by risk. The confusion can only be resolved by assuming that all industries being studied are purely competitive. If this assumption is contrary to fact, then it is not sufficient simply to speak vaguely about factors influencing risk, and about monopoly re- turns, as if they were two different things. Instead one must devise a better theory to relate rate of return to risk and monopoly power more explicitly and satisfactorily. TV. COMMENTS ON THE STATEMENT OF GORDON CONRAD, AND ON THE A. D. LITTLE STUDY, "TRENDS IN MARKET SHARE FOR ETHICAL PHARMACEUTI- CAL PRODUGTS'~ A. General comments (1) On page 1 of his statement, Conrad states that the Little study shows "a significant degree of interproduct competition." But we do not know how significant the data are for drugs until we have data from other industries with which to compare them. Risk, being subjec- tive, is a relative matter and until it ~ shown that other industries have less interproduct competition, the Little study will remain in- conclusive. (2) Even more important, this so-called interproduct "competition" cannot be beneficial to the consumer unless it results either in price competition or in genuine improvements in the quality of the prod- ucts. Otherwise we have change, and perhaps wasteful rivalry, but no progress. Yet on the very first page of the Little report we read: "This report does not explain the reasons, for competitive changes over the time period since this would require revealing product names and company strategy." This effectively prevents the study from making any real contribution to answering the real question: is the eco- nomic performance of the drug industry beneficial to the economy and the consumer? 1 (3) Conrad's statement concludes: These results illustrate one aspect of the potentially high risks facing pharma- ceutical manufacturers, that of the genuine uncertainties as to the length of time any one product can be expected to contribute to the company's profits. 1 This same point was debated during Kefauver's hearings on the drug Industry anti- trust bill in 1961. At that time Professor Markham placed great emphasis on the amount of turnover or change in the rank order of market shares by products In a particular therapy category. But to assess the degree of workability of competition evidenced by such turnover, one should determine how it was brought about: by price competition? by product improvement? or by less beneficial means? But Markham seemed to believe that turnover was a good thing for its own sake, and at least at the time of his appearance had not analyzed its causes. When asked just what was the value to the consumer of turn- over if there were no price competition, he responded: "~ * * I would still prefer, even if the prices are the same, and this I know nothing about, that the firms that are trying to serve my needs as a consumer feel that somehow or other they, through product innovation, or by whatever means-the development of new products, new processes, new drugs-are getting my consumer's outlay In terms of competitive activity" (pp. 2105-2106). During questioning, Markham conceded that he had not examined the facts as to whether or not any drug firm had experienced a change in relative sales rank because of price competition (p. 2096). Markham agreed that price competition is of paramount importance to the consumer, but concluded his contribution to the hearings with this statement: "I have not made any careful study of the workability of com- ietition in the ethical drug industry. I was examining primarily these particular issues that seemed to be important" (p. 2111). This suggests that to Markham the issue of workability of competition was not important-but since he Is known as one of the foremost students of the problem of workability of competition, the statement remains an anomaly. PAGENO="0394" 1944 COMPETITIVE PROBLEMS IN THE DRUG INDUSTRY Again, it should be emphasized that this is a risk to which the in- dustry intensively contributes, and in fact fosters, by its policy of imitative research and product development programs, sales promo- tion strategies, and the like. And very little of this is likely to be socially productive. B. Specific comments (1) The summary section of the report bristles with unsupported "plugs" for the drug industry. On page 4 it is stated that "the devel- opment of effective products with fewer and less severe side effects" is one of the most outstanding features of the market. This is a more sanguine opinion than was stated by many doctors during the Kefauver hearings, particularly those whose statements are contained in part 18 of those published hearings. L. Meyler's book, "Side Effects of Drugs," is highly educational in regard to the general failure of later modified drugs to have side effects much different from the earlier drugs. (2) The summary section also contends that combination products are improvements. Again, there is much medical opinion to the con- trary. The U.S. Pharmacopeia does not list combination drugs; the National Formulary has only a few. Dr. Maxwell Finland stated during the Kefauver hearings that combination drugs lacked flexi- bility and compounded the problems of dosage and toxicity. The summary refers to "combination products in which the in- gredients provide synergistic effects." Dr. Finland referred to syner- gism claims as "incorrect and misleading," observing that such activity "is a highly specialized property related to individual strains of bac- teria and is recognizable only after special tests. Thus, so-called syner- gistic drug combinations can only be tailormade to an individual strain of bacterium after such tests" (p. 13928). (3) It is also contended that there are "obvious economies" in pro- ducing and distributing combination products, but no evidence of this is given. But even if economies are realized~ of what value are they to the consumer? If drug prices are based on the "value" of the medication-that is, what the market will bear-then costs are ir- relevant, and cost savings simply contribute to larger profit margins. V. COMMENTS ON THE STATEMENT OF PROFESSOR FIRESTONE A. Geneiai comments (1) Much of this paper is similar in style and approach to that of Professor Cootner: It is a straightforward pedagogical exercise in which some of the rudiments of index number measurement methods and problems are discussed. (2) But it differs from Cootner's statement, and resembles the "Trends in Market Share" study in that Firestone has a. tendency to compliment the drug industry for alleged advances which have been regarded somewhat more skeptically by medical men. For example, on page 7 and again on page 20, he refers with approval to sustained- release medications despite the fact that Dr. B. IV. Burack previously stated before this subcommittee that such a preparation "remains un- predictable at best" (part 1, p. 330). And on page 21 he speaks with approval of combination drugs, the criticism with regard to which PAGENO="0395" COMPETITIVE PROBLEMS IN THE DRUG INDUSTRY 1945 have already been mentioned in connection with the comments on the "Trends in Market Share" study. (3) On page 23, Firestone develops an argument akin to Whitney's that the price of drugs should be judged in relationship to their "value" to the consumer rather than their production cost. Firestone's presen- tation, although less provocating phrased, is open to the same crit- icism that was made of Whitney~that use of the "value" criterion betrays a monopolistic pricing strategy. B. Specific comments (1) On page 8 it is argued that because there are 110 drug prod- ucts in the wholesale price index, this "probably results in the drug category being more representative than most other categories of like weight" since the inclusion of only 21 products would be required to give the drugs category adequate coverage in terms of the relative "im- portance" of this category in the WPI compared with the total num- ber of products included in the WPI. But it should have been made clear that the number of products included does not necessarily guar- antee the répresentativeness of the data reported unless the reported products are themselves representative of the category in question. (2) On page 11 it is stated that the PMA wholesale price index is limited to brand-name drugs only, which in itself makes it less than perfectly representative of all drugs. Firestone should have esti- mated how much bias might be introduced into the PMA index by ex- cluding generics, and should have indicated how much of the indus- try's sales are in generic form. (3) On page 16 it is claimed that patented drugs declined in price by 24.8 percent between 1949 and 1966, while nonpatented drugs had increased by 1.1 percent. While it is not surprising to discover that prices in the competitive segment of the industry's market presumably increased, and while it is not likely that any segment of the industry could have experienced substantial price declines during an inflationary period unless initially uncompetitively high prices prevailed, the contrast between the price movements would seem to require some additional explanation. There are two questions. First, what sort of criterion was used to distinguish patented and unpatented drugs at various points in time? And, second, why did the price index for patented drugs seem to drop so appreciably? Neither question is answered by: Firestone, although one suspects that to answer the first question may largely answer the second. A de- tailed explanation of the criteria used to distinguish patented and unpatented drugs should have been provided. Senator NELSON. When they use the term "patented drugs" here, are they referring to drugs that were patented but subsequently the patent ran out? Dr. STEELE. That would be true, of any drug which was patented. Senator NELSON. It is one thingto say patented drug prices declined while the patent was still in effect and another to say they declined, if they declined, after the patent expired, is it not? Dr. STEELE. Quite true. Senator NELSON. Then if during a certain period of time, as you sug- gest, when costs are rising, the nonpatented drug prices rose, you might very well expect that, as in most products. PAGENO="0396" 1946 COMPETITIVE PROBLEMS IN THE DRUG INDUSTRY Dr. STEELE. Yes. Senator NELSON. Would it not indicate that if during the same pe- riod, patented drug prices declined, they had set an arbitrarily high price in the first place, since drug industry costs are increasing, too? Dr. STEELE. Yes; I would agree they must have set the price arbi- trarily high so there is only one direction in which it could change. Firestone does refer to large drops in hormones and antibiotics prices, but one must not forget that a major antibiotics price cut was announced in 1960 just before the Kefauver hearings were to investi- gate the antibiotics market. Since this was the first price change in almost 10 years, the motivation for that price cut at that particular time is somewhat problematic. The significance of the reported price trends for nonpatented drugs, however, may not be very great. Since the PMA index is restricted to brand name drugs, it is doubtful if it is very representative of nonpatented drug prices, since much of the sales of these clrugs-in ph3~sical if not in dollar terms-is probably under the generic name. (4) On page 19, Firestone states: "An index cannot tell us whether prices are too high or too low." Quite true. This should be kept in mind when appraising the significance of reported declines in price indexes for brand name drugs. (5) On page 14, Firestone refers to "initial exploratory prices" for drugs during the introductory period of their use. This is an interest- ing phrase and its implications should have been enlarged upon since it suggests market testing by a seller possessed with discretion over prices and an interest in experimenting with prices to determine what the traffic will bear. (6) Although Firestone states that the 17 items included in the BLS Consumer Price Index for all prescriptions since 1967 "should be rea- sonably reliable to measure price movements for all prescriptions"; (p. 18) it is apparent from the study by Agnes Brewster, "Examina- tion of Anomalies in Prescription Drug Prices and Utilization," that neither the drugs included nor the method of pricing them is such as to permit the BLS index to make any claim of being representative. Of the 14 drugs in the 1964-66 index, 11 could be purchased generi- callv. Only 12 of the 200 most frequently prescribed drugs in 1965 were not sold under trade names, and these 12 drugs constituted only 6.2 percent of all prescriptions. Yet six of the generic drugs included in the BLS index were members of the group of the 12 most frequently prescribed generic drugs. It is likely that these six drugs accounted for only 3 or 4 percent of all prescriptions. Yet they constituted over 40 percent of the num- ber of drugs in the 1964-66 index. More important, brand-name drugs account for 90 percent or more of all drug sales, but only about 57 percent of the number of drugs could be purchased under brand names. Hence generic drugs were overly intensively represented, and brand- name drugs underrepresented. The method of pricing the drugs is even more unsatisfactory. All drugs were priced generically, even though over 90 percent of the prescriptions are written by brand name, and probably half or more of the remainder are dispensed by brand name. PAGENO="0397" COMPETITIVE PROBLEMS IN THE DRUG INDUSTRY 1947 For only three of the 14 drugs would the generic and brand-name prices be identical, since these three drugs were sold exclusively by patent monopolists. (7) The entire issue of the relative, movements of drug prices and average prescription charges merits more unraveling than Firestone has attempted. The price of a prescription is based on two factors: The cost of the drug to the druggist, and the druggist's monetary makeup. The cost to the druggist varies with the identity of the drug, the dosage form of the drug, the unit cost of the drug, and the number of units embodied in the prescription. The druggist's markup is a monetary sum added to his cost, a]- though it may be determined eithe r as a percentage markup added to cost, or as a. service charge whk h does not vary directly with the drug cost itself. In either case, the amount of the markup will be influenced by the druggist's costs and the state of competition in his marketing area. (a) As Firestone suggests, one possible explanation for the decline in drug prices as measured by certain indexes and the increase in average prescription charges is that an increase in the distributor's cost has more than offset the decrease in the price of drugs to the distributor. (b) Elsewhere Firestone makes a point of expressing his convic- tion that pronounced variations in prescription prices for the same item cannot be due to "the misdeeds of the manufacturer" (p. 22). One may infer that this conviction includes a concern to show that increases in prescription charges are similarly not to be attributed to misdeeds at the drugmaking level. ("What has been most serious in the misuse of average prescription prices is the use of these prices for measuring what has happened to manufacturers' prices," Fire- stone states on page 21.) (c) However, certain drug firm practices might increase prescrip- tion charges even though drug prices were declining. In terms of the above analysis of the factors determining prescription charges, the following avenues might be exploited to increase prescription charges despite declining prices for each dosage form of each in- dividual drug. These points are possibly quite significant. First, with regard to the nature of the drug: Sales promotion may succeed in changing prescribing habits so ~s to increase both the pre- scribing of drugs in general, of the more expensive drugs in particu- lar, and even of the more expensive dosage forms or modes of each drug. Drug industry critics have claimed that overselling the doctors means overprescribing and overmedication. Last year, I think Dr. Frederick Wolff testified before the subcom- mittee that in his opinion, which he thought was shared by most of his colleagues, something like 60 percent of all prescription drugs pre- scribed were unnecessary. Thus doctors may be persuaded to prescribe more expensive dosage modes, such as sustained-release and combina- tion forms, instead of the simpler mode. They may also be induced to prescribe the drug in situations where they would previously have recommended the use of a proprietary preparation, or simply no medi-. cation at all. PAGENO="0398" 1948 CO~ETITIVE PROBLEMS IN THE DRUG INDUSTRY (2) With respect to the identity of the drug: Doctors may be in- duced to use newer products as rapidly as they are marketed, and these drugs may be marketed at successively higher prices. If this is the case, then steady declines in the price levels of individual drugs may be quite consistent with increases in average prescription charges since new drugs are being prescribed at higher prices in the place of the older drugs the prices of which are declining. Senator NELSON. So if the price index being used is composed of older drugs and this process is in effect, it does help explain at any rate the average prescription cost increases, is that correct? Dr. STEELE. Yes; and this could be a real factor in the BLS index. Firestone's index is the chain index type, so possibly some of this may be allowed for. But not immediately. Senator NELSON. So you would have to keep updating the drugs in the index in order to have a really accurate reflection of a large pro- portion of the prescriptions in the country, right? Dr. STEELE. That is true, but there are also some unavoidable biases in the construction of chain indexes-well, in any indexes. An index has a number of problems which cannot be allowed for satisfactorily. I will say a chain index is better than an index which does not ever change its weights and allow for such biases, but even a chain index can't allow for all of them. Senator NELSON. But if you use an index to look at drug prices and the number of dollars spent for drugs, and the index includes only old drugs no longer being widely prescribed, that does not give you a very informative view of what is happening to drug prices; does it? Dr. STEELE. If as you say the drugs now included in the drug index are no longer being used, this is quite true. With respect to the quantity of the drug prescribed: the greater the number of chronic conditions which doctors can be persuaded to treat with continuous drug therapy, the greater the average prescription quantity is likely to be, to the extent that such patients buy less fre- quently but in larger individual sizes, to take advantage of any econo- mies of the "large economy bottle" variety. (Presumably this trend could continue until long-term drug therapy was available for all chronic conditions-and possibly even beyond that, to the extent that the progressive superannuation of the population as a whole increased the incidence of chronic disorders.) However, the emphasis should be placed upon the possible difference between the need for such therapy, and the degree to which drug firms are successful in creating the impression of the need for such therapy. As Dr. Frederick Wolff, of the George Washington School of Medi- cine, has stated before this subcommittee: It is known that with appetite suppressants, the patients become tolerant to them after some eight weeks and they have no effect. This is not generally recog- nized by physicians, so they are being prescribed indefinitely, and occasionally with very harmful side effects (Part 3, p. 836). d. Each of the above avenues would tend to increase total revenues from drug sales as a conscious end in itself. But the net effect of the various means used to gain this end: a steady stream of "new" drugs; an ever-increasing number of combination drugs, and a medically aim- less proliferation of brand names-all these would act as an uninten- PAGENO="0399" COMPETITIVE PROBLEMS IN THE DRUG INDUSTRY 1949 tional "misdeed" on the part of manufacturers in that they would in- crease the druggist's inventory and other costs of a related nature, and give him an incentive to charge higher prices to cover these increased costs. VI. COMMENTARY UPON THE STATEMENT OF G. R. CONRAD AND J. W. MAEKHAM The argument that "somewhat peculiar" riskiness attaches to the drug companies because of their failure adequately to discharge their responsibilities for conducting every stage in the product develop- ment process in sufficient depth to eliminate all hazards of "unantici- pated side effects," of deficient quality control procedures, of addiction potentials, and of toxicity or inefficacy in use, is in itself somewhat peculiar. If drug development were undertaken from the perspective of the best interests of the public health in the long run, such evaluation might easily take 20 years before a suitably conscientious manage- ment could satisfy itself that the drug merited general commercial use. During such a time period, many related drugs might be evaluated concurrently. If during this period a new drug were discovered which proved clearly superior to other drugs being tested, none of the inferior drugs would ever be marketed. Hence the risk of "the development of a competing product superior to" another already on the market should scarcely exist if drug development were as thorough as it should be. (As far as present market risks per se are concerned, however, the word "superior" can be replaced simply with "newer".) In describing the operation of risk in drugs, the author states: "The types of collapse we refer to do not offer hope, in most cases, of subse- quent recovery of the product's market position" (p. 3). But in his oral presentation, Markham illustrated his somewhat peculiar risk No. 2, "the discovery of unanticipated side effects," by reference to Parke, Davis' brand of chioramphenicol, the so-called Chloromycetin. This was an unfortunate choice. When, as early as 1950, this drug proved that its lethal potential extended to the infected, as well as the infect- ing, organism, some apprehension regarding its use developed, and it was even taken off the market for a 2-month period during 1952. When reinstated, it was only on condition that strong warnings regarding its use be placed on the label and the package insert. For a considerable period of time, chloramphenicol sales were greatly reduced. But Parke, Davis marketing strategies rose superior to FDA precautions. The firm's detail men were given instructions which included memoriz- ing clever and misleading sales spiels and gambits-see the report on the Kefauver administered prices in drugs hearings, pages 192-198- and before too long chioramphenicol sales had risen to the point where it was the most lucrative single brand name drug. It would appear that the period of temporary shrinkage in sales, even though limited in duration, had benefited the drug in the long run. Micro-organisms had less exposure to it than to other antibiotics, and fewer strains resistant to chlorampherncol developed in the early 1950's. In this instance, the somewhat peculiar risk turned out to be a w-indfall in disguise. Senator NELSON. Thank you very much, Dr. Steele. PAGENO="0400" 1950 COMPETITIVE PROBLEMS IN THE DRUG INDUSTRY Do you have any questions, Mr. Gordon? Mr. GORDON. No. Senator NELSON. Your critique and your original statement were very well done and a very fine contribution to these hearings. They will be very valuable in the hearing record. The committee appreciates very much your taking the time to come up here today to testify. Thank you very much. Dr. STEELE. Thank you, Senator Nelson. (The supplemental information submitted by Dr. Steele follows:) [From the Journal of Law & Economics, Vol. V, October 1962] MONOPOLY AND Co,IPETITI0N IN THE ETHICAL DRUGS MARKET 1 (By Henry Steele, Rice LTniversity) Great concern has recently been expressed in regard to the ethical drugs industry as an object of public policy.la At first glance such concern seems dis- ijroportionate in view of the relatively small fraction of the consumer's dollar which is devoted to drug purchases. In 19~9 total sales by manufacturers to dealers of all pharmaceuticals were estimated at ~2.25 billion. Of this total, $1.7 billion consisted of ethical drugs, which are available only upon a doctors prescription, and S560 million of proprietary drugs and medications which can be purchased over the counter without any prescription.2 The conventional mark- up imposed by the druggest (66% per cent of invoice costs) would increase the price to the drug consumer of ethical pharmaceuticals by $1.13 billion, making the total cost of ethical drugs to the final consumer $2.83 billion. This figure is only .9 of one per cent of the 19~9 total personal consumption expenditures of $313.8 billion. Concern with the policies of ethical drug makers is, however, not misdirect ed. since the industry occupies a position of importance in the economy out of proportion to its size. There are several reasons why it is very important to analyze and evaluate the performance of the ethical drugs industry. First, if competition is desirable and monopoly undesirable as a matter of principle, industry size is immaterial, and there is no reason why smaller industries should be permitted to misallocate resources even if, for example, increases in their prices do not have a dramatic effect on aggregate price indices. Second, for the policy maker, there are compelling reasons why the quality of the performance of the ethical drugs industry should be given greater relative w-eight in his social welfare function than its dollar volume of sales might seem to justify. During a given interval of time, the incidence of disease is far from uniform. and those w-ho pay the largest drug bills are, because of their disgbilities, likely to have incomes substantially below the average. The industry shares with the medical profession much of the responsibility for the maintenance of health, and health is naturally a value in itself as well as an important determinant of the productivity of human resources in the economy. The existence of ding regulation and inspection by the Food and Drug Administration testifies to the social premium placed upon adequate performance in this industry. Third, the industry operates under certain unique conditions which provide an interesting case study for analysis by the economist. These conditions include (1) the separation of the authority to prescribe from the responsibility to pay, w-hich in inherent in the status of the prescribing physician as an independent purchasing agent for his patient, and which minimizes the influence of prices on the volume of prescription sales; (2) the use of research to effect often minor product changes which can be marketed with surprising success as major thera- peutic advances; (3) the extreme degree to which product differentiation may be The author is indebted to Professors E. 0. Edwards, M. A. Adelman, G. V. Rimlinger. and J. H. Auten for valuable criticisnis of earlier drafts. in See particularly, Report on Administered Prices of Drugs, Subcommittee on Antitrust and Monopoly of the Senate Comm. on the Judiciary, S. Rep. No. 448, 87th Cong., 1st Sess. (1961). (Hereinafter cited as "Subcomm. Report.") 2 Data obtained from a survey made by Arthur D. Little, Inc., and published in Silber- man. Drugs: The Pace Is Getting Furious, Fortune, May 1960, p. 139. PAGENO="0401" COMPETITIVE PROBLEMS IN THE DRUG INDUSTRY 1951 accomplished by different firms merely assigning different brand names to iden- tical substances arid then advertising such brand names intensively; and (4) the presence of a very high degree of concentration in the production of most of the important ethical drugs, in the absence of any apparent economies of large-scale production. An investigation of the production and marketing of ethical drugs compels the conclusion that the industry may easily be considered delinquent in regard to its impact on the allocation of economic resources. This paper will attempt to demonstrate that such misallocation of resources has been made possible by the existence and abuse of the patent privilege for ethical drug products and proc- esses and by a number of measures which the industry has taken to foster and exploit the remarkable degree of imperfection of market information which the structure of the ethical drugs market pernuits. Such measures can, however, be understood adequately only in the context of actual drug marketing practices, which will be described below. I. THE ECONOMIC FRAMEWORK OF THE ETHICAL DRUGS INDUSTRY Broadly speaking, the economic framew~ork within which the ethical drugs in- dustry operates is conditioned by the nature of demand, the structure of costs, the existence of the patent privilege for drug products and processes, the eligi- bility of branded drug products for protection under the trademark laws, and the existence of the Food and Drug Administration's drug product inspection facilities as an external means of insuring quality in drug products. The nature of demand is such as to distinguish the prescription drugs market from virtually all other markets.3 Sole purchasing authority, as well as the neces- sary initiative, lies with the prescribing physician, who orders the specific drug for which the patient must pay. Chemically identical drugs may be selling under different names at widely varying prices, but the physician has no direct motiva- tion to prescribe the lowest-priced brand or even to become aware of prices at all. (Indirectly he may reason that his fee for services may more readily be capable of collection if drug prices are low, but this would seem to be a distinctly second- ary consideration.) Ideally, the total potential market for a drug or for a group of related drugs consists of the total need for medication (either for cure or for the relief of symptoms) on the part of all individuals afflicted by the various disease entities or clinical syndromes which are capable of treatment by the drug or group of drugs. Economically, total effective demand at given drug prices may fall short of total physical need in the case of individuals with very low incomes and no access to welfare case status, and it may exceed total physical need to the extent that individuals not suffering from those disorders for which the drugs are of use may nevertheless be treated with them. For any given drug, therefore, the actual relevant market is àomprised of the total effective demand for medication on the part of all individuals who can be induced to consult physi- cians arid who are afflicted by those disorders for which physicians may be in- clined or persuaded to prescribe the particular drug. Either for single drugs, or for a group of related drugs as a whole, the demand curve is likely to be extremely inelastic. It has been reported that individuals with severe inflammatory diseases and lOw incomes sometimes do without food in order to buy drugs.4 Shifts in the demand curves for individual products are brought about by direct-mail advertising, medical journal advertising, and by the insistence of itinerant sailesmen or "detailmen," employed by the major drug firms. Advertising cannot manipulate the total incidence of diseases (although to a limited extent, news articles in newspapers and magazines may mention that a given drug can treat a given condition, and thus make more people who suffer, or imagine that they suffer, from a given condition, aware that drug treatment is available) ; but it can shift the existing effective demand from one product to another, and such shifts can perhaps be visualized as parallel rightward or lef t- Exception might be taken to this statement by college professors, who may be said to "prescribe" textbooks for their students. Here, however, the student has alternatives not open to drug buyers. such as the existence of a second-hand textbook market and the possibility of textbook sharing by two or more students. Testimony of E. P. Bransome of the Arthritis and Rheumatism Foundation. Hearings on Administered Prices before the Subcommittee on Antitrust and Monopoly of the Senate Comm. on the Judiciary, 86th Cong., 1st Sess., pt. 14, at 7992-93 (1960). (Hereinafter cited as "Hearings on Administered Prices.") 81-280-68-pt. 5-26 PAGENO="0402" 1952 COMPETITIVE PROBLEMS IN THE DRUG INDUSTRY ward movements of nearly vertical demand curves. The absence of price-con- sciousness on the part of the physician, the inability of the patient to purchase any but the specified drug, and the marked inelasticity of the individual patient's demand curve, all are conducive to the possibility of charging a price which is extremely high relative to marginal cost. The cost structure of the industry is such that there seem to be few, if any, notable economies of scale in the production process. In this regard, a sharp dis- tinction must be drawn between the production of "heavy" or "bulk" chemicals which are commonly produced in enormous quantities by continuous-flow proc- esses (sulfuric acid would be a good example) and the production of "fine" chemicals, such as pharmaceuticals. Batch methods, which allow far less scope for economies of scale, are the rule, especially where fermentation is the key phase in the production process, as is the case in the production of antibiotics (except chioramphenicol) and synthetic corticosteroid hormones.5 These two cat- egories of ethical drugs are the largest in terms of total sales volume. Fermenta- tion vessels are "amazingly standard." 6 Large antibiotics makers usually employ 10 to 15 such vessels.7 Changes in output levels for such products may be accom- plished by the employment of a differing number of individually identical ferm- entation vats, a circumstance conducive to constant returns to scale. A second reason for the absence of important economies of large-scale production is the frequent absence of truly large-scale production. The physical volume of output of the active ingredients in drugs is typically very small, total national output being perhaps in the neighborhood of a ton pear year.8 Actual cost data are kept secret by the drug firms. The inconclusiveness of some of the economic aspects of the 1059-1960 Senate hearings on the ethical drugs industry is due to the "trade secret" status afforded certain crucial cost data and the consequent failure to insist upon the publication of actual production costs for at least some of the high- er-priced patented drugs. To the economist, this is a most unfortunate omission, for such data are otherwise entirely unobtainable. On the basis of the costs sub- mitted to the Senate Subcommittee (but not published in the record), the Report comments upon the absence of economies of scale and the resulting "extremely small size of plant required for economical production." ° It is certainly true that very small firms can purchase the active ingredient for a given drug in bulk (the so-called "bulk powder") from the large firms and tablet and sell it at prices greatly below those which the large firms see fit to charge; there is also con- siderable evidence that very small firms can actually manufacture their own fine chemicals as bulk powder and sell them to small firms in bulk at prices equal to or lower than those charged by the major firms.'° Similarly, on the basis of anti- biotics production costs revealed to the Senate Subcommittee, but not published in the record, Senator Kefauver said at one point in response to Eli Lilly & Com- pany's request for confidentiality as to its production costs, "I will tell you frankly you all have about the same cost figures. . . . For a difference of one or two cents . . . I do not see any sense in all the secrecy. . . . Your costs, Bristol's costs, Upjohn's costs, are all within a very, very narrow range." `~ This tends to support the contention that there are no important economies of scale in pro- duction, inasmuch as the level of output of Bristol's product was almost twice as great as Lilly's. Granted that the ethical drug firms can purchase bulk chemicals from heavy chemicals firms which do themselves produce under conditions of economies of scale, within ethical drugs proper, there seems to be no evidence of anything but roughly constant returns to scale in the production process. The most important economies of large size seem to lie in the area of large-scale sell- ing and advertising, to be discussed below. In the absence of important economies Cortisone Quest: The Right Process Bug. Chem. Week, Aug. 25, 1952, as quoted in Hearings on Administered Prices. pt. 14, at 8291. 6 Gaden. Fermentation. Chem. Engr.. April 1956, p. 159. FTC. Economic Report on Antibiotics Manufacture 118 (1958). 8 Production in 1958 of the most important synthetic corticosteroid hormones (the sec- ond largest category of ethical drugs by sales volume) was, according to the U.S. Tariff Commission. as follows: cortisone, 2.15 tons; hydrocortisone, 3.21 tons; prednisone, .63 tons; prednisolone, 1.4 tons; methyiprednisolone, .5 tons. Hearings on Administered Prices, pt. 14. at 8285-86. Subcomm. Report 4. 10 According to Dr. Phillip Berke, president, Formet Laboratories, a small producer of drug bulk powder, with estimated sales of about $500,000 for all products, could in 1959 sell bulk prednisone at prices competitive with those charged by Merck and Pfizer, firms with over $200,000,000 In sales of all products. Hearings on Administered Prices, pt. 14, at 8056. l11~7, pt. 24, at 14125-26. PAGENO="0403" COMPETITIVE PROBLEMS IN THE DRUG INDUSTRY 1953 of scale in production, however, it might be conceivable that the ethical drugs industry could approximate to the condition of pure competition. Patent protection is available in the drug field for process improvements and for the development of new products. Although a product patent cannot be ob- tained without an accompanying process patent, this is no obstacle to product patents. "Patent medicines" are by no means peculiar to the proprietary drugs field. Unrestricted patent privileges on drug products are, however, virtually peculiar to the United States. Most countries do not award product patents for drugs. Only 28 of the 77 countries for which the Senate Subcommittee could ob- tain patent law data granted pharmaceutical product patents. Twenty-five of these countries have provisions for compulsory licensing. Only Panama, Belgium, and the United States are without any such limitations.12 In the United States, the definition of a patentable drug has been broadened in recent years. Prior to 1940, naturally occurring substances could not be patented; product patents on cortisone and hydrocortisone, for example, were impossible to obtain. In 1946 a patent was obtained on the antibiotic streptomycin, on the grounds that, although it was a product of nature, it was transitory in nature, had never been isolated, and its therapeutic use was unknown.'3 In 1955 a patent was issued for tetra- cycline, even though this compound was jointly produced in the yields from the process for making chlortetracycline, a previously patented drug, and even though tetracycline had previously been marketed by another firm (which did not receive a patent) ~ Patents, therefore, are becoming increasingly easy to obtain in the ethical drugs industry. Trademarks, in the drug industry as elsewhere, may be employed to extend a privileged market position for a larger period of time than the term of the relevant patent. A good example is the case of the barbitu- rate phenobarbital, which the patent holder sold as "Luminal", at $9.80 per ounce. After the patents expired in the 1930's, competitors produced phenobarbital for sale at $.85 to $90 per ounce, but for years "Luminal" continued to sell at $9.80.'~ The Food and Drug Administration has numerous responsibilities with regard to ethical drugs, chief among which are the control of the composition and quality of ethical drugs, the investigation of the safety of new drugs, the certification of antibiotics and insulin, and the prevention of the illegal distribution of prescrip- tion drugs. The laws are very complex. For the purposes of this paper, it may be sufficient to note that the purely economic impact of such regulation is, first, to supplement the quality control of individual firms through the authority of the Food and Drug Administration to inspect samples of ethical drugs (and its more limited authority to inspect production facilities) and, second, to prevent the marketing of dangerously lethal new drugs.1° Violative samples of drugs may be simply confiscated and destroyed, or legal proceedings may be instituted, and civil and criminal penalties may be imposed by the court, including fines large enough to put a small firm out of business. The certification of new drugs, if done with proper regard for the standards necessary to insure drug safety, can be a time-consuming process. Economically, the consequences of regulation are there- fore to provide an additional guarantee to the consumer that all drugs on the market are of acceptable quality and to delay the introduction of new drugs in the interests of safety. II. DRUG INDUSTRY RESPONSE TO THE SPECIFIC ECONOMIC FRAMEWORK The ethical drug industry's response to the given economic framework may be interpreted as a successful strategy to exploit the monopolistic potentialities in- herent in trademarks, the patent privilege, and the great inelasticity of market, in such a way as to overcome the potential for competition implicit in the absence of economies of scale and, to a lesser extent, in the presence of governmental in- spection to insure the safety of all drugs placed on the market. "The vast majority of European drug discoveries have been made In countries where no patent protection was available. Germany, Switzerland, Sweden, Austria, the Netherlands, and Italy are without product patents. France had none until 1959. Great Britain allows product patents on new drugs, but not on combinations of known drugs, and all such patents are subject to compulsory licensing at reasonable royalties. Subcomm. Report 105-6. "Id.,at 141. 14 Id. at 146. `~ Hearings on Administered Prices pt. 18, at 10433. `~ Testimony of Food and Drug Administration Commissioner G. P. Larrick, id., pt. 22, at 12110-36. PAGENO="0404" 1954 COMPETITIVE PROBLEMS IN THE DRUG INDUSTRY The most prominent monopoly element in the market, the patent privilege, may, as is well known, be employed in many ways to limit competition beyond what is explicit in the specific grant of an individual product or process patent. In the chemicals industries the necessary expedients are of easier access than is gener- ally the case. Original patent grants may be extended indefinitely by a judiciously timed succession of improvement patents.17 Improvement patents in chemicals are particularly readily devised, especially where there are process patents on a con- tinuous-flow production method. On the other hand, it is relatively easier to "pat- ent around" existing patents by engineering small chemical changes in product composition. Whenever the product or process developed as a result of patenting- around efforts is more efficient than that developed by the original patent holder, a situation results w-hich requires cross-licensing.18 The negotiating of cross- licensing agreements has as an attendant hazard the introduction of the possi- bility of such meetings of the minds as may be incident to the making of agree- ments involving the mutual compromise of patent monopoly positions. Patents in the drug industry may bring about few absolute monopolies (few patented drugs are uniquely efficacious, even for a single disorder) but may provide the oc- casion, and the economic basis, for the development of a greater sense of corn- munity of interest in matters of price and production. So much is speculation. However, the terms of the actual patent licensing and cross-licensing agreements obtained by the Subcommittee, together with the evidence of price uniformity (not only between different brands of the same drug but also among different drugs in the same field of therapy), lend some support to such speculation.19 It is of course claimed by the drug makers that patents promote competition. Several relatively large drug firms have asserted that patents are a small coin- pany's best friend.20 By this it is meant that the company in question gained monopoly power in some market and was enabled to out-grow its small, competi- tive beginnings.21 This is consistent with the theory implicit in much current anti- monopoly legislation: that the remedy for the existence of monopoly power on the part of the large firms in a given market is to confer a similar degree of mo- nopoly power on the small firms. Really small firms, on the other hand, have testified that a patent is merely an invitation to costly litigation and the only advantage a really small firm can take of a useful patent is to sell it to a larger rival, usually for a lump sum paid-up royalty.22 Even if a small firm were willing to face the hazards of litiga- tion, it would lack the funds necessary to introduce even a hypothetical new product of superlative efficacy because of the needed outlay on sales promotion.23 17 Insulin is an excellent case in point. See subcomm. Report 141. 18 This is often the case where the production of a final product requires as inputs a number of intermediate products, some of which may be capable of chemical synthesis as well as of organic cultivation or fermentation. An important example is the case of the broad-spectrum antibiotic, tetracycline, independently discovered by several drug firms at about the same time. Bristol Laboratories used a method of direct fermentation to produce tetracycline. Lederle had patented an earlier antibiotic, chlortetracycline ("Aureomycin") anti had never licensed any competitor. Lederle found that by dechlorinating chlortetra- cychine. the resulting compound, tetracycline, had unusual antibiotic properties. The chemical dechlorination process was more efficient than direct organic fermentation, but Bristol apparently never applied to Lederle for supplies of chlortetrncycline, presumably in the interest of secrecy. or in anticipation of refusal. Pfizer had done tetracycline research, and had filed a patent application, but found that Lederle had also done so. Pfizer con- vinced Lederle that Pfizer's claims were superior; Lederle, however, had patents on chlor- tetracycline, which Pfizer needed as an input for tetracycline. A mutual agreement was made, insuring both Pfizer and Lederle of a market for tetracycline, regardless of tile award of the patent. Hearings on Administered Prices, pt. 24, at 13848. 13697-701. 10 The currently pending Federal Trade Commission complaint, FTC v. Cyanamid, con- cerns such a point in connection with the tetracycline patent negotiations mentioned just above. The FTC charged the respondents with a conspiracy in connection with the award- ing of the patent to Pfizer, and with a conspiracy in restraint of competition, operating through the subsequent patent licensing arrangements. Subcomm. Report 145-47. 20 Testimony of Dr. P. I. Bowman, president of Bristol Laboratories. Bristol failed to get tile tetracycline patent, but the mere award of a license (in 1956) seems to have been lucrative. Bristol's profits after taxes as a per cent of net worth averaged 2.9 per cent for the four years prior to licensing, but 10.3 per cent for the four years following. Hearings on Administered Prices, pt. 24, at 13835, 13858. 21 Testimony of Henry Hoyt, president of Carter Products. In the five years prior to the marketing of its patented tranquilizer, meprobamate ("Miltown"), Carter had earned 16.8 per cent after taxes on net worth. In the next five years (1955-1959) the rate of profit after taxes averaged 41.0 per cent. Id., pt. 16, at 9113, 9170. 22 Testinlony of Seymour Blackman, executive secretary, Premo Pharmaceutical Prod- ucts. Id., pt. 14, at 8254. 93y of Myron Pantzer, vice-president of Panray Corporation, Id., pt. 16, at PAGENO="0405" COMPETITIVE PROBLEMS IN THE DRUG INDUSTRY 1955 And the most certain evidence of the lack of that sort of competition which is beneficial to the consumer is the rigidity of the prices of patented antibiotics, compared with the steady price decline displayed by the few unpatented anti- biotics, such as the earlier penicillins.24 The extent to which patents are used to limit the number of producers of a given drug can briefly be illustrated by coni- paring the number of producers for 42 drugs, in the categories of hormones, dia- betic drugs, tranquilizers, sulfa drums, and antibiotics. Of these 42 drugs, which account for over half of all ethical drug sales four were unpatented, while 38 were produced under patent or patent-application licensing agreements. Of the unpatented drugs, two were made by three firms, and two by seven firms. Of the patented drugs. 24 were made by a single producer, 8 were made by two firms, 4 by three firms, and 2 by five firms.25 Twenty-four of 42 drugs or 57.1 per cent, were therefore produced by a single patent holder. The concentration of sales is not as great as the concentration of production, but 16 of the 24 patent holders who license no others to produce, also license no others to sell, and therefore market their products as patent monopolists. One firm which licenses another to produce for it licenses no other sellers, and thus constitutes a seventeenth patent monopolist. Of the remaining eight products for which there is a pro- duction monopoly, two producers have licensed only one other seller, three pro- ducers have licensed two other sellers, and the two other three producers have licensed six, seven, and eight other sellers.26 Even in an industry without econo mies of scale, patents can act as a barrier to entry and enable monopoly positions to be attained. An inelastic industry demand curve does not in itself insure monopoly power on the part of the seller, if there is workable competition among sellers. In the absence of economies of scale, small firms could compete with large firms on a price basis. To minimize the effect of suCh competition, the ethical drug in- dustry has taken measures to influence the normal means of dissemination of market information so as to prevent the physician from becoming effectively aware that there are lower -priced sellers in the market, to prevent the -recogni- tion of lower priced alternatives to higher priced drugs wherever the aware- ness may exist that there are lower priced drugs, and to convince the physician that all lower priced drugs are of dangerously poor quality. These efforts to maximize the degree of imperfection of market information have been extremely successful, in large measure because of the nature of the market in which drugs are bought and sold. It is essential at this point to examine that market in some detail. The direct buyers in this market consisted, in 1959, of some 2,800 drug whole- salers, 53,000 retail pharmacies with 110,000 pharmacists, 9,000 public and pri- vate general hospitals, and 4,000 veterinary hospitals.~ The individual patient is the final consumer, but the physician acts as his purchasing agent when writing the prescription. The physician may write the prescription in terms of the generic name of the compound, or by use of a particular firm's brand name. The retail pharmacist may fill generically written~ prescriptions by using any firm's brand of the drug, but any prescription written in terms of brand names must be filled with the brand called for, unless the pharmacist obtains specific permission from the physician to make a "substitution" of another brand. Unless the pharmacist obtains permission in each instance, he is subject to punitive action by the state pharmacy boards in 44 states under the "substitution" doctrine.28 It must be admitted, however, that the courts are not always willing to concur in this doc- trine. In Michigan State Board of Pharmacy v. B. L. Casden the state court dismissed the contention that any meaningful "substitution" had taken place St For the ten-year period 1951-1960 the bulk price of penicillin dropped from $2.50 to $0.21 for ten million units: the bulk price of streptomycin dropped from $3.24 to $0.36 for ten grams; the prices of the patented antibiotics chlortetracycline ("Auteomycin"), oxy- tetracycline ("Terramycin"). chioramphenicol ("Chloromycetin") and tetracycline (intro- cluced in 1955 and sold as "Polycycline." "Achromycin," "Steclin," "Tetracyn," and "Pa n- mycin") remained constant at $5.10 per hott1e of sixteen bbiliirram capsules, until Au- gust 1960 (one month before the scheduled hearings on antibiotics) when they were re- duceti by 15 per cent. Id. pt. 24, at 13664. It should be observed that all of these anti-~ biotics are produced by the same general process of fermentation, with the exception of chioramphenicol. which is produced by a cheaper synthetic process. Subcomm. Report 67. Id. at 67-68. Testimony of Dr. Austin Smith, president of the Pharmaceutical Manufacturers Asso- ciation. Hearinge on Administered Prices, pt. 19, at 10728. 26 Letter from H. S. McNeil, president of McNeil Laboratories, id., pt. 21, at 11713. PAGENO="0406" 1956 COMPETITIVE PROBLEMS IN THE DRUG INDUSTRY when the identical chemical substance was sold under two different brand names.~° Most hospitals operate under the formulary system, which authorizes the hos- pital pharmacist to fill prescriptions with any generic equivalent of the pre- scribed drug, regardless of the brand name written by the physician. Such for- mularies operate under agreements involving the written consent of all resident physicians. In practice it also applies to visiting private physicians treating hos- pital patients. The drug trade associations have attempted to challenge this practice, but without marked success.3° The lack of success here stands in marked contrast to the effectiveness of the campaign of the National Phar- maceutical Council in influencing state boards of pharmacy to adopt "substitu- tion" laws. "Substitution" used to refer to the dispensing of a different chemical substance than the prescription required. The National Pharmaceutical Council undertook to influence state boards of pharmacy to adopt antisubstitution laws which defined substitution in terms of brand names. In 1953 only four states had any sort of antisubstitution law-s, but by early 1959 the efforts of the National Pharmaceutical Council had increased the number to 44,31 Hospitals and dispensaries in the Armed Forces operate under the Military Medical Supply Agency, which purchases on the basis of competitive bids. The bids received often fall well below the wholesale price of the drug in some lines. The Veterans' Administration customarily enters into negotiations with indi- vidual suppliers, although it has also solicited competitive bids, particularly after the Military Medical Supply Agency had demonstrated how this could lower prices. The market consists, in addition, of some 170,000 practicing physicians with the M.D. degree (allopathic physicians), 12,000 osteopathic physicians, and 16,000 veterinarians, all of whom act as purchasing agents for the drug consumer.~2 A vast selling effort is directed at these purchasing agents, primarily at the allo- pathic physicians, to create brand preferences which will result in the writing of brand-name prescriptions. The types of selling effort include (1) advertise- ments in over 300 medical journals, (2) direct mail advertisements, (3) the employment of chemists and pharmacists as highly trained itinerant salesmen, and (4) the sponsoring of medical conventions and meetings, in connection with the exhibition of wares. For the 22 largest firms in the ethical drugs industry. selling expenses were in 1958 the second largest component of costs, averaging 24.8 per cent of sales, as compared with a cost of goods sold of 32.1 per cent of sales. Average selling expenses were therefore 77.3 per cent of the cost of goods sold. Sixteen companies had a ratio of more than 75 per cent, nine had a ratio of more than 100 per cent, and four had a ratio of more than 150 percent. Total sell- ing outlays amounted to about ~3,200 per physician per year.33 The result of this volume of selling effort on the structure of the market has been to eliminate price competition except in the hospital and government bid- ding markets, and to substitute product differentiation and brand preference.34 Any physician who wishes to put himself to the trouble of looking up the generic name of a compound advertised by brand name can write any prescription ge- nerically. The retail pharmacist may then fill the prescription with any available brand. For patented drugs there is little likelihood that there will be any great variation in price, although there are a few such instances.3~ For unpatented drugs, however, the unadvertised, generically named products of small firms may sell for 10 per cent or less of the price of the large firms. ~ Text of Wayne County, Michigan circuit court opinion no. 301799, id., pt. 21. at ii 761-62. ~° Memorandum iS of the National Pharmaceutical Council, 1957, id., pt. 21, at 11835-36. ~1 Subcomm. Report 236. ~ Testimony of Dr. Austin Smith, Hearings on Administered Prices, pt. 19. at 1072S. ~ Data submitted by each firm to the Subcommittee. Subcomm. Report 31. ~ The extent to which advertising. salesmanship. and the bestowal of free samples is regarded by drug makers as capable of creating lasting brand preference is indicated in an article by the vicee-president of Abbott Laboratories, where it is stated : "Practically tIle only ethical drur products whose sales will not be benefited greatly by sampling [the dis- tribution of free samplesi are the rare medicameats distinguished by advantages so great that in simple justice to the patient their actual use is called for." Downs, Advertising, in Drug Research and Development 4S5 (Smith & Herrick ed. 1948). Reserpine, a tranquilizer patented by CIBA. is sold under its brand name, "Fierpasil.' at a wholesale price of 84.50 per hundred .25 milligram tablets. CIBA has licensed several other firms, some of whom make sales of the bulk powder to small companies who in turn can undersell the large firms' advertised brands. The 1959 Drug Topics' Red Book listed 41 firms selling this product. Two firms were selling at over $4.00 per bottle (CIBA's $4.50 was the highest price), two were selling from $3.00 to $4.00 per bottle, two were selling from $2.00 to $3.00 per bottle, eleven were selling from $1.00 to $2.00 per bottle. and 24 were selling at less than $1.00 per bottle. The lowest price, Winsale's $0.45. was exactly 10 per cent of CIBA's price. Hearings on Administered Prices, pt. 18, at 10595. PAGENO="0407" COMPETITIVE PROBLEMS IN THE DRUG INDUSTRY 1957 It is conceivable that some degree of price competition might gain a foothold in the prescription drug market. This is nOt likely to occur, however, even where generic name products are sold in competition with brand name products, for several reasons. First, the physician is probably only infrequently very price- conscious.3° It has even been argued that the physician who prescribes the more costly drugs will be accorded the greater prestige. Second, drug advertisements virtually never mention prices, and the various physicians' reference manuals never give competitive prices.37 It is, in fact, apparently illegal for a generic name product price catalog to compare its prices with those of identical prod- ucts sold under brand names, or even to mention any of the brand name equiva- lents of a generic name product.38 Third, the pharmacist has no economic interest in selling the cheapest brand of drug, since the absolute magnitude of his cus- tomary markup of 66% per cent over invoice cost will thereby be minimized. Indeed, it may even happen that the pharmacist will buy generic name drugs and sell them at brand name prices, if the product itself is not clearly identified as to manufacturer, and if the policing policies in use among druggists are not adequate. Fourth, and perhaps most impoi-tant, the large firms who do the adver- tising, support the trade associations, and hire the detailmen, make every effort to disparage the products of their smaller competitors who sell at lower prices.39 There is probably no other industry in existence where the disparagement of the quality of lower priced products can so completely substitute for actual price competition.4° Physicians are the chief targets for such "educational" disparagement efforts, but the pharmacists and the general publiè are also exposed to it whenever oppor- tunity allows. Several representatives from the large drug firms, and from the Pharmaceutical Manufacturers Association, showed no reluctance to use the Hearings as a forum for the disparagement of drugs selling at low prices. The intent behind these various efforts is clearly to deprive the physician of access to all the information necessary tO allow him to function as a competent purchasing agent for his patient, and to influence favorably his attitude toward brand names. To the extent that the physician, who does not pay, can be con- sidered to have a demand curve for the drugs he prescribes, the efforts of the drug makers can be interpreted as attempts either to give the physician's demand curve an upward slope within the relevant region of drug prices, or to reduce his demand to zero for prices lower than those charged by the major firms. The demand curve of the patient is perhaps nearly vertical up to prohibitively high prices if he trusts the judgment of his physician; at any rate, it will be very inelastic. The patient has no alternative but to purchase the medication at the specified price, or go without, and thus forego the benefit of the medical advice he is paying for. No one wishes to be exposed to the hazards of inferior drugs, regardless of their price. Is there reason to suspect that lower priced drugs are of poorer quality than higher priced drugs? Actually, there is very little to be gained by cutting corners in production and quality control, and much to be lost if such a practice is detected by Food and Drug Administration inspectors. The reasons why no quality differences should be expected to exist between the drugs of large and small sellers may be briefly listed. (1) The same standards of identity. potency, and purity are enforced with regard to all makers for drugs listed in the United States Pharmacopoeia and the National Formviary. The Food and Drug Administration is empowered to inspect any drug produced by aimy firm at any time, and to impose criminal penalties upon violators. (2) The Food and Drug Administration has concentrated its drug regulation efforts on the smaller firms. From January, 1950, until June, 1960, 8,376 samples of drugs produced by the 22 major firms were examined, and 8,621 samples from some 1.200 smaller firms were examined. The larger firms produced 87 per cent of all ethical drugs, ~ Mr. Blackman of Premo testified that it once took him two and one-half hours of argu- ment to convince one physician of his acquaintance that the physician was, in fact, acting as a purchasing agent for his patient. Id., pt. 14, at 8223. ~ Detailmen are also instructed to ignore prices. One doctor persuaded several of his colleagues to interrogate detailmen on drug~prices and costs. If pressed, they would gen- erally quote prices, but no one ever mentioned a cost figure. Id., pt. 18, at 10455,. 3815., pt. 14, at 8212-13. ~ An almost prototypical utterance of this genre is the characterization of the lower priced drug seller as providing "the unfair competition of the unscrupulous one who manufactures under unsanitary conditions, cheapens his product with low quality in- gredients, and falsely labels his product as being better than it is Statement of Dr. Austin Smith, president of the Pharmaceutical Manufacturers Association, iS., pt. 19, at 10738. , 4° See the testimony of Dr. Solomon Garb, 15., pt. 18, at 10476-77. PAGENO="0408" 1958 COMPETITIVE PROBLEMS IN THE DRUG INDUSTRY the smaller firms only 13 per cent.41 The likelihood that a specific drug prepara- tion made by any small firm will be inspected is therefore about 6.9 times as great as in the case of any large ñrm. It must be conceded, however, that during this period the average large firm had 381 samples inspected; the average small firm, 7.2. Each large firm is virtually certain of one or more samples per year; a small firm may hope to go uninspected for a year or more. (3) Most small firms buy the drugs in bulk form from large firms, and merely tablet and bottle the finished dosage forms. Pejorative comments by large firms, therefore, often imply at least some criticism of their ow-n bulk drugs. (4) The United States Pharmacopoeia sets tolerance intervals above and below 100 per cent of labelled strength for the active ingredient in many drugs. If the lower limit is, for example. 05 per cent, it has been claimed that the producer might aim of 95 per cent strength and thereby achieve economies. For a small firm, however, the cost of goods sold may average about tw-o-thirds of sales ~ and the cost of raw materials may constitute about half of the total cost of goods sold ~ so that raw materials costs may be in the neighborhood of one-third of total costs for a small firm. Hence a reduction of 5 per cent in raw material input w-ould cut total cost by only about 1% per cent. or by much less than it would require to begin to justify a procedure which would result in turning out a substantial number of tablets below the 95 per cent low-er limit of tolerance.44 (5) It is claimed that low priced sellers economize on quality control. Again, there seems to be little room for cost savings, for the cost of quality control is probably no more than 21/2 to 3 per cent of the total of all costs 45-surely not enough to justify a procedure which would make the violation of the drug laws a certainty. Perhaps the best argument w-hy quality differences might be expected to exist between higher and lower priced drugs is the contention that the Food and Drug Administration is understaffed and cannot make enough inspections. It is true that insufficient inspections are made with regard to both large and small firms. From January, 1950. to June, 1960, 7,699 samples of drugs produced by large firms ($10 million or more in annual sales) and 9,298 samples from small firms were inspected. During the same period, 84 incidents of irregularities developed in connection with drugs made by large firms, 79 of which were handled by the "drug recall" procedure and five of which led to legal actions. For small firms, 690 such irregularities developed, 206 of which involved drug recalls, and 484 of which led to legal actions. The ratio of legal actions is almost 100 to one, while the ratio of drug recalls is only five to two. Clearly, those irregularities involving large firms are much more frequently negotiated than those involving small firms.44 For the large firms, there was a ratio of 1.1 percent of irregulari- ties to total samples inspected. For the small firms, the ratio was 7.4 percent. The difference is a matter of degree; both figures are too high. The remedy is not to prohibit or to discourage by propagandistic activities the sales of drug at 41 Data presented by Food and Drug Administration Commissioner G. P. Larrick, id, pt. 22. at 12147. ~ The only data available are 68 per cent for Panray Corporation. Id., pt. 16, at 9375. 4~ The only data furnished by any one company for any product show a 52 per cent ratio for Carter's meprobamate. Id., pt. 16, at 9157, 9161. ~ The products of large firms commonly sell at prices as much as several hundred per cent above those charged by small firms, as has been indicated in the case of reserpine. Small firms can undercut large ones by economies in selling costs without any risk of violating the food and drug laws. Large firms spend about 25 per cent of sales revenue in sales promotion, according to data submitted to the Senate Subcommittee by the major firms during the hearings. It is clearly preferable to save 20 to 25 per cent of total cost by foregoing sales promotion in favor of price competition, than to save 12/s per cent in raw materials cost and risk a violation that might result in a fine large enough to put a very small firm out of business. ~ Panray Corporation's quality control costs were 3 per cent of sales revenue; its profits before taxes were about 10 per cent of sales. Hence quality control for this small firm accounted for about 2.7 per cent of total cost. That this is adequate is evident from the fact that no samples of Panray products have been judged violative by Food and Drug Ad- ministration inspectors. Hearings on Administered Prices, pt. 16, at 9375, 9378. ~ It must not be assumed that irregularities involving legal action were more serious, on the average, than those settled by drug recall. If anything, the reverse is true. The nature of 46 violations taken to court, for 32 large and small firms appearing in the record of the hearings, had to do almost entirely with deficiencies or excesses of the active ingredient; only three samples in 46 contained adventitious deleterious substances. Of 285 drug recalls, most were for lack or loss of potency. but several very serious irregularities appeared. For small firms, there were four instances of contaminated drugs, and one Instance where the drug was suspected of causing agranulocytosis. For large firms, there were six instances of contaminated drugs. and one instance of suspected agranulocytosis, three instances where the drugs had caused liver damage or jaundice, one instance of induced acidosis, and one instance (again a recall, not a legal action) where polio vaccine had caused widespread paralytic poliomyelitis. Id., pt. 22. at 12137-61. PAGENO="0409" COMPETITIVE PROBLEMS I~ THE DRUG INDUSTRY 1959 lower prices. This would merely increase the monopoly power of the large firms in those few markets where price competition exists, and would therefore in- crease that part of the profits of the large firms which constitute moflol)O1Y re- turns. The obvious remedy is to provide adequate inspection. Food and Drug Administration Commissioner Larrick testified at the hearings that adequate ethical drugs inspection could be obtained at a cost of an additional $3.418,000.~° For the 22 largest firms in the industry in 1958, net profits before taxes amounted to $562 million.48 A simple computation may throw some light on the nature of the alternative. A reduction of net profits before taxes of $7121 million would cut federal income tax receipts by 52 percent of this amount, or by $3703 million. The net gain would therefore be the $7.121 million saved by drug cus- tomers, less the $3.703 million lost in tax~receipts, or the needed $3418 million. If adequate drug inspection could establish confidence in lower priced drugs to the extent that the resulting competition would lower major thug firm prices and profits by as little as 1.27 percent before taxes, the savings realised would pay for the expanded enforcement program. To this, the substantial benefits ob- tained by the elimination of inferior drugs must be added. Nowhere is the policy of fostering imperfection of market information so much in evidence as in the controversy surrounding the issue of prescription by brand name instead of by generic name. Drug advertisements must, according to law, mention the appropriate generic name, but the generic name is universally given in much less prominent type face,4° and is sometimes concealed by its appearance in an unlikely place in the advertising copy, or else is suppressed in favor of the full and formidable chemical nomenclature. These generic names do' not lend themselves to practical use. Drug makers are free to designate the generic name for any new compound they market, and generally do so in such a way as to insure minimum use of the generic name in favor of the brand name.5° To take a single example from among thousands, the compound with the brand name "Darvon" is burdened with the generic name dextropropophyphefle hydrochlo- ride.5' Dr. Frederick H. Meyers of the University of California epitomized the predicament of the medical profession by commenting that `. . . unless these generic names are disciplined, the United States Pharmacopoeia will eventually become the dictionary of a nonsense language." 52 Such generic names are at the very least inappropriate, since the generic name need serve no other purpose than simple identification of the compound. The definitive and adequately descriptive name is the chemical name, which is a verbal summary of the molecular structure, and which may considerably ex- ceed in length even the most prolix examples of generic names. The generic name is usually merely an essentially unenlightening encumbrance, useless to chemist, pharmacist, and physician alike. An interesting contrast is provided by the procedure for designating new names for insecticides. Rules exist which require that names be short, distinctive, easily spelled, and in conformity w-ith accepted scientific usage.53 No such rules exist in the case of pharmaceutical preparations for human use. The efforts made to suppress the use of generic names have been very effective. Medical schools teach doctors to prescribe by generic name.54 in practice, hiysi- clans have proved vulnerable to the usual advertising appeals, ~nd about 88 per- cent of all prescriptions are written in terms of brand names.'~ By this means, 88 percent of the prescription market is~removed from the sphere of price com- petition. ~° Id., pt. 22, at 12132. 4° Data submitted by firms to the Subcommittee. Profit estimate obtained by multiplying total sales by the 25.8 per cent weighted average sum of profits after taxes and taxes. Subcornm. Report 31. 5° Dr. Austin Smith of the Pharmaceutical Manufacturers Association was at one point during the hearing unable to locate the generic name on an advertisement for a drug caJed "Dimetane" until given a magnifying glass by the Subcommittee counsel. Hearings on Administered Prices, pt. 19 at 10931. 5° Drug-naming procedures are described in id, pt. 21, at 11499, 11675, 11868. 51 Id., at pt. 21, at 11775-76. 52 Id., pt. 18, at 10401. ~ Testimony of Dr. C. 0. Wilson of Oregon State College, id., pt. 21, at 11505. ~° A survey sent to the 82 medical schools in the United States in 1960 brought 77 replies. Sixty-four schools teach only generic terminology ; three teach generic and brand names together; and ten used brand names only under certain circumstances, such as the monopo- lization of a patented generic compound by a single firm. Subcomm. Report 226. 5° 1958 National Prescription Survey, Pharmaceutical Extension Service, Rutgers Uni- versity, as reported in Hearings on Administered Prices, pt. 15, at 8776. PAGENO="0410" 1960 CO1\'IPETITIVE PROBLEMS IN THE DRUG INDUSTRY The industry's reaction to the presence of Food and Drug Administration regulation has been apparent primarily in the area of the certification of new drug applications. In order to secure permission to market a new drug, the applicant must submit a sufficient amount of clinical and experimental data to establish that there is no significant danger connected with the proper use of the drug, i.e., that its acute toxicity is sufficiently low to let it be considered a safe drug. It takes time to conduct sufficient experiments and carry out enough clinical studies to determine probable toxicity in general use. It takes still more time to have the application studied and processed. In drug market- ing, many firms are often working on the same product at the same time, and e~c'h desires to cut the period between discovery and marketing to an abso- lute minimum, for the order of priority in market appearance usually deter- mines the relative sales ranking for different brands of a given drug. Conse- quently, the motivation is to limit experimental and clinical work to the mini- mum acceptable level,57 to skip stages in product development, such as the pilot-plant stage. and to influence the staff of the Food and Drug Administra- tion in such a way as to facilitate rapid approval.57 The response of the ethical drug industry to its specific economic frame- work may be summarized as follows: the employment of the patent privilege to erect barriers to entry into the production of specific patentable drugs, the substitution of product differentiation for price competition, and the use of the accompanying techniques of sales promotion to minimize the impact of the price competition that might be offered by smaller firms. This is accomplished in three ways: the vast bulk of advertising done by the major firms tends, by its mere magnitude, to obscure the very existence of small, non-advertising generic sellers; the employment of opaque brand names for advertised drugs makes it formidably difficult for buyers to detect the existence of lower- priced, generic equivalents; and the campaign of disparagement renders sus- liect the quotation of a low price. No price competition need ever develop for patented drugs; for non-patented drugs, product differentiation, the adoption of deliberately confusing nomenclature, and the waging of a never-ending campaign of disparagement against low-priced drugs can effectively substitute for price competition, and prevent small, low-priced sellers from taking over any appreciable amount of the prescription drugs market, even though the absence of economies of scale Or other barriers to entry will permit small sellers to undersell large by remarkable margins. This has been accomplished in the face of the presence of governmental inspection to insure the quality of all drugs nerely by the extension of the policy of disparagement to include the adequacy of the Food and Drug Administration's facilities for making inspections.n III. EVALTJATION OF MARKET PERFORMANCE IN THE ETHICAL DRUGS INDUSTRY A. Varieties of competition among drugs There seem to be three main dimensions of competition: price competition, product competition, and product differentiation. While price competition is not ~ This Is naturally deplored by physicians and medical educators. Dr. C. D. Leake of Ohio State University noted pertinently: "There is no shortcut from chemical laboratory to clinic, except one that passes too close to the morgue." Id., pt. 18, at 10418. Dr. Barbara Moulton. formerly with the Food and Drug Administration, testified that excessive deference was at times shown to impatient applicants. She testified that if the medical officer in charge of the evaluation of a new drug application Is not satisfied as to the evidence of its safety, the applicant will frequently make an appointment with the medical director. She continued: "I have known such conferences to be followed by an order to the medical officer to make the new drug application effective, with the statement that the company In question has been evaluating new drugs much longer than the medi- cal officer, and should, therefore, be in a much better position to judge their safety." Id., addiction danger warning on the label of a tranquilizer was refused by her superior, whom pt. 22, at 12025. She related an experience of her own in which her request to place an addiction danger warning on the label of a tranquilizer was refused by her su'~erior. whom she quotes as saying. "I will not have my policy of friendliness with Industry interfered with." I'd., pt. 22 at 12032. 5S The large firms, acting through trade associations, have been Instrumental in con- tributing to the inadequacy of governmental inspection. Prior to 1953, the Food and Drug Administration had broad powers to inspect plants as well as products. Tbe Factory In- spection Amendment of 1953, supported by the major firms, made it possible for drug makers to refuse to allow inspection of significant phases of drug operations. Large as well as small firms have not hesitated to avail themselves of this privilege. Testimony of G. P. Larrick, Id., pt. 22, at 12113. PAGENO="0411" COMPETITIVE PROBLEMS IN THE DRIJG INDUSTRY 1961 apparent, closer study reveals that even in the case of sales to retail druggists, there is some price competition via quantity discounts and the shipment of "free goods" in excess of the invoice quantities ordered. In the hospital purchases markets, bids are frequently requested, and in some product lines, sharp price competition may develop. Product competition proper is more easily achieved in the chemical industries than in some other fields. Research effort, whether domestic or foreign, public or private, may culminate in the disc'ovei'y of a new compound with important therapeutic applications. If one firm is awarded the patent, other firms will at- tempt to modify the molecular structure of the compound in order to discover a "different" and patentable therapeutic agent which will in all probability be of use in the treatment of the same classes of disorders as the original drug, but which will hopefully be more potent, or less toxic, or will at least yield a dif- ferent variety and incidence of concomitant side effects. Product differentiation is seen in its purest form when several firms obtain licenses from the patent holder and then proceed to market the identical com- hound under different names, each of which must then be intensively advertised. The choice of a name for such a drug constitutes an exercise in product differ- entiation at a higher level than is usually encountered in the commodity market, since the name typically abandons or repudiates description of the good and refers instead to nothing outside itself; stressing only its abstract (and pre- sumnably unique) identity.59 Product differentiation efforts are of course also en- listed in behalf of products which are~ physically different from each other. Since "ethical" drugs are ethical in the sense that they cannot be bought over the counter or advertised to the public at large, the market consists of the private physicians and hospital pharmacists who can order them. This is a relatively small and well-defined market which can be intensely saturated by advertisements and the employment of itinerant salesmen. There is, in fact, some evidence that the intensity itself, at least in some quarter.6° The role of research as the driving force behind both the search for new prod- ucts and the devising of minor molecular modifications on old ones should be considered separately at some length. Lack of space, however, allows only a brief summary. Spokesmen for the major drug firms invariably defend the height of their prices, or the gap between computed costs and market prices, by reference to the vast sums spent on research in the interests of advancing the cause of health and medical science. It is also asserted that in the absence of the patent incentive, research would disappear. Both statements may be questioned. It may be doubted that they can be reconciled. It is probably difficult to convince an economist that the primary purpose of drug research is not to increase profits. This is economically desirable only if no cheaper way can be found to insure an adequate supply of new and improved drugs. That research is the monopoly of the large and profitable firms, or that the most profitable firms do most of the research, may be questioned. That research outlays are a major or even a very significant factor in the total cost picture for most major drug firms may further be questioned. Finally, that the patent privilege is a necessary incentive to elicit truly productive drug research may be disputed. As a brief summary of the evidence in regard to these issues, the following facts must suffice: 1. The size of the research budget as a per cent of the sales dollar varies inversely with the size of the firm. For a group of 22 large firms and two representative smaller firms, the most profitable firm (Carter, with a profit before taxes of 43.8 per cent of sales) had the smallest research budget (2.7 percent of sales), while the least~ profitable firm (Panray, 10 per cent of sales) is estimated to have had the highest relative research budget (15 per cent of sales), during the year 1959. ~ An example may make the contrast with the usual type of product differentiation by brand name more apparent. Dr. Solomon Garb of the Albany Medical College of Union University submitted that if the drug makers took over the manufacture of canned beans, then rather than selling "Pfizer's Beans" or "Parke, Davis and Co. Beans" they would prefer to coin novel and unique anagrams such as "Sneabs" or "Nabes," or adopt unde- scriptive slogans like "Lo Cals" or "Hi Pro's" and abandon entirely the generic noun. Id., pt. 18, at 10481. ~° See particularly the testimony of Drs. Bowes, Console, Seidell, Weinstein, Bean, Garb Leake, and Meyers, id., pt. 18. PAGENO="0412" 1962 COMPETITIVE PROBLEMS IN THE DRUG INDUSTRY 2. For the 22 largest drug firms in 1959, research outlays comprise the smallest part of the budget. Of the sales dollar, 32.1 cents represented cost of goods sold, 25.8 cents was net profits before taxes, 24.8 cents went for sell- ing costs, 11.0 cents for general and administrative expenses, and only 6.3 cents for research.6' Selling costs average about 394 per cent of research costs. If the cost of drugs is excessive, it is not because of the magnitude of the research budget. 3. The growth of a large American drug firm has typically depended but little upon the quality of its own research. The largest firm Pfizer did not begin to grow rapidly until 1941, when it w-as 92 years old, and then its growth was largely in the nature of a w-indfall gain. Pfizer was one of only a few firms with experience in fermentation methods when w-artime demand arose for fermentation-produced penicillin (discovered in England). The same is true for Bristol, which owes its present position largely to the pur- chase in 1943 of a small firm with some experience in fermentation tech- niques.62 Smith, Kline, and French ow-es its growth not to its own research, but to its ability to take advantage of promising Europeon discoveries. To a lesser extent, this is true of Carter. Additional examples could readily be given. 4. More basically new drugs have been found in European countries with- out drug product patents than in the United States with its unrestricted drug patent privileges. On the basis of an inclusive list (including molecular modifications as well as basic new- discoveries) drawn up by the Senate Sub- committee staff and revised by the Pharmaceutical Manufacturers' Associ- ation, since 1886, 82 drug discoveries have been made in countries without product patents, compared with 79 in the United States, only (30 of which ~sere found in the laboratories of drug firms. Fifteen drugs were found in foreign countries w-ith drug patents.63 Hence, 75 drugs were found by Ameri- can commercial firms and in foreign countries with product patents, w-hile 101 drugs have been found in countries without product patents, and by noncommercial American investigators. 5. Most American drug firm research is not basic research and is not very fruitfu1.~ With regard to four drug categories, American industry is respon- sible for the basic discovery in only one area, cortisone in the corticosteroid field. In oral antidiabetic drugs. the basic drug was found in Germany; in tranquilizers in France, Switzerland, and England; and in antibiotics, in England. Minor modifications have been made by American firms in all four areas. for it is in this sort of research (or more properly, product develop- ment) that the domestic industry excels: the further exploitation of an already accomplished basic discovery. It is in this sort of product develop- ment-research that the risks are least.' Such activities in this country have been fruitful, as in the field of broad-spectrum antibiotics, where important new drugs were found, but no basic research was involved. The basic dis- coveries bad been made in England in 1929. the problems of large-scale pro- duction had been solved by government scientists during the second world u-ar, and the observation w-as current that many molds occurred in nature. some of which might be capable of yielding new antibiotics. All that re- mained to be done w-as to collect and screen soil samples. a tedious, trial- and-error procedure of a routine nature, w-hich the technicians involved have 61 Subcomm. Report 31. - 62 Hearings on Administered Prices, pt. 24, at 1382o. ~ Subcomm. Report 116-18. 64 Dr. H. J. Weinstein, former medical director of Pfizer, testified that much drug firm research "has little relationship to what most people engaged in academic and research activities would consider to be scientific research." He gave as examples the practices of molecular manipulation, the aimless proliferation of combinations of existing drugs which lack flexibility and compound the problems of dosage and toxicity, and the "battle of the additives" where much energy is devoted to showing that the addition of certain substances to an existing drug will intensify its activity. Hearings on Administered Prices, pt. 18, 10243-44. Dr. Console testified: "I think the majority of it [research] is in that cate- gory . . . with many of these products. it is clear while they are on the drawing board that they promise no utility. They promise sales." Id., pt. 18, at 10380. 65 Dr. A. D. Console, former medical director of Squibb, testified that "they [the drug firms I stress that there are many failures for each successful drug. This is truo since it is the very essence of research. The problem arises out of the fact that they market so many of their failures." He continued, "I doubt that there are many other industries in w-hicli research is so free of risks . . . with a little luck, proper timing, and a good promotion program, a bag of asafetida with a unique chemical side chain can be made to look like a wonder drug." Id., pt. 18, at 10372-37~3. PAGENO="0413" COMPETITIVE PROBLEMS IN THE DRUG INDUSTRY 1963 sometimes referred to as "making glorified mud-pies." ~ Furthermore, since no basic research was involved, the returns to more and more intensive exploitation of the same method have approached zero. No important anti- biotic has been found in the United States since 1955, and the same is true in other drug fields. B. Degree of product substitutability among, drugs With this background, which cannot pretend to do justice to the many com- plexities of the research issue, it is possible to turn to the evaluation of market performance in the four drug product categories investigated during the Senate Subcommittee Hearings, corticosteroid hormones, tranquilizers, oral antidiabetic drugs, and antibiotics. These four categories accounted for about 42 per cent of all ethical drug sales in 1959. The corticosteroids market will first be described, and will then be used as the point of departure for parallel comparisons with other drug product categories. By virtue of the emphasis which it places upon the continuous obsolescence of existing corticosteroids through the development of new synthetic steroid hormones eligible for patent protection, this market may be taken as typical of ethical drugs as a whole. The important corticosteroids in 1959 were seven in number: cortisone, hydro- cortisone, prednisone, prednisolene, methyiprednisolone, triamicinolone, and dexa~ methasone. All of the corticosteroids are to some extent substitutes for one another in the treatment of many disorders. Striking evidence of this is avail- able on the basis of a 1959 study by the Pharmaceutical Manufacturers Associa- tion, showing that all seven of the major cor-ticosteroids were prescribed in every one of the seven disease groups requiring such therapy, with but two exceptions.67 No one of these drugs i-s apparently characterized by advantages so great that its actual use is always exclusively in order in a certain disease group. Obviously, for individual diseases or for individual classes of patients this may not be the case, for the disease groups are not very narrow-ly defined, and even within the limits of a single disease, patients of different ages or with different degrees of prone- ness to complications might require different specific drugs. Nevertheless, it is striking that in no instance is as much as -two-fifths of any disease group treated by a single drug. In such a situation it may well be the case that no drug has any commanding advantage over all the others for any major disease group; hence the patterns of demand may be greatly influenced by sales efforts, and the prod- ucts may be differentiated in terms of their claimed freedom from side effects, their relative potency, their prol)able impact on certain types of patients, and in fact upon anything but their relative prices. Similar evidence is available for the tranquilizers market from the same source. This time, each of the seven major tranquilizers is prescribed in every one of the five disease groups requiring such therapy.68 The maximum usage in any disease group is 41.4 per cent. If all drugs were prescribed at random for all disease groups, each would be used in 14.3 per cent of -all treatments. In only four instances is the frequency of use more than twice what one would expect on the basis of random choice. Substitution among drugs is apparently as much -the rule in tranquilizers as in corticosteroids. The oral antidiabetics drug market is more readily described. There are three such drugs, tolbutamide, chioropropamide' and phenformin. All seem to be fairly close substitutes, and tolbutamide (Upjohn's "Orinase") and chiorpropamide (Pfizer's "Diabinese") have been subjected to strenuous competitive product differentiation programs. The market shares follow the order of priority of mar- ket appearance, the earliest to appear having the largest share. These products niay not be perfect substitutes, however, for it has been claimed that chiorpro- pamide is effective in some cases where tolbutamide is not,69 and there is evidence to show that side effects are more frequent and severe with chlorpropamide.7° It is impossible to summarize the extent of possible substitution among anti- biotics in brief compass. The main areas in which substitution is possible are among the five "broad spectrum" antibiotics, all of which are considered inter- changeable in many uses, and all of which are highly advertised; and between ~ Subcomm. Report 124. 67 Hearings on Administered Prices, pt. 19, at 10826. (One drug was prescribed In only five groups.): 68 Id.. pt. 19, at 10827. 69 Testimony of Dr. C. J. Donovan, formerly of Upjohn', id, pt. 20, at 14070-72. 70'Testimony of Dr. Henry Dolger, Mount -Sinai Hospital of New York, id, pt. 20, at 11146. PAGENO="0414" 11964 COMPETITIVE PROBLEMS IN THE DRTJG INDUSTRY penicillin, which is unpatented and not highly advertised, and the broad spec- trum antibiotics. Within its range of activity, penicillin is usually more potent than the broad spectrum drugs.1 But micro-organisms tend to become resistant to certain antibiotics; many of the newer antibiotics are said to be effective against strains resistant to the older ones. Among the broad spectrum drugs, chlor- amphenical may currently be regarded as somewhat distinct from the others, in that fewer strains have become resistant to it, and for two reasons. First, the use of chlora.mphenicOl was discouraged to some extent in the early 1950s when publicity was given to the incidence of its lethal side-effects; hence its less fre- quent use during that period provided less occasion for the development of resistant strains. Second, this drug resembles penicillin more than do the other broad-spectrum antibiotics in that microorganisms do not easily develop re- sistance to it ;72 chloramphenicOl is, in fact, the most lethal of the broad spectrum antibiotics to micro-organism and patient alike.78 The incidence of side effects is often more important than the spectrum of primary efficacy of the antibiotic as far as drug demand is concerned. Advertising may influence prescription demand by stressing the alleged absence of side effects as well as therapeutic efficacy. The absence of any advertising may conversely prevent the wide-scale use of extremely efficacious antibiotics. Bacitracin is a case in point. The drug is so fiee of side effects that some of its preparations may be bought over the counter. The patent is ow-ned by the federal government, and royalty-free licenses may be obtained. The only ethical drug firm currently producing this drug, Pfizer, also sells its own patented antibiotics, tetracycline and oxytetracycline, and does no sales promotion for bacitracin, save of a negative sort. Originally the Food and i)rug Administration restricted its use to hospitals, but later removed the re- striction, but as of September, 1960, Pfizer (despite protests from the discoverer of the drug) continued to print the former restriction on bacitracin labels and brochures.74 C. Pi-odiict competition in individilal drug markets The earliest of the corticosteroids, cortisone, was first synthesized in 1944 by Merck, and commercial production began in 1948.~ The drug first proved its effectiveness in relieving the symptoms of rheumatoid arthritis and related clini- (al syndromes, although concomitant side effects presented complications. Corti- sone is a naturally occurring substance in the human body (adrenal cortex) and no product patent could be obtained. Process patents presented no real barrier to entry, and Merck soon had competition from a number of large and small firms.76 Squibb discovered fludrocortisofle in 1953, by modifying the hydrocortisone molecule. Squibb obtained a patent, sells the drug under the name "Florinef" and has apparently licensed no other sellers. More potent than its parent, its side effects are also more severe.77 Prednisone was the next corticosteroid to be found, along with the related compound prednisolone. These discoveries are credited to the Syntex Corporation of Mexico in 1955. By nature still further modifications of the cortisone molecule they are generally felt to constitute some actual improvement over the parent substance. They are used to treat the same disorders, and have about five times the antirheumatic activity of the parent steroids.78 No patent has yet been awarded for either drug, and several large firms and many small firms sell both drugs. Some medical authorities hold that no essential advances have been made in the corticosteroids field since prednisorie and prenisolone were introduced. Up- john found methyiprednisOlofle in 1957 as a further modification of the pred- 71 L. S. Goodman & E. Giliman, The Pharmacological Basis of Therapeutics 1322--23 (1056). 72 Testimony of Dr. Maxwell Finland of Harvard University, Hearings on Administered Prices, pt. 24, 13947-49. ~ In a nation-wide survey of all severe antibiotic reactions from late 1953 to early 1957, 25 deaths from severe skin reactions and blood duscrasias were associated with the use - of chiorampheflicol. and only two deaths from the same causes were associated w-ith tile use of the tetracyclines. L. Meyler. Side Effects of Drugs 129-30 (1960). ~ Testimony of Dr. F. L. Meleney of the University of Miami, Hearings on Adminis- tered Prices. pt. 24. at 14195. ~ Testimony of Dr. H. C. Kendall of Merck. id., pt. 14, at 8018-20. ~ Testimony of J. T. Connor, president of Merck. id., pt. 14, at 8025-26. Modern Drug Ejlcyclopedia and Therapeutic Index 471 (B. P. Jordan ed. 1958). 78 Meyler, op. cit. supra note 73, at 148. PAGENO="0415" COMPETITIVE PROBLEMS IN THE DRUG INDUSTRY 1965 nisolone molecule, and claimed that it had i!ewer side effects than other cortico- steroids, while having the same general field of use. Upjohn obtained the patent and then licensed Schering to produce the drug. Lederle brought out triamcinolone in 1958, another molecular modification of cortisone, with a potency about equal to that of methylprednisolone, with the same general therapeutic applications as the other steroids, and with the same claim of fewer side effects. Lederle obtained the patent, and licensed only Squibb. The latest modification of the cortisone molecule to be marketed is dexa- inethasone, brought out by Merck and Schering in 1958. This drug is said to be the most potent of the corticosteriods, being about seven times as potent as prednisone,7° and is also claimed to have fewer side effects. No patent has yet been awarded. Merck, Schering, and CIBA were producing this drug in 1959. A Survey of worldwide experience with these newer drugs prompted the opin- ion that their side effects are".. . certainly no less than those of prednisone and prednisolone. .. . With the newer drugs, digestive accidents, fractures, adrenocor- tical inhibition, and indfections are still the main dangers. In contrast to the manufacturers' statements, risks are in no way eliminated by the new deriva- tives." 88 Since it takes years to determine the true incidence of side effects for a given drug, medical opinion is as yet inconclusive, but the most frequently ex- pressed view is that all the corticosteroids will prove to have the same side effects in the long run.8' Thus product improvement has here taken the form of product competition by means of trial-and-error molecular manipulation, and product differentiation by means of massive selling efforts, rather than by revolutIonary advances which would make all previous corticosteroids obsolete. In other drug markets, the development of new products also relies heavily upon the devising of molecular variations on former products. In oral anti-dia- betic drugs, tolbutamide was found in Germany and licensed by. Upjohn for sale in the United States as "Orinase." Upjohn carried out extensive clinical testing, and the drug was released for general prescription use in the United States in June, 1957. Other firms conducted research in order to produce a patentable sub- stitute compound. Lilly and Pfizer came up, with chiorpropamide, a slight molec- ular modification of tolbutamide, and Pfizer began to market this product as "Diabinese" in November, 1958. Lilly withdrew its patent application in favor of Pfizer; at present no patent has yet been awarded. United States Vitamin and Pharmaceutical Corporation later discovered phenformin, a molecular modifica- tion of certain compounds subjected to study by many scientists prior to the dis- covery of insulin in 1920, and marketed it as "DBI." 82 Tolbutamide appears to be unusually free from side effects. Chlorpropamide, on the contrary, is not; nor is phenformin.83 Some physicians are of the opinion that advertising, rather than therapeutic superiority, is responsible for the sales of the later oral antidiabetic drugs. In the tranquilizers market, the basic research on the more potent tranquilizers was done largely in France, with the development of certain phenothiazine com- pounds with sedative effects. Smith, Kline and French was licensed to sell chior- promazine ("Thorazine") and prochlorperazine ("Compazine") in the United States. "Thorazine" was released for sale in 1954 as an anti-emetic; the tran- quilizer properties were advertised only later when it was found that a market existed in the mental institution area. "Compazine" was introduced in 1956 as a tranquilizer, and is a phenothiazine compound similar to chlorprornazine. The market for tranquilizers grew, and so did the number of such drugs. Promazine was marketed as "Sparine" by Wyeth in 1957; again the drug is a molecular modification. Amusingly, Wyeth began to advertise promethazine hydrochloride ("Phenergan"), an antihistamine originating in France, as a tranquilizer.84 78 Report from Pharmaceutical Manufacturers Association, Hearings on Administered Prices, pt. 19, at 10845. 80 Meyler, op. cit. supra note 73, at 148. 81 Testimony of Dr. Russell Cecil, consulting medical director of the Arthritis and Rheu- matism Foundation, Hearings on Administered Prices, pt. 14, at 7985. 82 Subcomm. Report 123. 83 The use of chiorpropamide is associated with liver damage and other mischief. Dr. Henry Dolger, in testifying at the Hearings, concluded: "I deplore the premature prosecu- tion of unnecessary and inadequately investigated modification of sound drugs and their indiscriminate and unsubstantiated promotional campaigns." Hearings on Administered Prices, pt. 20. at 11149. 841n 1952, 20 per cent drowsiness was noted as a side effect. In 1959, it was stated in an advertisement that "Phenergan relieves apprehension, relaxes the patient, and produces light sleep." Id., pt. 16, at 9272. PAGENO="0416" 1966 COMPETITIVE PROBLEMS IN THE DRuG INDUSTRY Other molecular modifications followed: perphenazine ("Trilafon"), trifluopera- zine ("Stelazine"), and trimepazine ("Temaril") ~ In the mild tranquilizers field, a Dr. Berger had discovered mephenesin, a muscular relaxant, in England, and came to the United States to patent a molecular modification, meprobamate, as a tranquilizer, which was marketed by Carter ("Miltown") and Wyeth ("Equanil") in 1955. Both products were highly advertised, a condition of Wyeth's license from Carter being that Wyeth would spend at least 20 percent of total meprobamate sales revenue in promoting its sale.8° Smith, Kline and French found itself with the potent tranquilizers and no mild tranquilizer; it then began to advertise "Compazine" as a mild tranquilizer. "Compazine" is actually much more potent than "Thorazine." 87 Medical opinion inclines to the view that none of the later modifications of the original phenothiazines has been a marked improvement. Dr. Lehmann testified: "There hasn't been a very much better one than the very first ones that came out, in the six or seven years of frantic research since then." 83 No diminution in the incidence of side effects could be demonstrated.88 The situation in the antibiotics market is largely the same, but of greater complexity. The discovery of penicillin in England in 1929, and the further discovery of streptomycin at Rutgers University in 1943, demonstrated the pos- sibility that many naturally occurring molds might have antibiotic properties. Many drug makers set themselves the task of the trial and error screening of thousands of soil samples. The first success was encountered at Yale Uni- versity in 1947 under a Parke, Davis grant, the antibiotic chioramphenicOl ("ChloromyCetin") being isolated. Lederle next discovered chlortetracycline in 1948. and named it "Aureomycin." Pfizer then found oxytetracycline in 1949, and designated it "TerramyCin." Pfizer managed to elucidate the chemical struc- ture of oxytetracycline, and with the molecular structure known, molecular manipulation became possible in antibiotics, although to a more limited extent than is possible with drugs which can be made synthetically. In 1955, Pfizer developed tetracycline as a molecular modification of chlortetracycline.°0 The screening of soil samples continued. Lilly in 1952 discovered erythromycin. Pfizer in 1953 brought out a closely related analog, carbomycin, which affected the same bacteria as erythromyCin, but which proved less effective in human disease than in the test tube. and which was finally withdrawn from the market in 1990.81 In 1956. Pfizer introduced another closely related analog of erythromycin, oleando- mycin: and in 1957, a modification of oleandomycin, triacetyloleandomycin, which was advertised as a great advance, since the same oral dose as in oleandomycin llrocluced somewhat higher concentrations of the drug in the blood. To offset this, Lilly in 1958 modified its original erythromycin to market it in the form of its propronyl salt, which is said to produce a higher concentration of this drug in the blood than triacetyloleandomycin.°2 Hence, five drugs were produced to serve the purpose of one (since slightly higher doses of erythromycin could serve the same purpose as the newer derivatives). Another broad spectrum antibiotic was brought out by Lederle in 1959, demethylchlortetracyCline, or "Declomycin," a sPght modification ef Lederle's chlortetracycline. Fewer side effects than other broad spectrum antibiotics were claimed. Not all physicians agree.83 The later Tenlclilins are patentable variations of the earlier original penicillin w-hich the British discoverer did not bother to patent. In 1952, four American firms made a substitution of benazthine for procaine in the penicillin compound; the parties ~° Subcomm. Report 121. 86 Testimony of Irene Till, Hearings on Administered Prices. pt. 16, at 9208. ~ Testimony of Dr. Heinz Lehmann of Verdun Protestant Hospital, Montreal, id., pt. 16. at 9068-73. 83 Id., pt. 16. at 9029. 89 Physicians employed by drug firms who testified sometimes found it easier to ignore side effects than one suspects their drug users could. Dr. Gibson of Merck went so far as to introduce his aesthetic preference for facial contours as the determinative criterion for evaluating the side effect of "moon face" encountered in steroid therapy. In commenting noon this side effect (caused by redundant retention of cellular fluids, a circumstance con- clucive to additional strain on the cardiovascular system), Dr. Gibson testified: "When those people {malnourished patients] got a moonface I think it is a desirable effect." Id., pt. 14. at 8188. O~) Subcomm. Report 116-17, 124. 81 Testimony of Dr. H. F. Dowling of the University of Illinois, Hearings on Adminis~ tered Prices. pt. 24, at 14167. 92 Id., pt. 24. at 14167-68. 93 Several of the physicians who testified during the Hearings professed to see little dif- ference among the broad spectrum of antibiotics. See id., pt. 18. PAGENO="0417" COMPETITIVE PROBLEMS IN THE DRUG INDUSTRY 1967 agreed to concede priority to Wyeth in return for suitable cross-licensing assur- ances. In 1951, Lilly secured a patent which embraced phenoxymethyl penicillin among other compounds. Lilly apparently did not recognize its value; not until an Austrian firm applied for an American patent on it did Lilly learn of its potential. Lilly then entered into cross-licensing agreements with that firm, and did not market its product until 1955, four years after discovery. The next penicillin, phenoxyethyl penicillin, is a slight modification of phenoxymethyl penicillin, was found in England, and is marketed under four different brand names in the United States.°4 D. Price competition in individua' drug mairkets In general, the presence or absence of price competition in drugs depends upon the presence or absence of patent protection. Even where some price com- petition exists, however, it is rendered all but ineffective in the prescription market because of the inability of the lower price of the smaller producer to overcome the disadvantages of the obscurity inherent in small size and the dis- paragement associated with low price. In such cases, effective price competition is limited to the hospital formulary and government bid markets. Unpatented drugs compete with patented drugs in two out of the four markets under consid- eration, there being no important unpatented tranquilizers or oral antidiabetic drugs. In the corticosteroid hormones market, cortisone and hydrocortisone were ruled ineligible for patent protection, and no patents have yet been issued for prednisone, prednisolone, and dexamethasone; all other corticosteroids have been patented. For cortisone, the price was $20 per gram in October, 1951, but intense price competition forced the price down to $5.48 per gram in 1954, at which level it has since remained constant.95 But the case of prednisone is particularly instruc- tive and merits review at some length. Five major drug firms filed patent appli- cations for prednisone, and an interference proceeding was declared by the Patent Office. The parties involved undertook to make interim arrangements among themselves, involving the payment of "interim royalties" to one of the firms, and as a result all five firms began selling prednisone at the identical price of $17.90 per bottle of one hundred 5 milligram tablets. The Syntex Company of Mexico later became a party to the interference proceedings but was not allowed to participate in the interim arrangements because of its known reputation for price competition.96 Syntex then retaliated by selling the drug in bulk form to small drug firms in the United States, and two of the major firms met its lower price for bulk sales. Syntex then cut prices still lower, from $10.01 per gram in 1957 to $2.36 per gram in late 1959.°~ The major firms' prices at wholesale for the finished drug remained constant at $17.90 per bottle, but small firms were selling at prices as low as $1.75 per bottle by 1959.98 The lower-price brands, however, were unable to obtain more than eight or nine per cent of the prescription market. In the absence of extremely imperfect market information, such great differences in price would be impossible. The larger firms apparently found it much more con- genial to disparage the products of the lower priced sellers than to meet their competition. Had they cut prices, however, there would have been a great deal of room for price competition. The actual full cost of production for a bottle of one thousand 5 milligram tablets of prednisone is only $8.99.~~ This cost is but 5.07 per cent of the equivalen wholesale price of $179.00. No price competition whatsoever developed among the patented corticosteroid hormones. Drugs with increasing potencies per gram appeared from year to year, but the price per tablet of each new drug was set so as to achieve a price per equivalent therapeutic dosage unit identical with those of earlier drugs with different dosage units per ta'blct. In 1958, however, two firms applied for patents upon dexamethasone and marketed it (while patent interference proceedings were still pending) at a price ten per cent below that being charged for its pat- "Subcomm. Report 124-25. ~ Hearings on Administered Prices, pt. 14, at 7884. 9815., pt. 14, at 7920. ~ IS., pt. 14, at 8045. 98 Subcomm. Report 17. 9° Hearings on the Drug Industry Antitrust Act before the Antitrust Subcommittee of the House Comm. on the JudIciary, 87th Cong., 2d Sess. 15 (1962). (Hereinafter cited as "Hearings on Drug Industry Antitrust Act.") 81-280 O-68-pt. 5-27 PAGENO="0418" 1968 CO~ETITIVE PROBLEMS IN THE DRUG INDUSTRY ented subStitutes.10° No small firms entered production, however, because of the higher costs involved than in the case of prednisone. Here, lower prices could be quoted by major firms in spite of appreciably higher costs because of the very large profit margins made on substitute patented drugs. In the oral antidiabetic drugs market, no comparisons between the prices of patented and unpatented drugs can be made in view of the ainence of the latter. Costs may, however, be compared with prices in the case of "Orinase" (Upjohn's name for tolbutamide). The Senate Subcommittee staff estimated the cost of producing one thousand 500 milligram tablets to be $6.86, which is only 8.0 per cent of the wholesale price of $83.40. Upjohn, however pays a royalty of 7.5 per cent of the sales price to the German patent holder, which increases the cost to 15.5 per cent of the wholesale price.10' None of the tranquilizers is unpatented, but price competition has been allowed to develop in the case of a single drug, reserpine. CIBA obtained the patent, but elected to take the unique course of `licensing the patent widely, and also allow- ing the licensees to sell the drug in bulk powder form.'°' The licensees sold the bulk powder to small firms which fabricated their own capsules and proceeded to cut prices by as much as 90 per cent below CIBA's price. CIBA charged a wholesale price of $39.90 per bottle of one thousand 25 milligram tablets; other firms quoted prices as low a Panray's $2.65.'°~ The cost of production of such a bottle is a matter of record: sixty-three cents.'°4 This is only 1.6 per cent of the wholesale price. Price competition in the antibiotics market is brisk for unpatented drugs, such as the earlier penicillins, and for streptomycin, the patent for which was licensed by Merck to several other firms. Price competition is, however, extremely slug- gish for the later patented antibiotics. Penicillin declined in price by about 85 per cent between 1945 and 1948, and by almost 90 per cent between 1949 and 1952.'°~ Between 1952 and 1960 the price fell by another 80 per cent, although there was a marked increase from 1955 to 1957. Increasingly efficient production methods, coupled with the persisting competitive pressures resulting from free entry, made possible such a decrease in the price of this extremely important drug. Streptomycin prices declined by 95 per cent, 60 per cent, and 88 per cent for the same intervals, or by about the same amount as penicillin, and for the same reasons. Broad spectrum prices show very different patterns. Lederle set an initial price of $15.00 for sixteen 250-milligram capsules of chiortetracycline. This may have been above the profit maximizing price; within two months it was cut to $10.00, the same price at which Parke, Davis entered the market in March. 1949. On February 1, 1950, both sellers cut prices by 20 per cent, to $8.00. In April, 1950, Pfizer entered the market with oxytetracycline, selling at $8.40, 5 per cent above its rivals. In May, 1950, these rivals cut their prices by 25 per cent, to $6.00, but Pfizer did not meet the cut until November. On September 27, 1951, Pfizer cut its price by 15 percent to $5.10, and by November 1 its rivals had met the price cut. Tetracycline and demethylchlortetracycline came later, and sold also for $5.10. No further price cuts were made for almost nine years, when Pfizer in August, 1960, again acted as the price leader by initiating a 15 percent "trade discount" adjustment which eventually became a 15 per cent price cut to $4.34~106 (This price cut came one month before the Senate Hearings on antibiotics were scheduled.)'°' From 1948 to 1949, broad spectrum prices fell by one-third; from 1949 to 1951, by 49 per cent, and from 1951 to August, 1960, by zero per cent. It 100 Hearings on Administered Prices, pt. 14, at 8020. 101 Subcomm. Report 20. 102 Reserpine is the name given to the active substance isolated from the root of Rau- wolfia Sepentina, a species of tropical plant the medicinal properties of which had been known in India for centuries. CIBA was therefore perhaps uncertain as to the validity of its patent. 103 Hearings on Administered Prices, pt. 16, at 9369. 10~ Hearings on Drug Industry Antitrust Act 16. (Cost data from McKesson and Robbins.) 105 FTC, Economic Report on Antibiotics Manufacture 164, 179-80 (1958). 100 Hearings on Administered Prices, pt. 24, at 13664. `~ Mr. Duncan of Lederle may have failed to sense the humor of the situation at one point during the Hearings. He testified that Pfizer had cut its prices on a Saturday in order to "steal a march" on the industry. The following colloquy ensued: "Senator KEFAUVER: You mean after ten years of operation, they would suddenly steal a march on you `1" "Mr. DUNCAN: Yes, Senator. These things happen very rapidly. If you can get any kind of an advantage on your competitors, you try to do so." (Hearings on Administered Prices, pt. 24, at 13728.) PAGENO="0419" COMPETITIVE PROBLEMS IN THE DRUG INDUSTRY 1969 can be maintained that since 1951 the costs of producing broad-spectrum anti- biotics have fallen by about as much as the cost of producing penicillin, since the production methods employed are broadly the same.108 A decline of 90 per cent in penicillin prices since November, 1951 is to be contrasted with a 15 per cent decline for broad spectrum antibiotics. Even if broad spectrum costs fell only half as much as penicillin costs, the burden on the drug consumer due to the absence of pure competition in broad spectrum antibiotics would be an extra 30 percentage points in price. The Senate Subcommittee staff made production cost estimates for chioram- phenicol, and made public Bristol's tetracycline production costs. It cost Bristol $1.07 to produce a bottle of one hundred 250 milligram tablets which sold at wholesale for $30.00. Production costs were therefore 5.5 percent of the wholesale price.100 Bristol paid a royalty of $1.03 to Pfizer, increasing Bristol's costs to 8.9 percent of the wholesale price."0 The estimated cost of producing a bottle of one hundred 250 milligram capsules of chloramphenicol (the production and sale of which is monopolized by Parke, Davis) is $1.52, or only 5.0 percent of the whole- sale price of $30.00.'1' Limitations on space prevent adequate treatment of antibiotics price competi- tion in the government bid market. For unpatented drugs, competitive bidding has reduced prices steadily. For the broad spectrum antibiotics, only sporadic com- petition was manifest until foreign bids were accepted for tetracycline at less than half domestic bid levels. Price competition was absent until foreign firms, not encumbered by restrictive patent agreements, were allowed to bid.'12 iv. CONCLUSIONS It seems evident, particularly on the basis of experience in the corticosteroids, tranquilizers, and antibiotics markets, that genuine price competition among ethical drugs is effectively prevented, for the most part, by the existence of product patent privileges. Patent or patent application holders may exercise restrictions on output, and the resulting high level of prices used to finance sell- ing campaigns which contribute to the otherwise serious imperfections of market information and make it impossible for small sellers of generic name products to obtain any significant share of the retail prescription market, even for those products not (or not yet) protected by patents. It is possible that the ethical drug industry might be highly competitive, per- haps almost purely competitive, in the absence of patent protection and the brand name monopoly or oligopoly it makes possible. The only important economy of scale in most drug markets seems to be that of large scale advertising and distri- bution. The extent to which the costs of financing a sizable selling effort can in- crease the price of a product is well illustrated by the experience of the Panray Corporation in attempting to market its own brand of the tranquilizer reserpine, which is highly advertised by CIBA under the brand name "Serpasil" and is sold by CIBA at a wholesale price of $39.50 per bottle of one thousand 25-milligram tables. Panray found it impossible to sell to druggists for less than $20-$21 under its own brand name, "Serpanray." After it abandoned its selling efforts, it was able to make a profit by selling reserpine under its generic name at $2.~5 per 1,000 tablets."1 Simple yet radical remedies are required for the institution of competition in the industry. The first step is the abolition of the patent privilege as it applies to drug products. Most of the countries in the world with patent laws grant no patents on drug products as a matter of public policy. Provisions should be made for the compulsory licensing, at reasonable royalty rates, of drug process patents. It is often claimed that such measures would destroy the incentive necessary for 108 Subcomm. Report 82. Chloramphenlcol costs have probably declined by still more since Parke, Davis discovered a way of producing this drug synthetically. ~ on Drug Industry Antitrust Act 17. 110 HearIngs on Administered Prices, pt. 24, at 13876. luId. pt.24 at13971-78. 112 The hypothesis of the absence of real price competition Is afforded some support by the testimony of Admiral Kplckerbocker. The decision to seek foreign bids was made following a meeting of Military Medical Supply Agency officials with Pfizer representatives, during which meeting, according to Admiral Knickerbocker's testimony, "Mr. Cooney [Pfizer's sales representative] In rather an unguarded moment, made the statement that the price of tetracyline would stay where It was until Pfizer did something about It." Id., pt. 24, at 13819. Had all firms been yoked by patent restrlclons, his posilon would have been stronger. 218 TestImony of Myron Pantzer of Panray, Id., pt. 16, at 9366-68. PAGENO="0420" 1970 COMPETITIVE PROBLEMS IN THE DRUG INDUSTRY pharmaceutical research. A major shift of pharmaceutical research endeavor from private firms to public, university, and foundation channels could scarcely avoid resulting in gains. Private firms carry on relatively little basic research, and most of their so-called research budgets would probably better be described as applied research, product development, and product promotion. A large pro- portion of some very scarce human resources is wasted in the competitive duplica- tion of research programs carried on concurrently by virtually all the major firms as is attested to by the virtually simultaneous discovery of many of the cortico- steroids and antibiotics. In the absence of economies of scale, the removal of patent protection would induce new entry of many firms producing under con- stant cost conditions. The best remedy for a vertical or nearly vertical industry demand curve is a horizontal industry supply curve. The second step is to make Food and Drug Administration inspection and regulation more adequate to meet the needs of the drug buyer and prevent his exploitation. The Food and Drug Administration should be given sufficient addi- tional funds to allow it to inspect adequately the operations and products of all drug makers, large and small. It should further be given authority and addi- tional funds requisite for the control of all drug advertising, and the labelling laws with respect to drugs should be amended to require that all ethical drugs be identified solely by the generic name, in conjunction with the name of the seller employed as an adjectival modifier. All advertising as well as labelling should be undertaken with respect to the generic name, coupled with the name of the seller, rather than concentrating upon the promotion of quasi-chemical syllable assortments. All advertisements should prominently feature suggested retail prices. Both measures would tend to promote the same ends. The abolition of product patents would lead to the elimination of the monopoly profits which are neces- sary to finance intensive selling efforts and wasteful product-modification ori- ented research, and the departure of such profit possibilities would lessen the need for such advertising and the motivation for such research. The reform of drug nomenclature and advertising would enable physicians more easily to be- come aware of the nature of the drugs they prescribe, of the existence of lower- priced drugs, and of the precise nature of the alternative available in the drug market. The insuring of the adequacy of drug products inspection would render specious all activities devoted to the disparagement of lower-priced drugs. Uncompromising measures, rigorously enforced, are indicated in order to halt, and in time to remedy, the existing misallocation of resources in excessive selling efforts, duplicative research and product development programs, and exception- ally high profit levels-all of which are characteristics of the industry and much of which is paid for by individuals who can ill afford to do so. PAGENO="0421" Rep rinted from The Journal of Industrial Economics Volume XII JULY 1964 No. 3 HENRY STEELE Patent Restrictions and Price Competi- tion in the Ethical Drugs Industry. BASIL BLACKWELL BROAD STREET OXFORD ENGLAND Price: Thirteen shillings and sixpence net or $1.90. Annual subscription Thirty shillings or $4.20 1971 PAGENO="0422" PAGENO="0423" COMPETITIVE PROBLEMS IN THE DRUG INDUSTRY 1973 policy recommendations volunteered by witnesses varied all the way from a request that drug patent protection be extended from the present 17 years to a period of 25 or 30 years,4 to a suggestion that some personal violence against certain drug makers and their lobbyists might be in order.5 In part, however, the inconclusiveness of some of the economic aspects of the record is due to the `trade secret' status afforded certain crucial cost data, and the consequent failure to insist upon the publication of actual production costs for some of the higher-priced patented drugs. To the economist, this is a most unfortunate omission, for such data are otherwise entirely unobtainable. This paper is directed to the assessment of price competition among ethical drugs, and the extent to which such competition is prevented by the patent privilege. Before undertaking a detailed investigation of price competition in the varioñs drug markets, however, it is desirable to describe the factors influencing supply and demand for ethical drugs. II. ETHICAL DRUGS: THE MARKET CONTEXT The most striking features of the ethical drugs market are the extreme inelasticity of demand and the monopolistic restriction of supply, with the element of rivalry (rather than price competition) taking the form of product differentiation through research and development. As regards the nature of demand, the prescription drugs market is unique. Only a physician can order such drugs, but the patient must make the payment. Chemically identical drugs may be offered for sale under different names at widely varying prices, but the physician has no direct motivation to prescribe the lowest-cost brand, or even to become aware of prices at all. Furthermore, the demand for a drug which is efficacious in a given disease does not simply consist of the total demands of all individuals suffering from that disease at a given time. Instead, the relevant market for such a drug is composed of the total effective demands of all individuals who can be persuaded to consult physicians, and who suffer (or think they suffer) from whatever disorders physicians may be inclined or induced to prescribe the given drug in connection with. `Ethical drugs' by ~ Testimony of Dr. Chauncey D. Leake of Ohio State University, ibid., Part x8, p. 10433. ~ Testimony of Mr. Mike Gorman, executive director of the National Committee Against Mental Illness. In commenting upon a speech by F. C. Brown, president of Schering, and then president of the American Pharmaceutical Manufacturers Association, he concluded: `In fact, Mr. Brown went much further than Marie Antoinette, and he still has his head and his profits, too', ibid., Part i6, p. 8995. PAGENO="0424" PAGENO="0425" COMPETITIVE PROBLEMS IN THE DRUG INDUSTRY 1975 range of products than is allowed in the original grant. A series of improvement patents may indefinitely prolong the life of an original patent, as has occurred in the case of insulin,9 and these are usually easy to obtain. But it is also easy to `patent around' existing chemical patents, and if such efforts result in devising improvements on products made or processes employed by original patent holders, the logical solution is cross-licensing. Cross-licensing negotiations, in- volving as they do the mutual compromise of patent monopoly positions, supply the motivation for a greater sense of community of interest in price and production policies, and serve to further limit competition.'° Drug patents may bring about few absolute monopo- lies, since few patented drugs are without effective substitutes, but the licensing and cross-licensing agreements prevalent among patent holders facilitate a high degree of market control from the supply side. Lack of space prevents effective documentation of patent abuses, but it may be mentioned that of the forty-two most important patented drugs, twenty-four are produced by only a single supplier." Not all ethical drugs are protected by patents, and since there seem to be no important economies of scale in production for such drugs, it would appear possible for small firms to compete. Hence the larger firms have taken measures (i) to confuse the normal flow of market information in order to prevent physicians from knowing of lower-priced sellers in the market; (2) to prevent the identification of lower-priced equivalents of higher-priced drugs, and (3) to persuade the physician that all lower-priced drugs are of hazard- ously low quality. These efforts have been quite successful. Small firms with little or no advertising budgets cannot make their presence known in the deluge of major firm drug propaganda.'2 Marketing tactics produce economies of scale in advertising drugs, where none exist in producing them. The second objective is accomplished by making the physician brand name conscious, and by devising and advertising generic names of drugs in such a way as to minimize the use of cheaper equivalents. Generic names are designed to be lengthy and complex; brand names are brief and euphonious. Generic names, by law, must be included in all advertising, but are usually printed in minute type face and located in anomalous places. Brand names are given great prominence and are advertised inten- 9lbid.,p. `4'. 10 The currently pending Federal Trade Commission complaint, FTC v. American Cyanamid, et. al., charges the respondents with such a conspiracy in connection with the tetracycline patent negotiations, ibid., pp. 545-7. See also the discussion of the prednisone cross-licensing agreements given below. 11 Ibid., p. 67. 12 Selling expenses amounted to $3200 per physician in 5959, and constituted 24.8 per cent of sales revenue and was 77.3 per cent as large as the cost of goods sold, ibid., p. 3!. PAGENO="0426" 1976 COMPETITIVE PROBLEMS IN THE DRUG INDUSTRY sively. If a physician prescribes by generic name, the pharmacist may fill the prescription with any brand of that drug; if he pres- cribes by brand name, the pharmacist must dispense only the brand called for. Unless he obtains specific permission from the physician, prescribing another brand makes him liable to prosecution by state pharmacy boards under the `substitution' law.'3 The third objective is accomplished by a continuing program of disparagement of lower-priced drugs on the part of the major drug firms through their trade associations and salesmen, especially detailmen.'4 In no other industry is it likely that disparagement of lower-priced products has been able to supplant price competition so completely. Food and Drug Administration regulation has been surmounted by also disparaging the adequacy of FDA inspection. III. PRICE COMPETITION AMONG ETHICAL DRUGS Very little useful quantitative data exist outside the material published in the Hearings, which concerned themselves with four drug categories, antibiotics (about 20 per cent of total ethical drug sales), corticosteroid hormones (i o per cent), tranquilizers (~ per cent) and oral antidiabetic drugs (3 per cent). The corticosteroids market is easily the most representative of the four, since the pattern is not as clear in antibiotics nor as complete in tranquilizers or oral antidiabetics. Consequently, chief emphasis will be given to the corticosteroids market as a paradigm of the effect of patents and other drug marketing institutions on price competition. A. Corticosteroid Hormones There are seven closely related corticosteroid hormones which are used in the treatment of a number of related disease groups, particu- larly those with inflammatory symptoms of a chronic nature. All seven are to some extent substitutes for one another in the treatment of many disorders.'5 Cortisone, the first corticosteroid hormone to be discovered, was found in ~ and commercially produced in 1948. Its price history 13 Among pharmacists, `substitution' formerly signified the dispensing of a chemical substance other than that required by the physician. The National Pharmaceutical Council engaged in a crusade to induce state boards of pharmacy to draft antisubstitution laws defining substitution in terms of brand names, and succeeded in increasing the number of states with such laws from four to forty-four from ~ to 1959, ibid., p. 235. 14 Part i8 of the Hearings contains much testimony to this effect by physicians and medical educators. See particularly the testimony of Dr. Solomon Garb, Hearings, op. cit., Part i8, pp. 10480-5. 15 A study made in 3959 showed that all seven of these drugs were prescribed in every one of the seven disease groups requiring corticosteroids therapy, Hearings, op. cit., Part 39, p. 10825. PAGENO="0427" COMPETITIVE PROBLEMS IN THE DRUG INDUSTRY 1977 reveals a pattern characteristic of a new product in the chemical industry, with initially almost prohibitively high prices falling rapidly to reach a plateau. Merck and Company first developed cortisone, and although unable to obtain a product patent, it managed to keep a pure but transitory monopoly of production until 1952, largely because it alone had processing facilities. In the absence of patent protection, the cost savings from process improvements were passed on to buyers:16 July 1949 $200 per gram Aug. 1950 $50 per gram Dec. 1949 150 ,, ,, Oct. 1950 28 ,, Mar.1950 135 ,, ,, Apr. 1951 24 May 1950 115 ,, ,, Oct. 1951 20 ,, ,, July i~~o 95 But by 1953, Merck's brand, `Cortone' held only 54.8 per cent of the market. Intense price competition continued through 1954, at which time Merck's prices had dropped to $5.48 per gram. Merck, Schering (`Cortogen'), and Upjohn (`Cortisone') have produced virtually all of the United States supply of cortisone since 1954, and each has charged $5.48 per gram at wholesale since that time.17 Hydrocortisone prices fell commensurately with cortisone prices until 1956, when all producers quoted identical wholesale prices of $7.99 a gram.'8 A third corticosteroid hormone, prednisone, was first discovered by Syntex in 1955, but several other firms were working on corti- costeroids, and discovered the compound at about the same time. (Merck claims to have produced prednisone shortly after their discovery of hydrocortisone, but failed to realize its potentialities.19) Schering, Merck, Upjohn, Pfizer and CIBA all filed patent applica- tions on prednisone, and an interference proceeding was declared by the Patent Office. Syntex later became a party to the inter- ference, basing its claims on a 1951 patent covering a class of products which Syntex claimed included prednisone and prednisolone. The interference has not yet been resolved by the Patent Office, but the parties involved undertook to make interim arrangements among themselves. Apparently Schering could convince Upjohn, Merck and Pfizer that it had the best chance of obtaining the eventual patent. In 1955, Merck, Upjohn and Pfizer separately agreed on cross- licensing arrangements with Schering, each involving the payment 16 Testimony of J. T. Connor,~ ibid., Part i~, p. 8030. 17 Ibid., Part 14, p. 7884. 18 Ibid., Part 14, p. 7885. 19 Testimony ofJ. T. Connor, Hearings, op. cit., Part iz~, p. 8027. PAGENO="0428" 1978 COMPETITIVE PROBLEMS IN THE DRUG INDUSTRY to Schering of unconditional interim royalties of 3 per cent of sales, for a period of 3 years, in exchange for the right to make prednisone and prednisolone and market it only in finished dosage form. The royalties were `interim' in the sense that they could be collected by Schering until an interference was declared at the Patent Office, after which time they would simply accrue, but would not be paid to Schering until and unless a patent was issued to Schering. Parke, Davis signed a largely similar agreement in ~ and CIBA fol- lowed in 1958. Such licenses were granted on the basis of patent applications, rather than on patents held.2° The president of Schering defended the practice on the grounds that this was the only way to insure that all parties involved would be licensed to continue production, no matter who was awarded the patent.2' This is an admission that failure to pay Schering interim royalties might jeopardize the granting of a license if Schering did obtain the patent, and the size of the interim royalty is a measure of the degree to which the other companies considered Schering's patent claims to be superior to theirs. All of the cross-licensing agreements (except with Upjohn) provided for the sale of prednisone and prednisolone to third parties in finished dosage form only, thus preventing bulk sales of the finished powder to competitors who might tablet the powder and sell it, either generically or under their own brand names, at their own prices. Upjohn, although not bound by such a restriction, has in fact made no sales of bulk powder except to Schering.22 Schering's president saw nothing unusual in the interim royalty feature. Merck's president disagreed in principle, but con- ceded that there was probably nothing illegal about ~t.23 From the very beginning, the wholesale prices of both prednisone and prednisolone charged by Schering, Merck, Upjohn, Pfizer, CIBA and Parke, Davis have been identical, at $17.90 per bottle of one hundred s-milligram tablets.24 The sixth party to the inter- ference, Syntex, requested a license under Schering's process in 1955. A Schering spokesman is said to have refused because Syntex had a reputation `for knocking the pants off prices'.25 Syntex then began to make sales of prednisone bulk powder (i.e. the finished product in bulk powder form) in the United States in late 1956. Schering 20 Data taken from cross-licensing agreements submitted to the Suncommittee by Schering, ibid., Part 24, pp. 7918-20. 21 Testimony ofF. C. Brown, ibid., Part i~, p. 7928. 22Report, op. cit., p. 151. 23 Hearings, op. cit., Part 24, p. 8096. ~ Data from American Druggist Blue Book, as reported in the Hearings, ibid., Part i~, p. 7884. 25 Ibid., Part 24, p. 7920. Reference is to a remark allegedly made by Irving Jurow of Schering (who disavows the phrase) and reported in Fortune, August 1958. PAGENO="0429" COMPETITIVE PROBLEMS IN THE DRUG INDUSTRY 1979 brought suit for infringement under the process patent, and Syntex retaliated by suing both Schering and Merck for infringement of its 1951 product patent. Schering thought enough of the validity of the Syntex patent to agree in 1958 to pay Syntex a lump sum paid- up royalty in return for a release from all claims, present or future, under the 1951 patent.26 Merck made a similar agreement with Schering at about the same time.27 However, the bulk powder price structure for prednisone and prednisolone had already begun to deteriorate by the time such agreements were accomplished. When Syntex began to make its `uncontrolled' bulk shipments into the United States, Pfizer early in 1957 violated the terms of its cross- licensing agreement with Schering and made bulk sales of its own powder to third parties, meeting Syntex's low price. Merck soon followed suit, but Upjohn continued to sell in bulk only to Schering.28 This competition in the bulk market steadily forced bulk prices downward, from levels which were initially quite low in relation to the wholesale price of the tableted and bottled powder. The bulk prices charged by Syntex declined from $io.oi per gram in the first quarter of 1957 to $2.36 per gram in the third quarter of 1959. Merck and Pfizer are said to have met Syntex's price; after they entered the market at lower prices they made sales at Syntex's expense, and the records show that sales by Syntex declined, despite the great reduction in its price. In March 1959 Schering and Syntex made an agreement, under the terms of which Schering agreed, in the event that it was awarded the prednisone patent, to license Syntex (i) to make prednisone and sell it in bulk form to pharmaceutical manufacturers for use as a chemical intermediate in the manufacture of products other than prednisolone, or of those so closely related to prednisone or predniso- lone as to constitute their equivalents, and to pay a 6 per cent royalty on such sales; (2) to make and sell prednisone in finished form under its own label at a 6 per cent royalty rate; and (~) to sell prednisone in bulk to licensees of Schering, who in turn would pay royalties directly to Schering. If Syntex were to be awarded the patent, Schering would receive a license to make and sell prednisone at a royalty rate of 3 per cent.29 This agreement contrasts with those made by Schering with Merck, Pfizer, Upjohn, CIBA and Parke, Davis. The latter agree- ments provided for immediate áross-licensing, the payment of a 26 Testimony of F. C. Brown, ibid., Part 54, PP. 7921-2. 27 Text of Merck-Schering agreement, ibid., Part i~, pp. 8364-g. 28 Ibid., Part 14, p. 8095. 29 Text of agreement submitted by Schering to Subcommittee, ibid., Part 15, pp. 8878-9. PAGENO="0430" 1980 COMPETITIVE PROBLEMS IN THE DRUG INDUSTRY 3 per cent interim royalty (5 per cent for Parke, Davis), such interim royalties to be in effect for 3 years (the earlier agreements with Merck, Pfizer and Upjohn stipulated that in the event of an inter- ference proceeding,3° interim royalty payments would be suspended and allowed to accrue until a patent was actually issued; the CIBA and Parke, Davis agreements called for interim royalties for 3 years after the date of the agreement itself); and, finally, all of the agree- ments (except with Upjohn) restricted sales to finished dosage forms only.3' The purpose of restricting sales to finished forms only is clearly to prevent any shipment of bulk powder being made available to non-licensed competitors (chiefly the smaller firms who take the bulk chemical, tablet and bottle it, and sell it under their own brand name, or more frequently, under the generic name) who might make use of such an `uncontrolled' supply to introduce price competition into the market. It is by no means unusual in the drug trade to find licensing agreements predicated upon the basis of patent applications rather than issued patents, but it is somewhat unusual to have interim royalties paid on the basis of applications alone. Counsel for the companies pointed out that there was nothing illegal about interim royalties, since one firm might very well pay another a royalty in order to use a secret process for which no patent was sought. In such circumstances, the consideration for the royalty would be the disclosure of an otherwise unknown process. Schering's agreements with Pfizer, Upjohn, Merck and perhaps some of the others, could not be so felicitously construed, since all firms had been doing intensive research in the field and were roughly on the same footing. It was generally agreed, however, by Schering and its licensees, that it would be impossible to enforce the restrictions on the form in which the good is sold, until a valid patent had been issued.32 Nevertheless, a pattern of action which can be described as volun- tary compliance was observed until Syntex began making its ship- ments of bulk prednisone, at which time Pfizer and Merck violated the letter of their agreements with Schering by following suit. The initial effort at voluntary compliance may conceivably have been a gesture of goodwill toward Schering, or it may have been a collec- tively acceptable way of dealing with the problems posed by highly profitable prices (which could bear a great deal of shading) coupled 30 The prednisone and prednisolone interferences were declared in December i 956 and December 0957 respectively, ibid., Part i~, p. 8093. 31 Licensing agreements submitted to the Subcommittee, ibid., Part `5, pp. 8360-4. 32 of F. C. Brown, ibid., Part i~, pp. 7926-8; testimony of J. T. Connor, Part i~, p. 8094. PAGENO="0431" COMPETITIVE PROBLEMS IN THE DRUG INDUSTRY 1981 with the probable excess capacity on a product which as yet had no patent protection.33 When Pfizer and Merck violated their agree- ments by making bulk sales, Schering took no action, and its presi- dent indicated in the hearings that he realized that this part of the agreement was not enforceable.34 Price competition ensued. Small companies could obtain the bulk powder from Syntex, Merck, Pfizer or from a small domestic producer, Formet, and package their own finished dosage forms with no restrictions until such date as a patent might be issued. If a patent were to be issued to any firm other than Syntex, these smaller pro- ducers could expect to be eliminated from the market immediately.35 Until such time, however, small firms were free to cut prices as far below those of the major firms as they could. The lowest wholesale price reported in the hearings ($1.75 per hundred 5-milligram tablets) was less than one-tenth of the $17.90 charged by Merck, Schering and Upjohn. Given perfect market information, differences of 90 (or of iooo) per cent in the prices of identical products would be impossible. The larger firms found it much more congenial to disparage the products of lower-priced sellers than to quote lower prices to meet their competition. Nor was any price reduction necessary in the retail druggist prescription market, where well over 90 per cent of the sales were made by the three large firms. It may be appropriate to inquir~, in the perhaps extreme case of prednisone, what substance there may be to charges that lower-priced producers sell inferior prednisone. Prednisone could be purchased under the generic name in late 1959 from a number of smaller firms, for prices ranging from $1.75 to $12.00 per hundred ~-mil1igram tablets at wholesale. It could also be purchased from Schering as `Meticorten', from Upjohn as `Delta- sone', from Merck as `Deltra', or from Parke, Davis as `Pa~acort' ~ Excess capacity is very likely to have existed. Dr. Philip Berke of Formet Laboratories testified that an investment of only four or five million dollars would easily be sufficient to supply the entire world demand for prednisone and prednisolone, ibid., Part 14, p. 8056. ~ Ibid., Part `4, p. 7928. ~` It has not been the practice of the major drug firms to license smaller companies. Blackman of Premo, a small firm selling prednisone and other drugs at low prices under generic names, stated that he could produce the newer, and as yet unpatented, cortico- steroid dexamethasone in the interval before a patent was issued, but that after its issuance his investment in production facilities would be worthless. He declared: `With as much assurance as any human being can muster, I feel we would not get a license', ibid., Part i 4, p. 823!. Dr. Berke of Formét Laboratories, which is apparently the only small firm in the United States making prednisone and prednisolone in bulk form, was of the same opinion, and indicated that he would apply for a dexamethasone license `tongue in cheek'. He further indicated that it was cheaper to produce prednisone and prednisolone than dexamethasone; hence he could afford to produce the former while patent pro- ceedings were still pending, but not the latter, ibid., Part 14, pp. 8057-8. PAGENO="0432" 1982 COMPETITIVE PROBLEMS IN THE DRUG INDUSTRY at $17.90. Claims that brand name differences mean quality differences must be dismissed. Prednisone is listed in the United States Pharmacopoeia, which specifies standards of purity, potency and identity which must be met by all makers, and there is no therapeutic gain in producing drugs to a higher purity exceeding them.36 They are enforced by Food and Drug Administration inspection. No producer of prednisone, large or small, had ever been accused of irregularities in connection with prednisone products as of the date of the hearings. For prednisone, all sellers must stay within the potency tolerance, between 90 and 110 per cent of the level claimed on the label, whether their products are branded or are sold generic- ally. Since the cost of the active ingredient is typically about one- third of total factory cost, there can be no appreciable savings in aiming at the lower limit. Furthermore, the firm that sells the pro- duct may not manufacture the bulk powder, or even do the tableting. Parke, Davis, in 1958, made none of its own bulk powder, while the smaller companies bought their bulk powder largely from Pfizer and Merck (who are reported to have met Syntex's lower prices) as well as from Syntex and Formet. Further evidence of the reliability of at least some of the lower-priced sellers is given by the fact that both Premo and Chase were allowed to make sales to the Military Medical Supply Agency, which makes thorough inspections of the facilities of all low bidders on government drug purchase contracts.37 No brand name preference is given for purchases by government agencies or by hospital formularies. Here, small firms may compete with large firms in terms of price, and large firms, especially on government contract bids, may have to cut their prices drastically, and in such a way that something may be learned of their costs and pricing policies. Sales were negotiated by the Veterans' Administra- tion with Merck and Schering in February 1958 at a price of $136 per bottle of one thousand s-milligram tablets of prednisone. In March 1958 competitive bidding was instituted; Merck cut its bid to $95, Schering dropped to $68, but a small firm, Panray, obtained the contract with a low bid of $38.50, or 28.3 per cent of the price negotiated the month before. Competitive bids submitted to the Military Medical Supply Agency at about the same time showed Merck bidding $75, Schering $79.74, and another small firm, Chase, obtaining the contract with a low bid of $41.50. In April 1958 Merck and Schering entered their previous bids, and Premo was the low bidder at $38.40. In December of the same year, Schering cut its 36 Testimony of Dr. Walter Modell of Cornell University, ibid., Part 21, p. i i6io. `~ Testimony of Captain H. R. Fahlbusch of the Military Medical Supply Agency, ibid., Part 21, p. 15547. PAGENO="0433" COMPETITIVE PROBLEMS IN, THE DRUG INDUSTRY 1983 price to $23.62 and obtained the contract, underbidding smaller firms. At this time, Schering was charging $170 per thousand at wholesale to druggists, implying a discount of 86.7 per cent to the government.38 In February 1959 Premo was again the low bidder, with a price of $20.98. At that time, Premo was selling at $26 per thousand to the druggist, with the comparative discount to the government at 19.3 per cent.39 It is interesting that quoted wholesale prices and bids tend to vary inversely with the size of firm. The three large firms charged identical prices of $17.90 per hundred s-milligram tablets. Prices of the smaller companies named varied from Lannett's $12.00 to Premo's $2.35. A Washington chain drugstore dealer reported to the Sub- committee that he had been able to purchase the same goods at $1.75 from an unidentified `first line' company, and the 1959 Drug Topics Red Book listed twenty firms selling at prices below $3.oo.~° Such facts lead one to the question: what is the relationship between size of firm and cost of production? The presidents of Merck, Schering and Upjohn testified that the sales they made to the government were either not profitable, or did not yield a sufficient profit to cover the cost of an adequate research program. The smaller firms testified that they definitely made a profit at their prices. Syntex, however, spent about half again as high a percentage of sales revenue upon research as did the larger firms, and Formet spent relatively nearly twice as much.4' A comparison of selling costs is more instructive. Premo (the only small firm for which sales costs are available) devotes 2.0 per cent of its sales dollar to selling activities while the weighted average for the three large firms was 24.3 per cent.42 There is some reason to suspect that Premo might spend more of its revenues on selling efforts than do other small firms. Premo once attempted to increase its sales force, hire detailmen and compete with the large firms, using their own methods. In 1955 the company's sales outlays increased to almost 7 per cent of sales, but their campaign failed because of `the tremendous increase in the advertising dollars spent by our large competitors, to the extent that our efforts appeared, in the market place, as a mere spark in a vast conflagration'.43 Advertising outlays were cut, but may yet remain above the levels for small firms which never attempted large-scale selling campaigns. 38 Data submitted by the Military Medical Supply Agency to the Subcommittee, ibid., Part 28, pp. 11551-3. ~ Data from Premo 8959 Hospital Price Catalog, as reported in the Hearings, ibid., Part i~, p. 8709. 40 Report, op. cit., p. 87. 41 Hearings, op. cit., Part i~, pp. 8064, 8301. 42 Report, op. cit., p. 3!. ~ Hearings, op. cit. Part 84, p. 8215. - 81-280 0-68-pt. 5-28 PAGENO="0434" 1984 COMPETITIVE PROBLEMS IN THE DRUG INDUSTRY Production costs for prednisone are only a small fraction of the wholesale price. Rather than make public the actual production costs of individual firms, the Senate Subcommittee staff estimated the cost of a bottle of one thousand 5-milligram tablets of prednisone at $13.61 .44 This cost estimate was attacked by industry witnesses as being too low, as not including selling and research costs, etc., etc., despite its being made unmistakably clear in the record that the cost estimate referred only to production costs. Actually, this cost estimate is quite generous. It is based on a bulk powder cost obtained by using the prices charged by bulk producers for sales of bulk powder to other firms, and hence includes not only their actual direct and allocated costs, but also some margin for profit. To determine the cost of tableting and bottling, firms which perform such functions on a contract basis were contacted, and the median cost quotation (not the lowest) was employed. The actual production cost of McKesson and Robbins is only $8.99 for such a bottle.45 This cost is only 5.07 per cent of the equivalent wholesale price of $179 per thousand, and only 3.17 per cent of the retail price of $ 283.33.46 Price competition departs and then reappears on the scene with regard to the three newer corticosteroids. When Upjohn introduced methylprednisolone in 1957 (`Medrol'), it first had the market to itself~, and only later licensed Schering, which had no chance to market its brand until 1959. Being 25 per cent more potent than prednisolone (a standard dosage form is the 4-milligram tablet as compared with the s-milligram prednisolone tablet), its price per milligram was increased by exactly 25 per cent, so that the price per tablet is the same, and the introduction of the more potent product did not give rise to any price competition.47 As a new drug, however, ~ Report, op. Cit., p. 34. ~ Testimony of Dr. Herman Nolen, president of McKesson and Robbins, in Hearings Before the Antitrust Subcommittee of the Committee on the Judiciary, House of Representatives, Eighty-seventh Congress, Second Session, Washington, Government Printing Office, I 962, p. `5. 46 While the present paper is not concerned directly with the retailer's markup of 66~ per cent, it certainly helps to contribute to the magnitude of the final price to the consumer. It is well known that retail druggists, like retailers everywhere, dislike price competition in practice, and favor high markups-the role of the National Association of Retail Druggists as one of the prime movers behind the `Fair Trade' laws is notorious- and the complaints that high markups have failed to produce satisfactory profits for retail druggists is simply a consequence of the freedom of entry in the field and the over- capacity, low turnover rates and higher average unit costs which results from new entry in response to the high unit margins. It is pertinent to call attention to this point in view of the possibility that pressure may be exerted by organized druggists on the manu- facturers to discourage any tendencies toward price cutting, or at least that the drug manufacturers may be aware of such distributor sentiment when contemplating possible price cuts. It must be admitted, however, that no shred of evidence on this point was offered (or sought) during the whole of the hearings. `~7 Data from American Druggist Blue Book, 1959-1960, as reported in the Hearings, op. cit., Part 14, p. 8324. PAGENO="0435" COMPETITIVE PROBLEMS IN THE DRUG INDUSTRY 1985 methyiprednisolone began to take over a notable share of the market, and by 1958 `Medrol' had i8.i per cent of the new pres- cription market, while the leader in 1956, `Meticorten', had fallen from 53.9 per cent to 30.7 per cent.48 In 1957 Lederle introduced triamcinolone, a corticosteroid identical in potency to methylprednisolone, and licensed it to Squibb. Both firms set prices at a level identical with that of prednisone, prednisolone and methylprednisolone. Lederle's brand, `Aristocort', had 23.8 per cent of the new prescription market in 1958, while Squibb's brand, `Kenacort', had 8.9 per cent.49 Eighty-five per cent of the bulk powder, however, was made by Squibb, and only i~ per cent was made by Lederle. As a consequence of Lederle's prior and intensively advertised appearance in the market, it would appear that physicians' brand preferences crystallized in favor of Lederle's product so predominantly that Squibb had no outlet for much of its production except to sell it in bulk to Lederle. In 1958 Merck and Schering introduced dexamethasone, a corti- costeroid with a potency about six and two-thirds as great as prednisone, the standard tablet containing .75 milligrams instead of ~ milligrams. For the first time since the prices of cortisone and hydrocortisone reached a plateau, price competition was introduced into the market when Merck priced its `Decadron' at retail at 27 cents per tablet, 10 per cent below the prices of the other four competitive corticosteroids. Schering later marketed its brand, `Deronil', at the same price as Merck. Merck, intending to regain its earlier predominant place in the market, gave its dealers additional quantity discounts amounting to as much as 10 per cent.50 A patent interference was declared between Merck and Schering. Merck subsequently licensed CIBA and Organon of Holland. CIBA then began to market dexamethasone as `Gammacorten' in 1958. While dexamethasone has yet to be patented, no small firms seem to be producing this compound, because of the relatively high costs in- volved. In fact, it may be the case that the profit margins to the producers are considerably lower than in the case of the other corticosteroids. Sales of bulk powder by Merck to CIBA were made in 1958 at $6~ per gram. If the bottling and tableting charges used previously are added to the bulkpowder cost, a cost of $72.69 per gram is obtained, which amounts to 33.8 per cent of the wholesale price of $214 per gram.5' (The comparable figure for prednisone is 8.7 per cent.) This is only slightly more than the average ratio of 48 Data from the market research staff of Merck and Company, ibid., Part 84, p. 8028. ~ Ibid., Part i~, p. 8028. 50 Testimony ofJ. T. Connor, ibid., Part 84, p. 803!. 5' Ibid., Part 84, p. 8324. D PAGENO="0436" 1986 COMPETITIVE PROBLEMS IN THE DRUG INDUSTRY 32.1 per cent for cost of goods sold to total sales for the twenty-two largest firms in the industry, as cited above. Dexamethasone there- fore combines a lower price with much higher costs-a phenomenon impossible in pure competition. It is inappropriate to attempt to trace in corresponding detail the course of price competition, product by product, and in the light of existing patent licensing agreements, in the other drug group mar- kets. They all reveal the same general features, hence only a sketch will be given of prices, the relation of prices to computed costs and competitive bids, and of patents, in the other three markets. B. Oral Antidiabetic Drugs The oral antidiabetics market may readily be summarized in this regard. Each of the three products is either patented or produced under patent application arrangements. Upjohn set the price of `Orinase' in such a way as to be almost exactly equal to the price of insulin on a per dose basis, at 13.9 cents per 500-milligram tablet, comparable to 14.0 cents for 10 units of insulin.52 This price was not based on cost, for the cost of insulin is no doubt considerably greater than the cost of `Orinase', or tolbutamide, since animal pancreas is undoubtedly more expensive to obtain and process than the in- dustrial chemicals required to produce tolbutarnide.53 Yet the greater convenience of administration of an oral drug over an injectable drug would probably support a higher price than the insulin equiva- lent. The price was simply set equal to insulin's price, apparently in deference to the industry's tradition of pricing new products at (or very near) the price of existing drugs used to treat the same dis- order. (The price set is actually below the total cost of insulin therapy, including syringe and needle.) The price of insulin itself had not been changed since 1947. Had the practice of medicine not changed, we might still be paying second-century A.D. prices for the remedies of Galen. Pfizer priced its drug, chlorpropamide (`Diabinese') at i~ cents per 250-milligram tablet, but since Pfizer's drug is somewhat more potent than tolbutamide, the daily cost may run somewhat less. The third drug, phenformin, is made by one of the smaller firms, U.S. Vitamin. Phenformin is still more potent than chlorpropamide, a ~o-mi1ligram tablet costing i 2 cents. The range of possible daily cost (minimum to maximum recommended dosage) is from 14 to ~6 cents for tolbutamide, from 7 to 30 cents for chiorpropamide, and from 12 to ~ cents for phenformin.54 The drug buyer seems to have 52 Testimony of Dr. Upjohn, ibid., Part 20, pp. 11037-9. ~ Report, op. cit., p. g~. ~ Hearings, op. cit., Part 20, p. 11120. PAGENO="0437" COMPETITIVE PROBLEMS IN THE DRUG INDUSTRY 1987 real price alternatives, for once, but because of the priority in appearance of the higher priced drugs, and their intensive advertising, sales seem to be directly related to the height of daily average cost. The Subcommittee staff computed the estimated production cost for one thousand 500-milligram tolbutamide tablets (twenty bottles of fifty each) on the basis of bulk powder transactions plus contract tableting and bottling charges, as in the case of prednisone. The production cost estimate is $6.86; the royalty (71 per cent of the selling price) is $6.24. The wholesale price of twenty bottles is $83.40 and the retail price is $139.00 Upjohn's production cost is therefore 8 per cent of the wholesale price; the production cost plus the royalty brings Upjohn's total estimated cost of production and royalties to 15.5 per cent of the wholesale price and to io.6 per cent of the retail price.55 Production cost estimates for the other two drugs are not available. The only bid on a Military Medical Supply Agency contract that is available for any such drug is Upjohn's successful bid of $4.00 for 18,432 bottles of fifty 500-milligram tablets. Since the wholesale cost for one bottle is $4.17, it can be seen that in the absence of lower-priced competition from smaller firms, a large order from the government, with its attendant cost savings, could be obtained at only a 4 per cent quantity discount.56 C. Tranquilizers Price competition among tranquilizers varies with the type of tranquilizer. There are two types, the `potent' type used chiefly in mental institutions, and the `mild' type used for non-hospitalized patients troubled with anxiety.57 In the potent group, the first drugs to be introduced were chiorpromazine and reserpine in 1954. Both drugs were priced at wholesale at $3.03 per bottle of fifty tablets. Later during the year CIBA introduced its brand of reserpine at a price for fifty tablets 25 per cent below its existing competitors for that drug (Squibb and Riker), $2.25. When Smith, Kline and French introduced prochiorperazine in 1956, they priced it at $3.03 per fifty tablets, the same price they charged for their chlorpromazine. When Wyeth introduced promazine hydrochloride in 1957, it did so at a price of $3.00 for a bottleof fifty tablets.58 In an interesting ~ Report, op. cit., p. 20. 56 Hearings, op. cit., Part 21, p. 11551. ~ As is true in most drug categories, much substitution among drugs is possible in most disease groups treated by tranquilizers. A 1959 study by the Pharmaceutical Manu- facturers Association shows that all seven of the major tranquilizers (both mild and potent) are prescribed in all five major disease groups requiring such therapy, ibid., Part i~, p. 10827. 58 Ibid., Part i6, p. 8887. PAGENO="0438" 1988 COMPETITIVE PROBLEMS IN THE DRUG INDUSTRY move in 1958, Wyeth changed the wholesale price of its drug pro- methazine hydrochloride, which was originally an antihistamine with drowsiness-inducing properties, from $4.45 per hundred 25-milli- gram tablets to $5.94, thus making it competitive with its recent tranciuilizer promazine hydrochloride.59 An antihistamine was thus made a tranquilizer in response to changing market demand by changing the advertising to stress what were previously considered side effects as the desirable end result, and by increasing the price in recognition of the increased marginal revenue possible in the more lucrative market for tranquilizers. Later potent tranquilizers were priced at about $3.00 for the equivalent dosage. Apparently reser- pine was considered to be a sufficiently distinct product that CIBA's 25 per cent lower price did not call for price reductions for other potent tranquilizers. CIBA, perhaps uncertain as to the validity of its patent (reserpine is the name given to the active substance isolated from the root of Rauwolfia Serpentina, a species of tropical plant, the medicinal properties of which had been known in India for centuries), issued licenses to nine other firms in the United States, eight of which were allowed to sell in bulk powder form. These firms made sales to smaller firms, which in turn proceeded to cut prices down to as much as 90 per cent below CIBA's levels. CIBA continued to advertise its brand of reserpine and its price held up; other major firms dropped their advertising of this pro- duct, by and large, and their prices were cut to about $1.91 per fifty tablets at wholesale.6° No other potent tranquilizer has ever experi- enced a cut in price. Carter was first in the mild tranquilizer market with mepro- bamate, which it has sold since 1955 at $2.60 per bottle of fifty 400-milligram capsules at wholesale. Carter licensed Wyeth to sell meprobamate under Wyeth's brand name, and Wyeth charged the same $2.60, although its royalty payment to Carter would be 13 cents, and its raw material~cost (purchased from Carter) would be 25 cents greater.6' Carter licensed its very profitable product only because it did not have the selling facilities which larger drug firms such as Wyeth had. Little competition to meprobamate has devel- oped. Phenaglycodol is sold by Lilly at the same price to the drug buyer as meprobamate. The computed cost for reserpine production is $2.48 per bottle of one thousand 25-milligram tablets.62 CIBA's wholesale price for this quantity is $39.50; other firms quote prices as low as Panray's $2.65. For CIBA, estimated production cost is 6.3 per cent of its 5~ Hearings, op. cit., Part i6, p. 9274. 60 Ibid., Part i6, pp. 9387-90. 61 Report, op. cit., p. i8. 62 Hearings, op. cit., Part i6, p. q436. PAGENO="0439" COMPETITIVE PROBLEMS IN THE DRUG INDUSTRY 1989 price; for Panray, 93.6 per cent. It must be the case, however, that the computed cost figure is much too high, for in competitive bids to the Military Medical Supply Agency Panray has won bids at as low as $0.51 per thousand tablets on a 4080-bottle order, and CIBA has failed to get contracts on bids as low as $0.52-I .3 per cent of the wholesale price. When asked how CIBA could afford to bid successfully at $o.6o in February 1959 Mr. T. F. D. Haines, presi- dent of American CIBA, testified: `. . . we didn't anything like re- cover our out-of-pocket costs . . . it was perhaps a mistake that we did that'.63 This makes it all the harder to explain CIBA's later and lower unsuccessful bids. (Raw materials costs for one thousand 25- milligram tablets should be about 30 cents.) The president of Mc- Kesson and Robbins later volunteered his firm's production costs for reserpine: $0.63 per thousand tablets, or 1.6 per cent of the whole- sale price.64 The computed production cost for meprobamate to Carter is $7.32 for twenty bottles of fifty 400-milligram tablets, or~ the basis of a 20,000-bottle order. For Wyeth, the cost would be $12.40 plus $2.60 royalty. (Carter agreed that the Subcommittee staff's cost estimate was correct for its own production.) Both Carter and Wyeth obtain $52.00 at wholesale for twenty bottles; a retail drug buyer would pay $108.40 for this amount. Carter's production cost is 14.1 per cent of the wholesale price, Wyeth's is 24.0 per cent, and with the royalty to Carter, 28.0 per cent. (Carter's production cost is only 6.8 per cent of the retail price.)65 Bids to the Military Medical Supply Agency from Carter and Wyeth ranged from $22.50 for five hundred 400-milligram tablets in February 1958 to $1 9.845 in February i 960, or from 90 to 79.4 per cent of the wholesale price for this amount. In the absence of any domestic competition, the Military Medical Supply Agency purchased abroad in June 1960, obtaining a low bid of $3.95 from a Danish firm, a bid only 29 cents above Carter's production costs, and about 8o per cent below the bids of Wyeth and Carter.66 Danish costs may be below American costs, but Carter could at least have met this bid had it seen fit to do so. D. Antibiotics Price competition in the antibiotics market is intense in unpaten- ted drugs, such as the earlier penicillins, and streptomycin, the patent 63 Ibid., Part i6, p. 9430. 64 House of Representatives Hearings, op. cit., p. i6. ~ Report, op. cit., p. i8. 66 Testimony of Rear Admiral W. L. Knickerbocker of the Military Medical Supply Agency, Hearings, op. cit., Part 24, pp. 13784-5, 53794. PAGENO="0440" 1990 COMPETITIVE PROBLEMS IN THE DRUG INDUSTRY for which was licensed by Merck to several other firms. Price com- petition is, however, considerably less in evidence for the later patented antibiotics. Table I shows how progressively more efficient production methods and continual competitive pressures consequent upon free TABLE I PENICILLIN, STREPTOMYCIN AND BROAD SPECTRUM ANTIBIOTICS PRICES, 1945-60 Date (year; month for price changes) Penicillin (io mil- lion units, Strepto- n~ycin (io grams, Broad spectrum antibiotics capsules) (i6 250-milligram Chlor- Chlor- Oxy- Tetra- Demet/~yl- chlor- bulk) bulk) tetra- amphe- tetra- cycline tetra- cycline nicol cycline (~ brands) cycline (iso mg.) S S S $ S S $ 1945 May 6o.oo 1946 May 36.00 ,6o.oo 1947 January 21.00 45.00 1948 July 13.00 10.24 1948 December 9.50 8.19 15.00 1949 February 9.50 8.19 10.00 1949 March 7.50 5.00 10.00 10.00 1950 February 4.75 3.15 8.oo 8.oo 1950 April 1950 May 1950 November 4.75 4.75 4.75 3.15 3.15 3.15 8.oo 6.oo 6.oo 8.oo 6.oo 6.oo 8.40 8.40 6.oo 1951 September 1951 November 3.75 2.50 3.15 3.15 6.oo 5.10 6.oo 5.10 5.10 5.10 1952 1.15 3.24 5.10 5.10 5.10 1953 .95 1.70 5.10 5.10 5.10 5.10 1954 .75 1.70 5.10 5.10 5.10 5.10 1955 1956 .44 .52 .90 .75 5.10 5.10 5.10 5.10 5.10 5.10 5.10 5.10 1957 1958 1959 1960 .79 .70 `28 .21 .88 .88 .38 ~36 5.10 5.10 5.10 5.10 5.10 5.10 5.10 5.10 5.10 5.10 5.10 5.10 5.10 5.10 5.10 5.10 5.10 5.10 Source: Federal Trade Commission, Economic Report on Antibiotics Manufacture, Washing- ton, 1958, Section II, pp. 164, 179, i8o, 192; Senate Hearings on Administered Prices in the Drug Industry, Part 24, p. 13664. entry made possible the precipitous decreases in the prices of the extremely important drugs penicillin and streptomycin. `Broad spectrum' antibiotics prices have acted very differently. Lederle quoted an initial price of $15.00 for sixteen 250-milligram capsules of chlortetracycline, possibly overestimating the inelasticity of demand. Within 2 months, Lederle reduced its price to $io.oo, at which price Parke, Davis entered the market in March i~. ~n February 1st, 1950, both sellers cut prices to $8.oo. In April 1950 a new rival, Pfizer, entered the market with oxytetracycline, selling PAGENO="0441" COMPETITIVE PROBLEMS IN THE DRUG INDUSTRY 1991 at $8.40, 5 per cent above its competitors. In May 1950 Lederle and Parke, Davis reduced prices to $6.Oo, but Pfizer did not meet the cut until November. On September 27th, 1951, Pfizer lowered prices to $~. 10, and by November ist Lederle and Parke, Davis had met the price cut. Tetracycline and demethylchlortetracycline came later, and were also priced at $~.io. Prices were then stable and rigid for almost 9 years, when Pfizer in August 1960 again assumed the role of price leader by announcing a i~ per cent `trade discount' adjustment which eventually became a i~ per cent price cut, to $4.34.67 (This price cut was initiated only i month before the Senate Hearings on antibiotics were scheduled to take place.) From 1948 to 1949, broad spectrum prices dropped by 33 per cent; from 1949 to 1951, by 49 per cent, and from ~ to August 1960 not at all. It is reasonable to suppose that since 1951 the costs of producing broad spectrum antibiotics have deClined by approximately as much as the cost of producing penicillin, the production methods employed being largely identical.68 A comparison of the decline of 90 per cent since November 1951 in penicillin prices, with the 15 per cent decline for broad spectrum antibiotics measures the effect of res- triction of entry in broad spectrum antibiotics. The Senate Subcommittee staff made production cost estimates for tetracycline and chloramphenicol, and in 1961, during the House of Representatives hearings, Senator Kefauver made public Bristol's actual production costs. Bristol incurred production costs of $1.67 per bottle of one hundred 250-milligram capsules (in comparison with $2.88 as estimated by the Senate Subcommittee staff in the Senate hearings). The price of such a bottle to the druggist is $30.60; to the consumer, $~i.oo. Bristol's'production cost is ~ per cent of the price to the druggist; the addition of royalties to Lederle and Pfizer of $2.15 paid per bottle brings the cost up to 12.5 per cent. (Bristol actually received an average of $25.27 per bottle for such sales in 1958; hence its production costs and royalties totalled ~i per cent of its average price received.) Production costs are about 28.7 per cent of the wholesale price for Upjohn, which purchases bulk tetracycline from Bristol; its total production costs and royalties are $9.30, or 30.4 per cent of the price to the druggist. For Pfizer, ~ Mr. Duncan, the manager of Lederle, construed Pfizer's action as vigorous and aggressive price competition. He testified that Pfizer had cut its prices on a Saturday in order to `steal a march on the industry'. This explanation apparently surprised the Subcommittee chairman: Senator Kefauver: `You mean after ten years they would suddenly steal a march on you?' Mr. Duncan: `Yes, senator. These things happen very rapidly. If you can get any kind of advantage on your competitors, you try to do so', ibid., Part 24, p. 13728. 68 Report, op. cit., p. 82. Chloramphenicol costs have probably declined by still more since Parke, Davis discovered a way of making this drug synthetically. PAGENO="0442" 1992 COMPETITIVE PROBLEMS IN THE DRUG INDUSTRY which holds the patent and pays no product royalties, and which reportedly has a lower cost production process, the total costs of production might perhaps be as low as $1.50. If Pfizer pays the same royalty as Bristol for the use of Lederle's process, its production costs plus royalties would be about $2.75. For Upjohn, Bristol and Pfizer, the ratio of production cost plus royalties to the price to the druggist might be about 30 per cent, i~ per cent and 9 per cent. The gap between estimated production cost and price for chlor- amphenicol is predictably handsome. One bulk powder transaction did take place, between Parke, Davis and an Italian firm, Farmitalia, at $30.00 per kilogram. Using this price for the raw material, and Upjohn's actual costs for capsuling, finishing, and packaging, a cost of $1.52 for a bottle of one hundred 250-milligram capsules is obtained. This is 5.0 per cent of the price to the druggist of $30.60, and 3.0 per cent of the retail price to the consumer.69 The bids received by the Military Supply Agency reflected the status of market competition. For penicillin, competitive bidding was the rule, the price for 250,000-unit potassium penicillin G tablets falling from $1.29 per hundred in April 1957 to $o.6~ in June 1960. As in the case of prednisone, small firms selling penicillin regularly charge much lower prices than large ones. For this particular pro- duct, the smallest firm (sales of less than $ ioo,ooo) had the second lowest price ($3.30); the second smallest firm (sales of less than $i ,ooo,ooo) had the lowest price; and three of the largest firms (sales of $i~o to $250 million) had the highest price ($12.00).7° For broad spectrum antibiotics, the picture is entirely different, as can be seen from Table II. For this entire period the price to the druggist on all these drugs was $30.60, and the price to the consumer, $~i.oo. Lederle and Parke, Davis began the period of rigid prices to the trade by quoting a ~o per cent discount to the Armed Services Medical Procurement Agency (the predecessor to the Military Medical Supply Agency); Pfizer allowed about a 51 per cent discount off the druggist's price. Apparently the three firms recognized the high degree of sub- stitutability among their drugs, for prices fell for each seller with 69 Harry Loynd, president of Parke, Davis, objected at length that the Farmitalia price was much lower than American production costs. In order properly to appraise the reliability of Loynd's assertions, it is advisable to read the whole of his quite remarkable testimony; nevertheless, a consideration of the import agreement is instructive. Farmitalia agreed to supply Parke, Davis with up to 30,000 kilograms ofchloramphenicol during i 960 at ~30 per kilogram. ThroughJuly iith, however, only 6ooo kilograms had been imported. If this price were such a bargain, why did Parke, Davis not take greater advantage of it? The suspicion is that Parke, Davis might be able to make it for less in Detroit. However, even if the price were twice as high, it would only increase the cost per i oo-capsule bottle to $2.31, or 7.5 per cent of the sales price, Hearings, op. cit., Part 24, pp. 1397 i-8. 70 Report, op. cit., p. 84. PAGENO="0443" COMPETITIVE PROBLEMS IN~ THE DRUG INDUSTRY 1993 regard to its previous price until April 1956 when Parke, Davis repeated its previous bid of $12.50. At that time Lederle's price was $ii.oo on the first sale of tetracycline hydrochloride, pricing it at the same level as its chiortetracycline. Lederle was low bidder, and was much surprised to find that Pfizer and Squibb had both bid $19.58, and Bristol, $ I8.97.~' For the next 2 years Lederle always bid BROAD SPECTRUM AGENCY AND TO * ANT THE TABLE II IBIOTIC PRICES TO THE ARMED SERVICES MEDICAL PROCUREMENT MILITARY MEDICAL SUPPLY AGENCY, NOVEMBER 1951-JUNE 1960 (ioo capsules, ~o milligrams) Date Chlortetra- Chioram- ~ycline phenicol Oxytetra- cycline Tetracycline- hydrochloride (ASMPA) 1951 November 1952 March 1953 June 5954 January 1954 May 1955 March 1956 April 1956 October 15.30 15.30 15.00 12.50 ~ 12.00 11.00 52.50 15.00 12.84 11.47 10.97 11.00 (MMSA) 1957 February 1958 February 1958 March 5958 April 1958 June 1958 November 5959 June 1959 August 1959 December 1960 May 1960 June 11.00 12.50 ~ 11.25 11.00 11.25 11.25 11.25 ~ 10.97 10.75 10.11 17.24 19.19 17.24 17.15 14.36 17.15* 8.15 6,6 5.62 * Low bid-not accepted. Source: Federal Trade Commission, Economic Report on Antibiotics Manufacture, Washing- ton, 1958, p. 194 (i~~i -~6 data); Senate Hearings on Administered Prices in the Drug Industry, Part 24, pp. 13779-82; 13791-2 (1957-60 data). $19.58. On the second bid for tetracycline hydrochloride, Pfizer met Lederle's previous bid of $11.00, but Lederle and Squibb were both at $19.58. Pfizer, like Lederle, then learned that it was not necessary to make `ridiculously low' bids. Its prices on the next two bids went up to $17.24 and $19.19. In April 1958 Parke, Davis cut its chioramphenicol price by 10 per cent, to $i 1.25, after a con- ference with the Agency purchasing officer on the subject of its 71 Duncan of Lederle conceded that he had made a mistake in simply pricing tetra- cycline hydrochloride at the same level as chiortetracycline. He characterized the pressures of competitive bidding in terms of carelessness: `In other words, one gets a little sloppy in bidding for this kind of business. You sometimes simply bid a ridiculously low figure', ibid., Part 24, p. 13690. PAGENO="0444" 1994 COMPETITIVE PROBLEMS IN THE DRUG INDUSTRY rigid price.72 In June 1958 the tetracycline hydrochloride price dropped back to $17.24, where it stayed for three consecutive awards. In November Squibb cut its bid to $17.15, to undercut the others, all of which bid $17.24. InJune 1959 Lederle bid $17.24, Bristol and Squibb bid $17.15, but Pfizer was low bidder at $14.36. Pfizer's bid is surprisingly low,73 and it encouraged Military Medical Supply Agency officials to believe that prices had at last fallen. Such expectations were disappointed, for in August ~ three bids were received at $17.15. Negotiations failed to shade the price, and the contract was cancelled because of the Agency's judgment that the price was unreasonably high. Foreign bids were received, and an Italian firm was awarded three successive contracts at $8.15, $6.i6 and $5.62 in 1959 and 1960. Rather than lower its price to compete with foreign bidders, Pfizer first protested the Italian procurements (the Comptroller General rejected Pfizer's contentions as being without merit) and then attempted unsuccess- fully to buy out the Italian firm.74 Had Pfizer chosen to engage in price competition, it could in all probability have undercut the Italian firm on all three bids. The case of tetracycline hydrochloride prices is rather puzzling in that the product is a close substitute for three other drugs sold at lower prices to the government. Perhaps the industry wished to receive higher prices for the new drug, and a certain amount of trial and error bidding was necessary to facilitate the learning process. It is interesting to note that some sort of division of the total tetra- cyclines market in total dollar sales was accomplished-even at varying prices-during the period October 1956 through October 1959, or from the first award until the date of the first foreign pro- curement. During this period Pfizer (the patent holder) received 46.6 per cent of all contracts in dollar terms; Lederle received i 7.8 per cent; Bristol, 17.6 per cent; Squibb, 17.5 per cent; and Upjohn, only 0.5 per cent. This may indicate nothing more than the varying 72 Ibid., Part 24, p. 13782. ~` Senator Kefauver expressed surprise that Lederle would continue to bid $17.24 after the last sale had been awarded at S 57.55. Mr. Duncan replied that he was against quoting lower bids because'.. . we had been doing that, and all that happened was that the price just went lower time after time until it got down to an uneconomic level, and we decided that we would simply establish that Federal Government price, and stick to it, which is what we did', ibid., Part 24, p. 13694. It maybe doubted that `uneconomic levels' implied losses. On the basis of the cost data obtained by the Subcommittee, and assuming that Pfizer paid the same royalty on Lederle's process that Bristol did, that Lederle's and Pfizer's production costs were somewhat lower than Bristol's, and that Squibb's costs were the same as Upjohn's, Lederle's production costs and royalties might be $3.60; Pfizer's, S4.oo; Bristol's, S5.o3, Squibb's and Upjohn's, S9.3o. This might aid in ex- plaining the lower initial bids of Lederle and Pfizer. 74 Testimony of Commander Arnold Wales, U.S. Navy Supply Corps, ibid., Part 24, p. 13815. PAGENO="0445" COMPETITIVE PROBLEMS IN THE DRUG INDUSTRY 1995 incidence of excess capacity or excess inventories, but does not reflect either relative production or relative sales. In 1958, Bristol produced 36 per cent, Lederle 33, and Pfizer 31. Sales are not available, but Bristol sold about ode-third of its output to Upjohn and another third to Squibb,75 so that relative sales might be: Lederle, 33 per cent; Pfizer, 3'; Squibb, i ~; Bristol, I 2~ and Upjohn, i i, unless Pfizer purchased significant quantities from Lederle, as it is entitled to do by agreement. It seems likely that, although prices fluctuated, real price competition never entered this market until foreign firms, not encumbered by restrictive patent agreements, were permitted to bid.76 IV. CONCLUSIONS AND POLICY RECOMMENDATIONS Evidence has been presented to, show that effective price com- petition among ethical drugs is seriously limited by the patent privilege. Holders of patents (or sometimes merely of patent applica- tions) may legally exercise restrictions on output and maintain prices at levels that are extremely high relative to production costs. The resulting gross profit margins are employed in large measure to finance enormous advertising and sales promotion campaigns which contribute materially to the already grave imperfections of market information. By this means, small sellers of generic name drugs are deprived of the physician's attention, and cannot obtain any significant share of the prescription market, even though they may be selling at prices which are a small fraction of their larger rivals'. This applies not only to products protected by patents or patent applications, but to virtually all advertised drugs. Monopoly profits from the sale of patented drugs thus finance advertising campaigns which extend monopoly power into other drug markets which may not be protected by patents. Drug firm monopoly and oligopoly could perhaps be rather readily supplanted by workable competition if two simple but radical reforms were effected in the institutional structure of the drug market: the abolition of the patent privilege as it applies to drug pro- ducts, and the expansion of the powers of the Food and Drug ~5 Data supplied by Bristol, ibid., Part 24, p. 13907. 76 Admiral Knickerbocker offered some interesting testimony which may tend to illuminate the chiaroscuro pattern of competition in the bid market from a somewhat different perspective. The decision of the Military Medical Supply Agency to accept foreign bids was not reached until after a meeting with Pfizer representatives, during which, Admiral Knickerbocker testified, `.. . Mr. Cooney [Pfizer's sales representative], in rather an unguarded moment, made the statement that the price of tetracycline would stay where it was until Pfizer did something about it', ibid., Part 24, p. 13819. Had the network of patent control been absolutely worldwide, this alleged `unguarded moment' might not have cost Pfizer any sales. PAGENO="0446" 1996 COMPETITIVE PROBLEMS IN THE DRUG INDUSTRY Administration to insure fully adequate drug plant inspection and control of drug advertising. It is something of an anomaly that the United States patent laws permit drug product patents; about two-thirds of the countries in the world prohibit them on grounds of public policy, and all but three of those which do permit such patents have provisions for compulsory licensing. Drug product patents should be abolished; drug process patents should be subject to compulsory licensing at reasonable royalties. This may or may not reduce the volume of research done by private drug firms. In turn, a possible diversion of drug research effort from private to public channels may or may not improve the efficiency of resource allocation in this sector. In many foreign countries, drug research has been immensely productive in the absence of drug patents. In the United States, patent protection has allowed prices to be maintained at very high levels and has detrimentally affected the nature of research. With prices enor- mously above costs (defined as raw material cost plus production cost plus a competitive rate of profit on the investment in productive facilities), the pressure to obtain more volume is inevitably diverted into lavish selling expenses, and into the `research' of molecular manipulation.77 Basic research by drug firms may be of question- able productivity, but the high salaries paid absorb all too large a fraction of those very scarce human resources qualified to engage in basic biochemical and pharmacological research. It is common knowledge in the industry that each major firm's research programs duplicate those of their rivals; witness the near-simultaneous dis- covery of many antibiotics and corticosteroid hormones by two or more firms. The total productivity of drug research efforts would arguably be increased by a partial shift from private to public and university channels. The Food and Drug Administration should be given sufficient additional funds to allow adequate inspection of all drug producers, large and small. It should also be given the authority to regulate all drug advertising and labelling, with a view toward eliminating brand names in favor of identification by generic name plus the name of the seller. Such reforms, properly implemented, might for the first time 77 Abundant evidence was presented during the Senate Hearings to show that much drug product `research' is misnamed, and could more appropriately be referred to as product development and product promotion. Dr. A. D. Console, former medical director of Squibb, testified to this effect, adding: `I think the majority of it [drug firm research] is in that category. . . with many of these products, it is clear while they are on the drawing board that they promise no utility. They promise sales', ibid., Part i8, p. 10380. See also the testimony of Dr. H. J. Weinstein, former medical director of Pfizer, ibid., Part i8, pp. 10243-4 PAGENO="0447" COMPETITIVE PROBLEMS IN THE DRUG INDUSTRY 1997 make the ethical drug market workably competitive. The elimination of patents would (in the absence of economies of scale in production) lead to the entry of new firms producing under conditions of approxi- mately constant costs. The inelasticity of demand would thus be offset by highly elastic supply. The existing economies of scale in selling would, in time, be eliminated by advertising reforms and the disappearance of those monopoly profits which have motivated and financed extravagant selling efforts. The advertising reforms would also remedy the imperfection of market information, allowing physicians to become aware of the precise nature of the price and product alternatives, and disparagement efforts would wane as the funds necessary to finance them dwindled due to increasing com- petitive pressure on prices, and as the manifest adequacy of FDA inspection rendered the substance of such contentions obviously specious. S RICE UNIVERSITY S HOUSTON, TEXAS PAGENO="0448" 1998 COMPETITIVE PROBLEMS IN THE DRUG INDUSTRY THE FORTUNES OF ECONOMIC REFORM LEGISLATION: THE CASE OF THE DRUG AMENDMENTS ACT OF 1962 (By Henry Steele 1) I. INTRODUCTION On April 12, 1961, the late Senator Estes Kefauver introduced a bill (S. 1552, 87th Congress, first session) to amend the antitrust and food and drug laws with regard to the prescription drugs industry. This bill, the outcome of extensive hearings held by the Subcommittee on Antitrust and Monopoly of the Senate Judiciary Committee in 1959-1962 regarding administered prices in the drug industry, embodied provisions to prevent monopolistic abuses in drug marketing. The Report issued by the Subcommittee established the existence of considerable monopoly and oligopoly power on the part of major drug firms, and indicated the extent to which such market power was obtained by virtue of patent privi- leges, excessive advertising, and other marketing practices which tended to augment the imperfection of market information in the interests of the larger firms.2 IT. ANALYSIS OF DRUG MARKETING PRACTICES An economist would briefly summarize the drug marketing situation as fol- lows: demand for many prescription drugs is almost perfectly inelastic. Supply can be restricted by drug firms which secure product patents and refuse to li- cense them. The resulting enormous profit margins on such products (factory costs may be less than 10 percent of wholesale price 3) stimulates entry, which take the form of imitative research activity on the part of other drug firms, aimed at devising a substitute for the patented drug which they can patent and advertise as superior to the original. Rivalry is thus diverted from price com- petition to product differentiation, as rival patented drugs attempt to take over the prescription market for a given group of disorders by intensive advertise- ment to doctors (the only agents qualified to prescribe) through direct mail, itinerant salesman (detailmen), and the sponsoring of medical conventions. High unit profit margins can finance large (and often imitative, duplicative, and wasteful) research programs, overpowering advertising campaigns, and still yield very high profits on investment. Since costs are low, and economies of large scale production apparently unimportant, small firms might be able to compete with large firms on a cost basis, but they suffer from two disabilities. Being unable to finance sales campaigns, their products do not come to the at- tention of the prescribing physician. Furthermore, the detailmen employed by the major firms make a practice of disparaging low-priced drugs, inducing the physician to equate low price with low quality. At best, the physician may not be very price-conscious, since he does not pay for the drugs he prescribes. The situation is aggravated by weaknesses in patent laws and practices, and in the regulations governing drug advertising. Patent protection is absolute as regards a given compound, but weak as regards slight variations from that compound. The game of "molecular manipulation" is a popular one with drug firm research personnel: the goal is to devise a derivate compound which has the same therapeutic effect as some primary patented compound, but which by virtue of its marginally different chemical structure can be patented as a different drug and advertised as a "new and superior" healing agent.4 Patents can also be obtained for naturally occurring substances if never previously isolated, for processes occurring in nature, such as fermentation, and even for combinations of existing patented drugs. But since most major drug firms have research programs which substantially duplicate each other, near-simultaneous discovery of the same drug is a not infrequent occurrence. When several patent applications covering the same compound are received by the patent office, an interference is declared, and the parties attempt to assist in resolving the question of priority to the satisfaction of the patent office. There is, however, 1 The author is assistant professor of economics at Rice University. 2 Study of Administered Prices in Tue Drug Industry, Report of the Subcommittee on Antitrust and Monopoly of the Senate Judiciary Committee, Pursuant to Senate Resolution 52, Eighty-Seventh Congress, First Session, Washington, D.C., Government Printing Office, 1961. 3lbid.,pp.15-16. 4Ibid., p. 16. PAGENO="0449" COMPETITIVE PROBLEMS IN THE DRUG INDUSTRY 1999 no barrier to private settlement of interferences by agreements among the parties themselves, involving concession of the patent to one firm in return for granting licensing rights and perhaps other privileges to the other firms. The patent office rather welcomes such private settlements in view of the pressure of work on its staff, but the antitrust division has been critical of the nature of many private interference settlements.5 Advertising abuses relate chiefly to drug nomenclature and to the quantity and quality of sales efforts. Four hundred ~ or more new drugs are marketed each year in the United States. Each new drug should be given a generic name to identify it; individual firms marketing the drug may confer upon it their own brand names. If a physician prescribes a drug by generic name, the pharmacist may fill the prescription with any brand of the drug; if the prescription is by brand name, only that firm's brand may be dispensed. Evidence indicates that generic names are designed to minimize use (excessive length, complexity, and clumsiness) while brand names are brief and memorable. Regulations require that generic names be displayed in all advertisements, by they are often obscured by the use of microscopic type face, by being concealed in obscure places in the advertisements, and are sometime simply omitted.6 Some drugs have no generic name; others have more than one. This deliverately cultivated confusion is evidently intended to suppress use of generic names in favor of brand names.7 By this means, trademarks are made to supplement patents as monopolistic devices. The same patented drug may be sold by ten different licensees under ten different highly advertised brand names, and each licensee through sales efforts may succeed in differentiating in the mind of the physician this physically homogeneous drug. In those rare cases where there is no patent protection, generically-named products sold by small firms may be priced at ten per cent or less of the price of the majOr drug firms, but price competition is rendered impossible unless prescriptions are written generically. However, the detailmen have been very successful in their disparagement campaigns; surveys show that almost 80 per cent of drug prescriptions are written by use of brand names.8 The quantity of drug advertisements is overwhelming in itself, includ- ing not only propaganda but also free drug samples, trinkets (toy urinals, chinese dolls, head cushions, etc.) and gifts such as golf balls engraved with the recipient's name. But the quality of advertising claims is the greatest obstacle to the enlightened practice of~ medicine. Dozens of examples of mis- representation were unearthed at the Senate hearings; one example must suf- fice: a drug firm mailed to physicians advertising copy showing X-ray photos clearly designed to imply a dramatic recovery in a patient's condition before and after the use of the advertised drug. Upon inquiry to the firm's medical director, it developed that the two X-rays were of entirely different persons with qualitatively different disease conditions, and that neither had ever used the drug being advertised.9 Ill. INDICATED REFORMS To an economist, the abuses in drugs seem amenable to relatively simple re- forms. Drug product patents should be abolished; drug process patents should be subject to compulsory licensing at reasonable royalties. Generic names should be simplified, the use of brand names should be eliminated, and firms should be required to advertise and sell their products under labels giving the generic name of the drug, followed by the name of the firm, e.g., chloramphenicol-Parke, Davis instead of "Chloromycetin." Food and Drug Administration drug plant inspection authority and funds should be increased in order to guarantee the safety of all drugs on the market, rendering specious all disparagement cam- paigns. Control of the quality of drug advertising should be made truly effective. If rigorously enforced, these reforms should suffice. Absence of patent protec- tion and mandatory use of generic names would allow price competition between large and small firms; excessive profits would disappear, and with them would disappear the ability to carry out wasteful and duplicative "research" programs, 5lbid., p.48. 6 study of Administered Prices, op. cit., p. 234. Firms may, however, on occasion find generic names of use as when the necessity arises for sending out circulars to warm physicians about side effects. Hearings on S. 1552, Part 5, p. 2938. 8lbid., Part 18, pp. 10481-10482. Ibid., Part 7, pp. 3301-3310. 81-280 0-68-pt. 5-29 PAGENO="0450" 2000 COMPETITIVE PROBLEMS IN THE DRUG INDUSTRY and to advertise so intensively as to prevent the products of smaller firms from coming to the attention of the physician. Much of drug research would thereby be shifted from private to public and university channels, with a probable in- crease in the efficiency of such efforts, since the patent-induced incentives to engage in duplicative research, molecular manipulation, and promote unpromis- ing drugs in the interests of their sales potential rather than their efficacy, would be removed. IV. S. 1552: FROM BILL TO ACT A. OriginaZ provisions Senator Kefauver, however, did not contemplate so sweeping a reform of existing practices, perhaps in view of the political difficulties. His original bill did, nevertheless, incorporate a number of valuable reforms, particularly of patents and drug nomenclature. The salient provisions of the bill may be listed as follows: (1) Drug patents shall take effect as of the date of the new drug applica- tion (or in the case of drugs for which no new drug application is necessary, the date for filing the patent application) and shall grant unconditional monopoly power for a period of only three years, after which such patents are subject to compulsory licensings at a royalty not exceeding 8 per cent of the sales price. (2) No patents are to be issued for compounds merely embodying molecu- lar manipulations of existing drugs, or combinations of existing drugs, unless the Secretary of the Department of Health, Education and Welfare deter- mines that such a drug has a significantly greater therapeutic effect than the unmodified drug, or of the combined drugs when taken separately. (3) The Sherman Act should prohibit private drug patent interference settlements or any other private arrangements whereby parties concede patent priority claims in return for royalty-splitting agreements, differen- tially favorable royalty rates for participants relative to third parties, or limitation of licensing to parties to the agreement. (4) The Food and Drug Administration should be given authority to establish generic names of drugs, with a view toward simplification and increased use of such names. (5) The Food and Drug Administration shall pass upon the efficacy as well as safety of drugs. (6) The Secretary of the Department of Health, Education and Welfare should publish and distribute to physicians and hospitals copies of drug firm package inserts which describe the action of drugs and give data on dosage, contraindications, and side effects. (At present, such inserts go by only to pharmacists, not physicians.) (7) All drug firms must be licensed by the Secretary of the Department of Health, Education and Welfare, and must submit to inspections of plant and equipment. (Plant inspection regulations are also to be made more adequate.' (8) All drug advertising shall include (a) the generic name in type face as large as the brand name, (b) a statement of the conditions for which the drug is an effective treatment, and (c) a statement of all side effects and other warnings. The reforms contemplated are rather modest. The provision for making patents effective as of the date of filing a new drug application is intended to eliminate any incentive for delay on the part of the applicant in patent proceedings, par- ticularly in negotiating interferences in the interest of prolonging effective patent protection, not only for the successful applicant, but also for the entire group of interim licensees. The three-year term of absolute patent monopoly is intended to allow the successful firm to recoup its research costs during an initial monop- oly period of high prices and no rivals; the compulsory licensing provision dur- ing the next 14 years is tempered by allowing the patent holder to charge an S per cent royalty, a very high royalty by drug industry standards, the usual rates being between two and six per cent. The prohibition of drug combination or molecular manipulation patents in the absence of evidence of superior efficacy is a laudable (although administratively difficult) attempt to halt wasteful du- plicative research efforts. The proposed Sherman Act amendment is rather strict in that it entirely pro- hibits private drug patent interference settlements, by making it illegal to with- draw any pending drug patent application. This is desirable in that private settle- PAGENO="0451" COMPETITIVE PROBLEMS IN THE DRUG INDUSTRY 2001 ments in effect transfer the determination~ of priority of invention from the Pat- ent Office to the rival parties themselves. Such private agreements often contain stipulations which restrain trade. One party is awarded the patent, and as a condition of surrendering their claims, the other parties are guaranteed licenses, and may even begin to pay royalties before the patent is issued. The agreements usually contain restrictive provisions under which the parties agree not to sell to outsiders, not to sell the drug in bulk powder form (which might get into the hands of non-licensed dealers who could tablet and bottle it and sell it at low prices), and to accept other limitations on their marketing practices. In addition, all of the parties to such agreements will almost invariably sell at identical prices.10 Most of `the other proposed reforms would alleviate abuses in drug marketing. Although the use of brand names is not outlawed, reforms in generic names would be a step in the right direction. The requirement that the FDA pass on the efficacy `as well as on the safety of drugs is a very reasonable one, but one which was bitterly fought by some drug firms during the original hearings on administered prices. In the absence of such a requirement, drugs which are not clearly harmful, and for which various physicians have written "testimon- ials" supporting a new drug application, may be allowed on the market, at which point the market success of the drug depends (`at least for a while) more on the skill of the sales department than on the intrinsic therapeutic merits of the compouii~L Several physicians testified at the hearings that not a few drugs cur- rently on the market are absolutely useless.11 The sixth provision is aimed at one of the many paradoxes in `drug marketing. Drug makers are required by FDA regulations to print up a leaflet describing the uses, dosages, and side effects of each prescription drug, *but the leaflet need only be included in the drug package, which goes to the pharmacist, who, having no use for it, throws it away. The physician, to whom such information is of the most vital importance, has no ready access to this information, but must routinely rely on advertisements and detailmen. Distribution of all such material to physicians is obviously an imperative necessity. Finally, the provisions for licensing all drug makers and for requiring more adequate inspection of drug plants is intended to insure the quality of all drugs, generic and brand name, and thus hopefully to reduce the effectiveness of generic drug disparagement an'd increase the physician's willing- ness to prescribe by generic name. B. Senate hearings on 5. 1552 The hearings on S. 1552 were held between July 19(11 and February 19(12. The battle lines were rather tightly drawn up on most issues.12 In the medical pro- j~ession, the American Medical Association (AMA) betrayed dedicated opposition to every section of the bill upon which they took an official position; on the other hand, of the eleven medical educators who appeared either individually, or as representatives of such groups as the American Public Health Association, ten favored the proposed reforms and an eleventh expressed opposition to the patent provisions, an area perhaps not strictly within his professional competence. Other physicians appeared in their capacities as administrators `of hospitals or health insurance plans, and indicated,' general approval. Ten patent attorneys and other patent spokesmen testified; nine of them, including representatives of the American Bar Association, The National Association of Manufacturers, the American Patent Law Association, etc., found little or no virtue in any `part of the `bill. The tenth, an attorney from a small town in a mid-western state, and representing only himself, regisitered substantial agreement with the aims of the bill. The Pharmaceutical Manufacturers Association opposed most of the bill's provisions, but showed limited agreement in some areas; other drug industry witnesses were not nearly so favorable. Testimony was given by two economics professors. One appeared at the request of the Subcommittee as an expert on the economics of patents, and gave testi- mony to the effect that the patent provisions of S. 1552 would probably not cause 10 Report on 5. 1552 op. cit., p. 45. 11 See, for example, .tTeariflg8, op. cit., Part 1, p. 285. 12 Battle lines were also drawn up within the Subcommittee itself, the majority (often represented by Kefauver alone) In favor of reform, and the minority (represented by Dirksen and Hruska) apparently concerned with obstruction of the proceedings and harrassment of witnesses favorable to the bill. Hruska in particular was adept at the latter, not scrupling to Insinuate Communist leanings in those with whom he disagreed. See Hearings, op. cit., Part 3, pp. 1410-1411. PAGENO="0452" 2002 COMPETITIVE PROBLEMS IN THE DRUG INDUSTRY the ruin of the drug industry if adopted; indeed, he saw- many faults in the present patent system, and suspected it w-ould be possible to increase the effi- ciency of drug research by transferring some of it from private to public auspices. The other was retained by counsel for the Pharmaceutical l\ianufac- turers Association, and he could find no evidence of any need for Kefauver's bill. Testimony was also obtained from a number of labor and consumer groups. Representatives from the AFL-CIO, the UAW, and the Internationa.1 Union of Electrical, Radio and Machine Workers supported the bill, as did spokesmen for Consumers Union, the National Consumers League, and the Cooperative League of the United States. Two representatives of retail Pharmacy appeared. One spoke for the American Pharmaceutical Association, a group of pharmacists. He articulated a viewpoint consonant w-ith that of the drug makers, and expressed great concern that the government not take the unwise step of distributing drug information to physicians. Another pharmacist, a former teacher and state phar- macy and drug law enforcement officer, found considerable merit in many of the reforms. The last parties to testify were advertising agency representatives who predictably saw no apparent need for any advertising reforms. It is likely that the most influential testimony w-as that given by the American Medical Association, the various patent law- groups, and the Pharmaceutical Manufacturers Association. It is instructive briefly to review the character and merits of the testimony given by these groups. 1. Testimony of the American Medical Association The AMA made no recommendations in regard to the antitrust law amend- ments, but took the view that none of the other proposed reforms were defen- sible, including the requirement that the FDA pass on the efficacy of drugs. Their spokesman, Dr. Hugh Hussey, recognized the need for certain improve- ments in regard to drug nomenclature and physician information. The AMA preferred, however, to carry out the reforms itself, in cooperation with the drug industry, but with no participation by any public bodies.n One may be for- given for entertaining the view- that the AMA position is simply a defense of the revenues it obtains from drug firms for advertisements in its journals. The roots of AMA opposition are more ramified and complex, but it is easy to trace the influence of advertising revenues, as seen against the background of prior AMA actions. First, the AMA reform program was adopted only five weeks prior to the sched- uled appearance of its representatives before the Subcommittee, and the imple- mentation of the program was to be gradual, extending over two or more years. Five years previously, however, a similar reform program in response to similar legislative demands, was outlined by the AMA, proposing cooperation between itself and the drug industry to control misrepresentation in advertising, but it entirely failed of implementation.~ Second, the AMA has become increasingly dependent upon drug advertising for its own financing. In 1949, medical journal advertising revenues comprised about 31 per cent of all AMA revenues: in 1955, about 44 per cent; and in 1900, a little over 50 per cent. Total advertising rev- enues, however, been augmented by royalties received from the leasing of the rights to use mailing lists of physicians.15 The sums received have increased -from small amounts to about 5.6 per cent of total revenues by 1960. Hence in 1960 about 56 per cent of AMA revenues came from drug firm advertising efforts. Third, the AMA in recent years has become increasingly permissive in its attitudes toward advertising standards, and it is likely that the increase in its advertising revenues is in good part attributable to this. The period of increasing leniency coincided with that of increasing advertising revenues. The measures by which this more lenient policy was evolved, or from which it can be inferred, may best be described and interpreted in chronological order. 13Ib Part 1, PP. 47-49. 14 Ibid., Part 1, p. 341. Testimony of Dr. Allan M. Butler, Professor Emeritus at Harvard University. 15 Obviously a royalty on direct mail advertisements ($2 per thousand mailings) creates - a direct financial interest in maximizing the volume of such traffic, and is hence undesirable in even greater degree than policies to increase advertising in the AMA journals; There is evidence that advertising standards in AMA journals are still higher than those in most (but not all) other medical journals, but there is no AMA control over the quality of direct mailings. The royalty income feature of AMA finances tends to justify in part the accusation of James Carey, president of the International Union of Electrical, Radio, and Machine Workers, that "The AMA, in our opinion, is just a business." (Ibid. Part 5, p. 2731.) It is ironic that, while most students of the AMA compare it to a trade union, a leading trade union spokesman sees It as a business. PAGENO="0453" COMPETITIVE PROBLEMS IN THE DRUG INDUSTRY 2003 In 1950, a rule (adopted in 1905) that only the inventor of a new drug could use his brand name of the drug in advertising in the AMA journals, was dropped. In 1952 the publication of a handbook, Useful Drugs (issued periodically since 1917) was discontinued. (The 1961 Ai\'IA proposals contemplated the publication of a similar book which, however, would include all drugs and would thus lack the discrimination between useful and less than useful drugs characteristic of the earlier publication.) In 1955, major advertising policy changes were made in response to a survey by Ben Gaffin and Associates addressed to the problem of increasing AMA advertising revenues. Before 1955, the AMA Council on Drugs, an expert technical body, had effective control over advertising. Only Council- approved drugs could be advertised. Approval required the obtaining of a "Seal of Acceptance" which was granted only after the drug firm had submitted to the Council all the data it requested, including both favorable and unfavorable reports. (By way of contrast, at that time there was no requirement that all test data, favorable and unfavorable, be sent to the FDA with new drug appli- cations.) The Council on occasion inspected drug factories. It exercised some con- trol over generic names, since only generic names approved by the Council could be used in AMA journal advertising. The Council examined and passed on all advertising. In 1952, Ben Gaffin and Associates began a survey to determine how AMA journals could get more advertising. It noted that AMA advertising had in- creased only 3 per cent since 1948, while other medical journals had increases of 40 per cent. The survey indicated that while physicians and small drug firms thought well of AMA advertising control policies, large firms were critical, par- ticularly of the "Seal of Acceptance" program. The Gaffin study recommended that the Seal of Acceptance program be dropped, and the advertising controls be liberalized.16 In 1955, several major changes were made in advertising policy. Advertising control was taken out of the hands of the Council on Drugs. The Seal program (in effect in that form since 1929) was abolished. The Council lost its influence over generic names by this means, and also its power to elicit unfavorable as well as favorable evidence on drugs. The Council lost all control over advertis- ing.17 While rules requiring generic name advertising were being abolished, edi- torial changes were made which drew increased attention to brand names.18 It is instructive to note that the 1961 reform proposals of the AMA did not include reinstatement of the control over drug advertising on the part of the Council on Drugs. Indeed, the preparing its presentation for the hearing, it did not even seek the advice of the Council, its own expert advisory body on the subject matter to which the hearings were addressed. This fact was testified to with dismay by more than one member of the Council.19 Had the Council been influential in drafting the AMA's testimony, it is hardly conceivable that the AMA would have gone on record as opposing the requirement that a drug need be efficacious in order to merit a new drug permit. During 1955 a microbiological laboratory, established in 1949 to establish purity standards for antibiotics and other drugs, was abandoned. And in 1959 a chemical laboratory, founded in 1906 to formulate and develope standards for drugs, was abandoned because of overwork and lack of finances.20 16 Ibid., Part 1, PP. 102-103; Part 2, pp. 490ff. Part II of the survey includes the following: ". . . while the possibility of increasing advertising revenue by several million dollars per year is a good motive for putting into effect the information gained from these two studies, there is an even more important reason for so doing. . . . the AMA has an opportunity to assume leadership in improving some $130 million worth of medical adver- tising per year." It thus becomes apparent that it was the duty of the A?~tA to enrich itself. 17 Mr. Stetler of the AMA denied that these changes had occurred in response to Gaffin's survey. Gaffin himself was convinced otherwise. In another drug survey, he links his recommendations and the AMA's actions: "The survey of pharmaceutical advertisers played a part in bringing about a number of policy changes . . . [including] the eventual dropping of the 58-year old council seal of acceptance program." Ibid., Part 1, p. 125. If the AMA disagreed with Gaffin's diagnosis, they took no disciplinary action. Gaffin continued to conduct their surveys throughout the 1950's. 18 Dr. Hussey volunteered that "Trade names, formerly listed at the end of monographs, also have been placed immediately after the nonproprietary titles of monographs and on the front page of the Journal to more readily catch the eye of Journal readers." Ibid., Part 1, p. 106. 19 Ibid., Part 1, pp. 216, 375. 20 The Council of Drugs also suffered from insufficient finances, its appropriations in- creasing from $135,000 in 1950 to $156,000 in 1954 (the last year of active Council adver- tising control) and declining to $75,000 in 1960. Curiously, the public relations department budget increased prodigiously, from $102,000 in 1950 to $494,000 in 1958. Ibid., Part 1, pp. 126-127. PAGENO="0454" 2004 COMPETITIVE PROBLEMS IN THE DRUG INDUSTRY It seems safe to conclude that on the basis of the recent record, the role of the AMA in contributing to high standards of drug advertising has steadily diminished. The adoption of the joint AMA-drug industry program outlined by Dr. Hussey would provide at best a questionable safeguard for the drug consumer. 2. Witnesses Testifytng on Patent Provisions The great majority of the testimony on the proposed patent reforms revealed the presence of a tropismic conservative reaction against any modification of the patent system, and the absence of any great evidence that the witnesses had studied the concrete operation of the patent system in the framework of the drug industry.boa The presentation given by Joseph Jackson, the chairman of the patent, trademark, and copyright law section of the American Bar As- sociation, may serve as an illustration. Speaking for the American Bar Associa- tion (ABA), Jackson reported that the ABA Board of Governors had adopted five resolutions in regard to S. 1552: (1) disapproval of the antitrust pro- visions limiting private patent interference settlements; (2) disapproval of the requirement of "non-obviousness" regarding the patentability of a drug product; (3) disapproval of the requirement of proof of significantly greater therapeutic effect in order to qualify molecular modifications for drug patents; * (4) disapproval of the distinctive treatment of drug patents with regard to the date of effective patent protection; (5) disapproval of compulsory licensing for drug patents.21 When Mr. Jackson was examined on his testimony, it developed that these resolutions, purporting to speak for the ABA, a `body with a membership of some 102,000, had been drafted in Saint LouIs, about ten days before the Hearings, by a group of 150 to 200 patent lawyers, with orders to "act with great aecelera~Uon" because of `the "emergency." ~ These resolutions were then submitted to the Board of Governors of the ABA (no member of which is a patent lawyer) and were PrOmulgated by them more or less over the head of the House of Delegates, the representative deliberative body of the ABA. (This is an extraordinary pro- cedure which is nevertheless permittof `by the constitution of the ABA.) Further- more, members of the ABA were apparently not given notice of this action by their Board of Governors, as evidenced by the surprise of several members of the Subcommittee staff belonging to ABA. It further developed that the ABA group required less than `an hour and a half to reach conclusions on matters which the Subcommittee and its staff had been studying for over two years. Jackson ad- mitted that there had been no discussion of the economics of the `drug industry and its relation to patents, no study of costs, no systematic consideration of prof- its, no attention to concentration or to the interdependence of major firms, no concern with entry conditions, no inquiry into price policies, and not even any consideration of the nature of incentives and the quality of research effort in drugs. Jackson explained that none of these matters had been considered, since "If we had discussed all these subjects you are presenting, we would still be in Saint Louis. We would not be here with ally resolutions at all." ~ But that is precisely the point. It may be inferred that the only `aim of the ABA action was to go on `record as condemning S. 1552 in time for the Hearings, regardless of the factual merits of the bill. The patent provisions stand or fall depending upon their application to the specific cireumstances of the drug industry; the entire patent system is not at stake. But patent attorneys in testifying almost universally took the position that any amendment to the patent laws in respect of a particular industry would necessarily imperil the patent laws with regard to all other industries. Jackson no doubt faithfully reported this attitude on the part of the ABA group when, in response to an Observation by the Subcommittee counsel that `S. 1552 was limited in its application to the drug industry alone, "We were afraid we would be faced with a special antitrust law and a special patent law for butter and eggs, and another one far milk and beer, and so on in different areas."24 One gets tsa There Is no evidence of any articulate "grass roots" support for the opponents of the patent provisions of S. 1552. Symptomatic of the testimony in behalf of the patent status quo is a communication from an individual styled "Clair V. Johnson, Newfane, Vermont, Patent Lawyer," who roundly condemns the bill. It appears that Mr. Johnson Is also a director of U.S. Vitamin and Pharmaceutical Corporation. (Part 3, p. 1601). 21Ibid., Part 3, p. 1473. ~ Ibid., Part 3, p. 1~72. ~ Ibid., Part 3, p. 1481. 24Ibid., Part 3, p. 1480. PAGENO="0455" COMPETITIVE PROBLEMS IN THE DRUG INDUSTRY 2005 the impression from the main drift of tile patent testimony that this rather naive contention is not advanced merely as a straw man, but that this consideration alone was enough to close the minds of patent attorneys regarding S. 1552. A further observation by Jackson leads one to wonder why his group should have debated even as long as an hour and a half on their resolutions: "There was only one man of the whole group who had anything good to say about the law . . . he said it might not be politically expedient to show the full extent of our disapproval." Had tile ABA group referred to the Subcommittee's Report on the drug in- dustry, they would have discovered that it is not at all unusual for nations to make exceptions of their drug industries in regard to patent privileges. Of some 77 nations with patent laws, 49 absolutely prohibit drug product patents on grounds of public policy, and 25 others have provisions for compulsory li- censing. Only Panama, Belgium, and the United States allow unrestricted drug product patent privileges.~° Indeed, to judge by a comparison of drug discoveries by drug firms in countries with and without product patents, it is `by no means clear that the patent incentive is necessary to elicit productive drug research. The patent privilege restrictions embodied in S. 1552, although more liberal, are closely related to those of Germany, long one of the world leaders in drug re- search. German patent law denies drug product patents, but drug processes may be patented, and such patents cover the products made by those processes. If, however, alternative processes are devised to produce the same drug, the drug in question then fails to retain its effective protection. Germany permits un- restricted drug process patents for a period of three years. After that, compul- sory licensing is required, with royalties of between 5 and 10 percent, as determined by the decision of a special tribunal.27 Professor Machlup of Princeton University testified that not only has the existence of such a patent law in Germany failed to halt productive drug research, but that its expediency and equity is not even questioned.3° One ominous tendency which came to light at the patent hearings concerns certain evidence that pressure groups are attempting to weaken the protection which drug buyers currently enjoy under the patent laws of many nations of the world. Professor Machiup testified that in recent years several countries, in con- sequence of pressure by industrial groups, have made their patent laws more favorable to such industries.29 It is ironic that at the time when efforts are being made to `bring United States drug patent policy into line with the more enlight- ened practices of other industrial countries, some of these very countries are experiencing a retrograde tendency. In 1949, England amended its patent law to allow drug products to be patented, but required compulsory licensing. France adopted a largely similar law in 1960.~° Drug and other chemical interests appear to have been active in connection with a certain diplomatic conference held in LiSbon in 1958 for the purpose of revising the International Convention for the Protection of Industrial Property. Prominent among the United States delegation were Roland Libonati, a Repre- sentative from New Jersey (where many drug firms have plants) and P. J. Federico, examiner in chief of the United States Patent Office, who testified at the hearings and was at pains `to take issue with the Subcommittee's conten- tions on the relative strength of patent protection in countries with and without product patents. The Lisbon convention agenda contained an item proposing to require all countries adhering to the, Oonvention to grant patents for chemical products, including pharmaceuticals. The United States delegation sponsored this resolution, and it fell to Mr. Federico's lot to expedite the proceedings. Tile resolution did not pass-12 nations voted against it-'but observers were some- what surprised to find that those voting in favor included the delegations from eight countries which proiibited drug product patents. After the resolution failed, Germany introduced a resolution to recommend that member countries study the question with a view toward revising their patent laws. This resolu- tion passed. Such efforts a't "study" may be bearing fruit: France, one of the countries opposing the original resolution in 1958, adopted its own drug product patent law in 1960, and the Scandinavian countries and Finland are trying to work out a common patent law which will extend to drug products. Ibid., Part 3, p. 1480. 2~ Study of Administered Price8, op. cit., p. 106. ~ Hearings on S. 1552, Op. cit., p. 1385. 28 Ibid., p. 1385. 29 Ibid., Part 3, p. 1371. 3°Ibid., PartS, pp. 1201, 1216. PAGENO="0456" 2006 COMPETITIVE PROBLEMS IN THE DRUG INDUSTRY 3. Testimony in Behalf of the Pharmacentical Manufactnrers Association The influence of the drug makers lobby in behalf of aborting reform efforts was not limited to their presentation of testimony at the hearings, although a survey of their manifold efforts a influencing-and even drafting-the legisla- tion on to control their own activities is beyond the scope of this paper.zoa Suffice it to say that the legislators must have found the PMA testimony the most con- vincing case with which they were presented, since the PMA recommendations agreed to a great extent with the Act as it was finally passed. Mr. E. N. Beesley, chairman of the PMA, presented his views as follows: (1) agreement with the the provision that FDA should pass on the efficacy of drugs, but judged on the basis of "substantial" and not "preponderant" evidence; (2) agreement that dis- tribution to physicians of full information on drugs is important, but that the provisions of S. 1552 in this respect are unnecessary, in view of the AMA's pro- motion of a cooperative program with PMA which will handle these details; (3) qualified agreement that some more systematic approach to the designation of generic names is desirable, but a preference for the announced AMA program of voluntary cooperation with PMA on generic names, with the concession that if such voluntary means do not achieve results, the FDA may designate a generic name; (4) disagreement with the compulsory licensing of drug manufacturers, but a readiness to compromise on simple registration, together with an expanded program of factory inspection by the FDA; (5) complete rejection of all phases of the patent program: condemnation of compulsory patent licensing, disap- proval of the restrictions of the patenting on molecular modifications and drug combinations, and disagreement w-ith the contemplated restrictions on private patent interference settlements, but with the concession that (as suggested by Patent Commissioner Ladd and Assistant Attorney General Loevinger) firms be required to file private patent interference settlements and other relevant data at the Patent Office for public inspection.3' The PMA testimony revealed a supporting willingness to compromise in regard to some parts of the bill, but the economic impact of these concessions would be but slight. It is also to be noted that while much is made of the prospects for co- operative programs with the AMA, nothing is included which would bring ad- vertising claims under any direct control. The prospects for constructive reform in these respects are made doubtful not only by the past record of the AMA but also by certain recent activities of organized drug makers in relation to attempts by a group of physicians, acting independently, to increase the standard of drug advertising.~ The PMA showed a clear intent to resist economic reforms which would reduce drug profits. Mr. Beesley testified directly to this effect: `Deletion of patent and trademark protection . . . would increasingly dilute the intensive competition 30a Kefauver, angered at the opposition of the subcommittee minority, and apparently sus- pecting it of connivance with the drug lobby, took the politically risky step of confiding in a journalist, Richard Harris, who was then doing a study of the drug industry. Kefauver made a detailed first hand report of the day by day political maneuvering. The resulting chronicle, summarized in the New Yorker, March 28, 1964, pp. 46 if., makes fascinating reading. It appears that PMA's initial reaction to Kefauver's bill was divided. One faction reportedly preferred to buy outright the necessary votes to defeat the bill; another thought it wiser to influence legislation in the making, conceding some points in order to moderate reformist pressures and prevent the eventual passage of truly substantial reforms. It appears that, under White House auspices, representatives of PMA were asked personally to aid in drawing up an alternative drug bill to replace that of Kefauver. At the subse- quently notorious "secret meeting" at which the substitute bill was drawn up, the only persons present were PMA lobbyists, legislative draftsmen from the Department of Health, Education, and Welfare, and counsel fbr the Subcommittee minority. Not a single senator was present. Nor had Kefauver, or any other member of the subcommittee majority been informed of the meeting. Fortunately, this attempt to railroad through an impotent bill was nullified when Kefauver succeeded in drawing public attention to it and the Adminis- tration saw fit to repudiate the substitute bill. ~` Ibid., Part 4, nn. 1997-2007. ~` The AMA seal of acceptance had been valued by some of the more responsible drug firms, large and small. With the demise of this program, some firms then supported a group of physicians who undertook to establish drug advertising standards and to review adver- tising copy, and who later became known as the Physicians Council. This group awarded approved advertisers with an insignia which could be displayed in advertisements. As the program became more widely accepted, the industry began to view it with suspicion, and in February 1959 the medical section of the American Drug Manufacturers Association (now part of PMA) adopted a resolution calling upon its members to refuse cooperation with the Physicians Council. The resolution claims that such outside review ould disrupt the normal, prompt, and effective processes of ethical drug promotion," a state- ment with which the Physicians Council would probably not find fault. The financial sup- port of the Council disappeared, and they abandoned their insignia program. Testimony of Dr. Julius Richmond of New York State University, Thid., pp. 361-368. PAGENO="0457" COMPETITIVE PROBLEMS IN THE DRUG INDUSTRY 2007 for superiority in discovery and manufacturing; it would leave price competition as the only remaining form of competition, and this would be a tragic result in an area so vital to the public health." ~ This is precisely the view of the subcom- mittee majority, although expressed from the point of view of an antipodal set of values. The Subcommittee was at great pains to prove that the "intensive com- petition for superiority" (i.e., in profits) stimulated by patent availability is wasteful, duplicative, and supports high prices, but that the replacement of prod- uct differentiation by price competition would be tragic only to excessive profit levels, and would improve resource allocation in the drug market and eliminate the inequity of burdening the sick with high prices. The drug industry tried to rebut these presumptions of noncompetitiveness by presenting a great deal of economic evidence regarding (1) price index data pur- porting to show a decrease in drug prices from 1949 to 1960, and (2) firm con- centration and turnover in firm rank order of sales volume in various drug sub- markets. Professor Markham of Princeton, retained by counsel for the PMA, presented a price index of prescription drugs, showing a decline of 7.6 percent for the period 1949-1960, or about one-half of one percent per year, compounded annually. This evidence, if valid, would argue for a remarkable degree of price stability in the industry, and might sustain the Subcommittee's loose usage of the loose term, "administered prices", one agreed-upon characteristic of which is that they do not typically increase during infiations to the extent that one would expect on the basis of short-run profit-maximizing theory. But what is the cause of this slight decline? Dr. Blair of the subcommittee staff pointed out that pat- ented drugs almost never change in price, while non-patented drugs typically experience a substantial price decline over time, and suggested that the chain index might reflect the presence of a number of patented drugs with fixed prices, and a number of unpatented drugs which declined in price. He asked further information be submitted with regard to the relative price changes of patented and unpatented drugs separately, and also for other information, including a tabulation of the drugs and prices employed in the index. Apparently the com- putations on patented versus unpatented drug price trends were never submitted, although the other material requested was supplied. Anyone who takes the trouble to look through the basic price data used in computing the index will find that of the 329 drugs included, no fewer than 193, or 59 percent of the total, had experienced no price change during the, period since their introduction in the index (not all drugs in the index were sold during 1949; many appear oily in subsequent years); 95 drugs (28 percent of the total) had experienced price in- creases (i.e., price in 1960 higher than price in initial year of appearance), and only 41 drugs (13 percent of the total) experienced price decreases.34 This tends to confirm the hypothesis of price stability. It is of further interest to note that the BLS index of the prices of drug and pharmaceutical materials declined by 32.7 percent during this time, while its index for the cost of retail prescriptions increased by 27.8 percent.33 It is perhaps evidence of lack of competition that prices of drugs dropped only 7.6 percent while prices of inputs dropped 32.7 percent; it is also curious that prescription costs increased 27.8 percent during this period if prescription drug costs actually fell by 7.6 percent, in view of the facts that (1) most prescription drugs are priced by druggists by adding a fixed markup (66% percent) over wholesale price, and that (2) if anything, the labor and other compounding costs of prescriptions should have declined since 1949 as a result of the marked decrease in the percentage of prescriptions which re- quird compounding by the pharmacist, rather than the transfer of capsules to different bottles. The concentration and turnover ratio data presented by Markham do indicate very considerable turnover among drug firms in relative sales rank. The con- clusions reached as regards concentration as a whole must be interpreted in the light not only of the usual imperfections of concentration ratio data (especially the implicit assumption of independence of action among firms) but also the fact that the data refer to sales of the finished product rather than to production of the bulk powder and there are great differences between concentration of these two in drugs. In 1958, for example, of the 51 most important prescription drugs, 27 were made by only one firm, 18 by two or three firms, and six by four to ~ Ibid., pt. 4, p. 2003. ~4Ibid. pt. 4, pp. 2576-d through 2576-k. ~ Ibid., pt. 4, pp. 2469-2470. PAGENO="0458" 2008 COMPETITIVE PROBLEMS IN THE DRUG INDUSTRY seven firms; with regard to sales, only 17 were sold by only one firm, 12 by two or three firms, and 22 by four to thirteen firms.~ Production of patented drugs is usually completely monopolized by the patent holder (unless there has been an interference settlement, with mutual licensing among contestants), but the sale of such drugs may not be monopolized. The patent holder thay not have a sufficiently developed marketing network, enough detailmen, etc., to fully exploit the produc- tion monopoly, and may hence license other firms with wider marketing facilities to sell the finished product, such licensees tableting and bottling the bulk powder which the patent holder sells to them. For example, Carter licenses meprobamate ("Miltown") to Wyeth in order to increase its total sales.~ While it is illegal for patent holders to dictate the pricing policies of licensees, price competition between such parties is almost unknown. But it is quite consistent with continual pure monopoly control of drug production that several of the licensees of the patent holder may alter their relative positions in the market from year to year. The turnover data presented by Markham do not indicate the extent to which this may be the case. The presumption made at the hearings that most of the turnover is induced by firms developing new products remains only a presumption; Mr. Mannis of Arthur D. Little, a participant in Markham's study, wasn't quite sure.~ Markham placed great emphasis on the amount of turnover, but to assess the de- gree of workability of competition evidenced by such turnover, it is necessary to determine whether it was brought about by price competition, by genuine product competition, or by mere product differentiation with its accompanying wastes in the drug industry. The evidence (admittedly uneven in quality) of some ten thousand pages of Senate hearings indicates that a very small weight be placed on the first factor, a quite moderate weight on the second factor (less important since 1955), and the preponderance of weight on the last. But Markham seems to believe that turnover in itself is a good thing, and didnot attempt to analyze its causes.n According to Markham, the purpose of his study is "a thorough and objective appraisal of all economic aspects" of the drug industry.40 But the study was spon- sored by PMA, and PMA had its schedule. Markham continued, "We have begun by organizing our research schedule so as to center attention on the specific issues raised by the Subcommittee's report no. 448, entitled `Administered Prices-Drugs,' published on June 27." (Markham's study began in July.) At the session of hearings when PMA presented its case in December, 1961, the study was apparently not yet complete. One of the issues slighted was price competi- tion, although this issue had certainly been raised in the Report. During the examination on his testimony, Markham conceded that he had not examined the question as to whether or not any drug firm had experienced a change in relative sales rank because of price competition,'~ and Mannis also confessed to ignoring the role of price competition in effecting changes in concentration.'2 Markham agreed that price competition is of paramount interest to the consumer, but con- cluded his contribution to the Hearings with this statement: "I have riot made any careful study of the workability of competition in the ethical drugs Industry. I was examining primarily these partIcular issues that seemed to be important."3 While a careful study of the workability of competition would admittedly take longer than the five months Markham had been able to devote to it prior to his ~ Study of administered prices, op. cit., p. 67. ~ Ibid., pp. 17-18. ~ Mr. Mannis testified, "Although I don't know, I would guess that back In 1951 or perhaps in the 1953 era, the leading product was quite different from the product that is leading the field now." Hearings, op. cit., pt. 4, p. 2102. : ~ `1lbid., Part 4, p. 2096. Other attempts to support the existence of price competition were also less than convincing. Professor B. V. Rostow of Yale University, testifying in behalf of PMA, was asked to give instances where patented drug prices had weakened. His "best example," upon closer inspection, pertained to an instance where prices had temporarily declined, not because of competition among different patent-licensed manufacturers, but as a result of manufacturers disciplining their own wholesalers (who had sold at low prices to hospital formularies and in other competitive bid markets) by undercutting their prices, then cutting off their supplies, and later restoring the original prices. Ibid., Part 4, pp. 2081-2082. 42Mr. Mannis attributed such changes as had occurred to "competition between and among products based on their merit. Ibid., Part 4. p. 2102. This would be more acceptable If genuine product competition-even at Identical prices-were prevalent In drugs; but when product differentiation becomes the road to sales, "merit" becomes what is measured by a DODularity contest. ~3fbi~., pt. 4, p. 2111. PAGENO="0459" COMPETITIVE PROBLEMS IN THE DRUG INDUSTRY 2009 appearance, it is quite another matter to seem to suggest that the workability of competition is not an important issue. On the contrary, the chief purpose of S. 1552 was to make the drug industry more workably competitive; it was not designed, for example, to stimulate or retard turnover in sales rank as an end in itself, or otherwise. But the point need not be labored. *B. Senate Judiciary Committe Action S. 1552 was reported out of Senator Eastland's Judiciary Committee as a much more innocuous document than when it was submitted.4" All the patent pro- visions had been rejected. The other contemplated reforms were considerably attenuated. The request for public control over generic names on the part of the Secretary of the Department of Health, Education, and Welfare survived in modified form: under certain circumstances, the Secretary is to hold hearings and establish generic names for certain drugs. The request that the FDA pass on the efficacy of drugs became a provision that drug efficacy claims be sup- ported by substantial evidence, but the wording adopted limited its jurisdiction only to initial claims made subequent to approval. The licensing requirement became mere registration, but the FDA was granted greater power to inspect plants. The requirement that advertisements contain generic names and state- ments on efficacy and side effects was entirely rejected. Kefauver and other members of the Subcommittee majority noted that the amended bill would perhaps improve the quality of drugs, but would have little if any effect on prices, and indicated their intention to propose several amend- ments to the bill in order to salvage some of its economic impact. First, they proposed that all drug patent license agreements be filed with the Commissioner of Patents. It will be recalled that 1~i~IA had expressed willingness to make pri- vate patent interference settlements a part of the public record; the amendment would not require public filing, but would extend to all patent licensing agree- ments. Second, it was proposed that the compulsory patent licensing feature of the original bill be retained, but that it be invoked only where wholesale drug prices exceeded 500 per cent of factory costs. Third, it was pointed out that the wording of the current bill would not restrict drug claims to demonstrated efficacy after the approval of a new drug application, and requested the necessary modification in wording. Fourth, a compromise on the contents of drug adver- tisement was put forth: generic names need appear in type face only half as large as brand names, and summaries, rather than complete statements, of efficacy and side effects should be included." C. Congressional action On August 3, 1962, the late President Kennedy communicated to Senator East- land a list of amendents for the Committee to consider, relating chiefly to drug safety. The Committee met a few days later, and in addition to adopting some of the amendments suggested by the President, accepted Kefauver's amendments re- garding advertising content and efficacy claims, but rejected his patent amend- ments. The bill was then transmitted to the House of Representatives, where further amendments were proposed. A conference report of October 3, 1962 shows that the conference between the Senate and House managers of the bill resulted in the substantial adoption of the Senate version, with only minor changes, except that the requirement that the Secretary of HEW publish and distribute to physicians all the detailed information required by law to be included in the drug package was dropped.'~ No reason is given for omitting this important pnovision. In the absence of information regarding motive or influence, the deprivation of easy access by the physician to the best information on drugs can only be inter- preted as an act of disinterest misanthropy. v. CONCLUSIONS AND POLICY RECOMMENDATIONS The amended bill was passed by both houses and became law on October 10, 1962, as the "Drug Amendments of 1962," more frequently referred to as the Kefauver-Harris Act. It will serve to improve the quality of drugs, but its Impact on competition will be slight. Requiring evidence of efficacy will, if properly en- Ma Rt. on S. 1552, op. cit., pp. 1-8. "IbId., pp. 36-48. ~ Drug Amendments of 1962: Conference Rept. No. 2526, House of RepresentatIves, 87th Cong., second sess., Washington, D.C., Government Printing Office, Oct. ~, 1962 pp. 10-18. PAGENO="0460" 2010 COMPETITIVE PROBLEMS IN THE DRUG INDUSTRY forced, keep off the market a certain number of useless but harmless drugs which presently encumber it. To the extent that drug firms are tempted to patent and push such inefficacious drugs in order to take advantage of the susceptibility of physicians to massive advertising, refusal to certify them will prevent further market confusion and reduce the amount of research devoted solely to pro vithng vehicles for sales promotion.40 But most research is probably not aimed at pure sales promotion of useless compounds; far more resources are wasted in duplicating existing drugs via molecular manipulation, and the failure of this part of S. 1552 to be enacted keeps incentives for such "research" as great as ever. Competition between generic and brand name products usually exists only where the drug is not patented, i.e., in only a small part of the prescription drugs market. The ability of generic drug firms to compete with brand name firms is slightly increased by the provisions for registration and Inspection of plants. This should give physicians more confidence in the quality of generic name products and increase their willingness to prescribe by generic name, but in view of the demonstrated ability of drug firm detailment to disparage generic producers, it is doubtful if the fact of more adequate FDA inspection, in itself, will eliminate the cultivated distrust of generic drugs on the part of physicians. In the past, detailmen have also disparaged the scope and the adequacy of FDA inspection, and ii rc~mains to be seen if they will have similar success even under the new program. The only sure way of eliminating disparage- ment is to eliminate the army of over 15,000 detailmen; eliminating monopoly profits will largely bring this about. The provisions giving the Secretary of HEW authority to establish generic names supplies the basis for sweeping reforms resulting in shorter and simpler names physicians can remember, spell, and hence can prescribe. But much depends upon the aggressiveness with which the Secretary carries out this mandate. At present, few generic names do not need~ reforming, but a universal housecleaning in nomenclature is probably not to be expected. Yet it would be most disappointing if the Secretary were only to exercise his power in those relatively few cases where no name exists, or more than one. The remaining reforms of advertising also proceed in the right direction, but not far enough. The requirements for summaries of side effects and efficacy in all advertisements will make for a more enlightened practice of medicine, but will not have direct effects on drug economics. The requirement that generic names be included in all copy will help, but the more such names are simplified, the more it will help. To an economist concerned with questions of public policy, the congressional performance in regard to S. 1552 is discouraging, but not entirely so.. The Act effected significant medical, if not economic, reforms. It now seems appropriate that renewed efforts be applied on the part of the public to induce congress to pass a more adequate act abolishing drug product patents and requiring generic name labeling. But the obstacles should not be underestimated. The influence of opposing groups is considerable, and the performance of several of these at the Hearings left much to be desired from the standpoint of concern for the public interest in health. Congress was apparently more sensitive to the need of protecting the consumer's physical health than his economic well-being; in fact the prime reason that any action was taken on the bill may have been its coincidence in time with the Thalidomide tragedy in late 1961, with the great publicity given the embryo-deforming effects of a new sedative drug. But it would seem preferable for the public to conduct a continuing campaign for more effective drug industry legislation, rather than to have such reform legislation come about piecemeal as a by-product of a series of tragic and avoid- able blunders on the part of pharmaceuticals makers. ~ Hearings on administered prices, op. cit., pt. 18, p. k0379 contains interesting testimony by Dr. A. D. Console, former medical director of Squibb, who in reply to Senator Kefauver's question as to how much drug firm research produces nothing worthwhile and is not in- tended to, answered: "I think the majority of it is in that category . . . with many of these products, it is clear while they are on the drawing board that they promise no utifity; they promise sales." PAGENO="0461" Income Opportunities arTd Physician Location Trends in the United States BY HENRY B. STEELE and GASTON V. RIMLINGER Reprinted from THE WESTERN ECONOMIC JOURNAL Vol. III, No. 2, Spring, 1965 2011 PAGENO="0462" 2012 COMPETITIVE PROBLEMS IN THE DRUG INDUSTRY INCOME OPPORTUNITIES AND PHYSICIAN LOCATION TRENDS IN THE UNITED STATES* HENRY B. STEELE and GASTON V. RIMLINGER Rice University INTRODUCTION In a highly industrialized society professional manpower is one of the most important resources, which implies that trends in its geographic location have important economic and social implications. The spatial distribution of physicians in relation to population has been highly uneven in the United States for at least several decades.' Since this situation must be considered less than ideal from both the economic and the social points of view, the trends in the location of physicians should be of particular interest. The present article seeks to answer three basic questions. First, is there any tendency toward change in the distribu- tion of physicians, and if so, is the long-run tendency toward a more even or a more uneven distribution pattern? Second, what is the role of physician location trends in the distributive pattern? Third, to what extent are the observed changes in physician location accounted for by such variables as population and per capita income, which may be taken to reflect income opportunities for physicians? It should be borne in mind that a more even distribution of physicians in relation to population can come about either because more physicians move to areas with a physician shortage or because more people move to areas with a relative physician surplus. The time period covered by the analysis is from 1950 to 1959. This is a rather short period, but its choice is dictated by the fact that adequate and comparable data on physician location are available only for these two dates. The data are from the health manpower surveys of the U.S. Public Health Service.2 The location trends over this period will be related to the following variables: (1) the degree of urbanization, (2) the regional shift in population, (3) the regional per capita disposable income at the beginning of the period, and (4) the increase in regional per capita income during the period. The choice of these variables is again limited partly by the availability of data and partly by the nature of the analysis, which is primarily statistical. A priori, the most important omission among vari- ables would seem to be the income of physicians. Relevant ~comparative regional ~ We are indebted to Mrs. Gloria Shatto for assistance in statistical computation. Respon- sibility for the article is solely that of the authors. An analysis of this distribution can be found in 0. V. Rimlinger and H. B. Steele, "An Economic Interpretation of the Spatial Distribution of Physicians in the U.S.," Southern Economic Journal, July 1963, PP. 1-12. 2The 1950 data are from M. Y. Pennell and M. E. Altenderfer, Health Manpower Source Book, section 4, U.S. Public Health Service (Washington, D.C.: U.S. Government Printing Office, 1954); and the 1959 data are from W. H. Steward and M. Y. Pennell, Health Manpower Source Book, section 10, U.S. Public Health Service (Washington, D.C.: U.S. Government Printing Office, 1960). PAGENO="0463" COMPETITIVE PROBLEMS IN THE DRUG INDUSTRY 2013 data on physician income are available only for the year 1949.~ An attempt will be made to investigate the significance of the physician income level at the beginning of our period in a separate comparison of standard metropolitan areas. TRENDS IN PHYsIcIAN~ DismiBul-loN AND LOCATION A brief introductory comment on the meaning of the physician distribution may be in order. Concern with physician distribution is based on the problem of access to medical services. The number of physicians in relation to population in a given area may be used as an index of the availability of physician services, but it should be noted that this is only an approximation. Availability depends on the relationship of the supply and demand for services. The demand is determined not only by the size of the population, but also by such interrelated factors as the level and distribution of income, the age structure of the population, the possession of health insurance, the level of educatiOn, and the degree of urbanization. The supply of physician services depends similarly not only on the number of physicians but on the amount of time they devote to patients and their effectiveness in treat- ing them. An area with a high concentration of retired physicians or with many physicians engaged in research or teaching naturally has a smaller supply of services for patients than a similar area where all physicians are engaged in active private practice. Unfortunately, we lack the data necessary to take these factors into account. The present analysis is based on the total number of physicians in an area rather than on the amount of service they provide. The areas used are for the most part groups of similar counties within a state; in one case standard metro- politan areas will be used. In each state counties are combined into the following groups adopted by the health manpower surveys: greater metropolitan, lesser metropolitan, adjacent, isolated semi-rural, and isolated rural.4 Although some information regarding income differentials is lost by aggregating the 3069 counties of the United States into a little over 200 groups, the group is considered a prefer- able area concept as far as the market for medical services is concerned. We turn now to the question of whether there was a change in the distribution of physicians between 1950 and 1959 and the direction of any such change. No change in distribution would mean that every area experiences the same relative increase or decrease in the physician-population ratio. In other words, if the national physician-population ratio increases by 5 per cent, an increase of 5 per cent of this ratio in every area of the country would leave the over-all distribution unchanged. Given the constant internal migration of the population and the entry and exit of physicians into and from professional life, it would be a very unusual coincidence if there were no change in distribution over a nine-year period. Hence, the question of whether there was a change in distribution is by itself not very meaningful. It is the extent and the direction of change that are `These data are used below in the analysis of 19 standard metropolitan areas. For the source of the physician income data, see W. Weinfeld, "Income of Physicians, 1929-1949," Survey of Current Business, July 1951, pp. 9-26. 4For a precise definition of each of these, see Penriell and Altenderfer, op. cit., section 4, pp. 18-20. PAGENO="0464" 2014 COMPETITIVE PROBLEMS IN THE DRUG INDUSTRY important. Physicians are most highly concentra~ed in high-income and in urban areas. After experimenting with several approaches, it developed that the most informative way to put the question of change in distribution is to inquire whether the degree of concentration of physicians in these areas has increased or decreased. The answer depends on how concentration is measured. It is interesting, and some- what unexpected, that relative concentration in higher income areas decreased between 1950 and 1959, while at the same time the physician-population ratios in more highly urbanized areas increased. To clarify the meaning of these findings a more detailed examination is necessary. To measure the change in relative concentration with respect to regional income, the writers constructed two Lorenz curves relating physicians to popula-. tion, one for 1950 and one for 1959, as shown in Chart I. On the horizontal axis CHART I LORENZ CURVES FOR DISTRIBUTION OF PHYSICIANS BY RANK ORDER OF COUNTY GROUP PER CAPITA DISPOSABLE INCOME' 0 0 C C, C, Per cent of total population Total physicians, 1950 and 1959; physicians in active private practice, 1959. PAGENO="0465" COMPETITIVE PROBLEMS IN THE DRUG INDUSTRY 2015 the population of the United States is arrayed in rank order of county group per capita disposable income,5 with incomes rising from left to right. The vertical axis represents the total number of physicians in the country. It will be noted that the Loren.z curve for all physicians in 1959 lies closer to the diagonal than the corresponding curve for 1950, indicating a change toward a more even distribution of physicians with regard to regional income. It is interesting to note also that when only physicians in private practice are considered, the distribution in 1959 is more even than when all physicians are taken into account. Unfortunately, we do not have the necessary data to analyze the change in the distribution of private physicians between 1950 and 1959. TABLE 1 CHANGE IN PHYSICIAN CONCENTRATION IN RELATION TO PER CAPITA REGIONAL INCOME, 1950-1959 Population ranked by income Per cent of physicians in 1950 Per cent of physicians in 1959 Per cent change 1950-1959 Lowest 10% 4.0 4.8 + 19.8 10-20 5.4 6.1 + 13.5 20-30 6.7 8.4 + 24.5 30-40 8.0 8.1 + 0.9 40-50 8.9 9.6 + 7.8 50-60 12.0 10.4 - 13.3 60-70 12.6 11.8 - 6.1 70-80 11.6 12.2 + 5.3 80-90 13.2 13.9 + 5.1 90-100 17.5 14.9 - 15.1 A clearer picture of the change in concentration can be obtained by examining Table 1, which shows the gains and losses of the regionally stratified income groups. On the whole the change was clearly in favor of lower income areas. The top 10 per cent of the income-ranked population suffered the greatest relative loss of physicians. The population with the greatest gain was in the third decile from the bottom of the income range. Of course, physicians still remain con- centrated in high-income areas. In 1950 the bottom 25 per cent of the population had 13.0 per cent of the physicians and the top 25 per cent had 36.5 per cent of the physicians. In 1959 this had improved to 14.4 per cent for the bottom quartile `on population and per capita disposable income are from Sales Management Annual Survey of Buying Power, May 1951, and May 1960. The income for the county group is a weighted average. The Sales Management staff terms its income measure "effective-buying income," but observes the same definition the U.S. Department of Commerce employs in defining disposable income, 81-280 0-68--pt. 5-30 PAGENO="0466" 2016 COMPETITIVE PROBLEMS IN THE DRUG INDUSTRY and 35.5 per cent for the top quartile. The top 10 per cent of the populatipn still had three times as many physicians in 1959 as the bottom 10 per cent, but in 1950 the top 10 per cent of the population had over four times as many physicians as the lowest 10 per cent. This improvement in the physician-population ratios of the lower income relative to the higher income areas should not be interpreted to mean that physicians relocated from high to low income regions. A more likely explanation is the tendency of the population to move to higher income areas, and the changes in regional income levels. A valid argument can be made that for a general evaluation of the distribution, the concept of relative distribution developed above is more meaningful than the measurement of distribution based on degree of urbanization. Relative distribu- tion takes into account the proportion of the population that gains or loses by a change in distribution. On the other hand an analysis in terms of the degree of urbanization shows what happens to specific social segments of the population and gives us another way of looking at the problem of distribution. It also has the advantage of isolating one of the important causal factors in the location of physicians and the relocation of the population as a whole. The results of this approach for the period 1950-1959 are summarized in Table 2. An important fact which can be noted in Table 2 is that in none of the areas did the number of physicians increase as much as population or decrease as little as population. This means that the over-all decrease in the country's physician- population ratio between 1950 and 1959 left all aggregated county groups with fewer doctors in relation to population. However, not all areas shared equally in this relative loss of physicians. There was a redistribution in favor of the more highly urbanized areas and against the rural areas. In isolated rural areas popula- tion decreased slightly (.7 per cent), but the number of physicians decreased about six times as fast (4.4 per cent). This is the meaning of the 5.9 coefficient of elas- ticity of physician mobility.6 This coefficient merely describes a numerical rela- tionship and is not intended to indicate a causal significance of the same magni- tude. Isolated semi-rural areas fared even worse than isolated rural areas, even though they had a slight increase (1 per cent) in the absolute number of physicians. Their problem is that population increased 11 times faster than the number of physicians. The other three county groups had somewhat comparable changes in physician-population ratios. Most favored were the greater metropolitan areas, where the relative increase in the number of physicians was equal to 70 per cent of the increase in population, as indicated by the elasticity coefficient of .7. In the lesser metropolitan areas, the relative increase in the number of physicians was second largest in the country, but the relative increase in population was the largest. As a result the lesser metropolitan areas had a rate of physician increase that was about half as fast as the rate of population increase. In the adjacent areas, where the rate of physician increase was much smaller, the rate of popula- tion increase was relatively still smaller. The result was a rate of physician increase equal to 60 per cent of the population increase. 6 Computed before rounding the percentage changes. PAGENO="0467" TABLE 2 CHANGE IN PHYSICIAN DISTRIBUTION IN RELATION TO DEGREE OF URBANIZATION, 1950-1959 Per cent change in County group physicians (1) Per cent change in population (2) Average population elasticity of physician mobility (1) ± (2) i (3) Per capita ncome in 1950 (4) Per capita income in 1959 (5) Per cent increase in income (6) Average per capita income elasticity of physician mobility (1) -~ (6) (7) Isolated rural -44 - 7 59 $ 849 $1,468 72 9 06 Isolated semi-rural 1.0 113 .1 1,046 1,506 44.0 .02 LTJ Adjacent 9.8 16.4 .6 1,287 1,975 53.5 .18 Lesser metropolitan ... 15.6 32.1 .5 1,439 1,888 31.2 .50 Greater metropolitan ... 18.7 27.9 .7 1,623 2,185 34.6 .54 I. PAGENO="0468" 2018 COMPETITIVE PROBLEMS IN THE DRUG INDUSTRY Table 2 shows also some relationships between changes in physician-population ratios and regional per capita incomes. Column (7) shows the average per capita disposable income elasticity of physician mobility and reveals a somewhat sur- prising negative coefficient for isolated rural areas. A 1 per cent increase in per capita disposable income in an isolated rural region would be associated with a decrease of .06 per cent in the number of physicians. With increasing degrees of urbanization, the income elasticity increases: for isolated semi-rural areas, an increase in per capita disposable income of 2 per cent will increase the number of physicians by .02 per cent; the ratio increases to .18 per cent, .50 per cent, and .54 per cent for adjacent, lesser metropolitan, and greater metropolitan county groups. There is a clear tendency toward higher income areas having smaller declines in the number of physicians in relation to population. This would seem to con- tradict the earlier finding about a change in relative distribution in favor of lower income areas. However, there is no necessary inconsistency, for two quite different ways of measuring changes in distribution are involved. Nevertheless, the apparent inconsistency does highlight the difficulty of describing changes in distribution in an unequivocal manner. Some questions may be raised also with respect to the relevant time period for which income is measured. This is especially the case when there is a change in the income rank over time. For instance, in 1950 lesser metropolitan areas had the second highest average income of the five county groups, but in 1959 they were in third place, behind adjacent areas. If the causal effect of income operates with a lag, the initial income level is more likely to be relevant. If it operates through anticipation, the final income level might be more appropriate, or even the rate of increase in income over time. Multiple regression and multiple correlation analysis is necessary to determine the respective signifi- cance of these different factors. This is the task of the following section. ANALYSIS OF PHYSICIAN LOCATION TRENDS The preceding discussion has shown that changes in physician location over time are an important aspect of the regional redistribution of physicians relative to population. A complete analysis of changes in distribution would have to account for the movements of the population as well as the movements of phy- sicians. In this paper, however, we are concerned only with the movement of physicians. Population shifts will be taken as an independent variable that has a causal effect on the location of physicians. One would expect that an increase in population in a given area, especially if it is through migration, makes it easier for physicians to open new practices and induces them to locate there. Of course, once a doctor has established his practice in a given area and built up his clientele, he is not likely to move. The highly personal nature of the doctor-patient rela- tionship is undoubtedly an important barrier to mobility. On the other hand, new doctors are constantly entering the profession while others are retiring and leaving vacancies. It is mainly through these entries and exits that physicians are relocated, which makes it important to add a substantial time dimension to the PAGENO="0469" COMPETITIVE PROBLEMS IN THE DRUG INDUSTRY 2019 analysis of their mobility. It is the mobility of the group rather than that of the individual that counts. In addition to increases in population, Table 2 suggests that the regional level of income may act as a strong inducement to new physicians. The simple prospect of a wealthy clientele might have much drawing power. People in the higher income brackets might be in the habit of paying their medical bills more promptly than persons with low incomes. But if the drawing power is based primarily on the physician's expectation to be able to earn a high income by charging higher fees, it will be greatly dampened if the high-income area already has a very large number of doctors in relation to population. On the other hand, a high concen- tration of doctors (provided that it does not affect income too adversely) might~be attractive because it offers more opportunity, for leisure. It is reasonable to assume that where there are many doctors it is easier to find vacation or weekend replace- ments than where there is a shortage. The results of an attempt to evaluate the relationship of changes in the regional number of physicians to changes in a selected set of variables are pre- sented in Tables 3 and 4, which show the coefficients of regression, and of correla- TABLE 3 REGRESSION COEFFICIENTS PERTAINING TO PERCENTAGE CHANGES IN THE NUMBER OF ALL PHYSICIANS IN Six SECTORS, 1950-1959 a County group Regression constant (a) Coefficient of per cent change in population (b') Coefficient of per capita p income in 1950 (c) Coefficient of er cent change in income (d) Isolated rural -40.43 .64 .02 .03 Isolated semi-rural - 29.34 - .22 .40 .05 Adjacent 48.26 1.42 - .50 .14 * Lesser metropolitan 56.72 1.00 - .08 .29 Greater metropolitan 48.86 .92 - .01 .61 . `~ ~ Coefficient of physician income in 1949 * Nineteen standard metropolitan areas 50.33 .96 .01 .00" a The value of the coefficients is derived from the linear regression equation: `b~P di~I - = a + + c It(o) + ~, where N represents the number of physicians; I, the Nt(o) Pt(O) it(o) regional per capita income; P, population; and t(o), the initial year of the period. b Less than .005. PAGENO="0470" TABLE 4 income ~ change ~ (R,.4) ~ .82 .06" CORRELATION COEFFICIENTS PERTAINING TO PERCENTAGE CHANGES IN THE NUMBER OF ALL PHYSICIANS IN SIX SECTORS, 1950-1959 County groups TOTAL CORRELATION PARTIAL CORRELATION SIMPLE CORRELATION All variables (R1.,,,) Population and income level (R1.,,) Population and income change (R,.,4) income level and income change (R1.,4) Population (R,,,) Income level (R1.,) Isolated rural .33 .16 .20 .14" .14 .06" Isolated semi-rural .63 .61 .59 .28 - .57 .26 Adjacent .87 .82 .82 * .08" .82 - .01" Lesser metropolitan .93 .90 .92 .88 .87 -32 Greater metropolitan .67 .48 .63 .64 .44 - .37 .53 Population and physician income Physician income and income change Physician income Nineteen standard metropolitan areas .95 .95 .95 .17 .95 .17 "Not significant at the .05 level. PAGENO="0471" COMPETITIVE PROBLEMS IN THE DRUG INDUSTRY 2021 tion of these variables. The tables relate the average rate of change in the number of physicians in each of the county groups to (1) the average rate of population change, (2) the average level of absolute income in 1950, and (3) the rate of change of the average per capita income during 1950-1959. These national averages are simple arithmetic averages of state data, which in turn are weighted averages of county data within each county group. In the case of the 19 standard metropolitan areas, the averages are simple arithmetic averages, but the level of income of the population in 1959 is replaced by the average level of physician income in 1949, which adds a fourth variable to the analysis. The fifth variable is the degree of urbanization, which is treated as a qualitative variable. It could be included in the regression and correlation equations by means of dummy vari- ables, but we chose the alternative of computing separate regressions and correla- tions for each county group and for 19 standard metropolitan areas. This has the effect in each case of keeping the degree of urbanization constant and provides a basis for comparison between urban and rural areas. One of the obvious facts indicated by Table 3 is that the impact of the popu- lation and income variables depends greatly on the degree of urbanization. It is well known that the rural environment :tends to discourage and the urban environment to encourage the location of physicians. There are many factors, especially differences in cultural and professional advantages, which account for this differential in drawing power. But here we are concerned only with variables that affect the income opportunities of physicians. To what extent do they have a differential impact in the different environments? We might like to know whether improved income opportunities would work in attracting more physicians to rural areas and roughly how much of ati improvement would be required. Of course, our data can only give crude indications of what the answer to these questions inight be. Let us compare now the effects of our variables in the different environments. Column (b) of Table 3 shows the percentage increase in the number of physicians associated with a 1 per cent increase in population in the different county groups. On the whole the effect of population change is quite strong, but it varies greatry between areas. It is greatest in the adjacent areas, where a 1 per cent increase in population is associated with a 1.42 per cent increase in the number of physicians. Somewhat puzzling is the case of semi-rural areas, where a 1 per cent increase in population is associated with a .22 per cent decrease in the number of physicians. In lesser metropolitan areas, population and physicians tend to increase at the same rate, while in more urbanized areas the rate of physician increase is somewhat smaller than the rate of population increase, and in isolated rural areas it is much smaller. The basic distinction is clearly between rural and nonrural areas. Assuming the observed association to be of a causal nature, we could say that `A few counties changed from one group to another over the nine-year period in question, usually moving from lesser to greater urbanization categories. In these cases the county was included in the category which it occupied in 1950, on the assumption that its condition in the initial year had a greater impact than its classification at the end of the period. The total effect of such changes, however, is negligible since only 34 out of 3069 counties were involved. PAGENO="0472" 2022 COMPETITIVE PROBLEMS IN THE DRUG INDUSTRY in rural areas population increases tend to lower the physician-population ratio substantially, whereas in metropolitan areas these ratios would be little affected and in adjacent areas they would increase. The extent to which the average income level in 1950 is associated with the increase in the number of physicians is shown in column (c) of Table 3. The variations between areas are again very great and this time rather unsystematic. In isolated rural areas, a $100 higher per capita income in 1950 is associated with a 2 per cent increase in the number of physicians over the 1950-1959 period. On the other hand, in isolated semi-rural areas, a $100 higher income is associated with a 40 per cent increase in physicians, and in adjacent areas with a 50 per cent decrease. These erratic results raise questions about the significance of the observed association, which will be explored below. With respect to the 19 standard metropolitan areas, the observed relationship indicates a 1 per cent increase in the number of physicians over 1950.4959 for a $100 higher level of physician income in 1949. Column (d) of Table 3 describes the associations between increases in phy- sicians and increases in regional per capita income over 1950-1959. In the 19 standard metropolitan areas a systematic relationship is for practical purposes nonexistent. But it is interesting that in the other areas the association becomes constantly closer with higher degrees of urbanization. In isolated rural areas a 1 per cent increase in income is associated with only a .03 per cent increase in the number of physicians, whereas in greater metropolitan areas it is associated with a .61 per cent increase in physicians. We have now measured the relationships between changes in selected vari- ables and corresponding changes in the number of physicians. But it must be remembered that all these changes are averages and that their significance depends on the scatter of the original values about the regression line. An indication of this significance is presented in Table 4, which contains the correlation coeffi- cients for the population and income variables. On the whole the correlations are highest in the 19 standard metropolitan areas and the lowest in the isolated rural areas. As a matter of fact, the variables considered in this study, other than the changes in population, are of little use in explaining changes in the number of physicians in isolated rural areas because the simple correlation coefficients do not appear to be statistically significant. In isolated semi-rural and in greater metro-. politan areas the situation is slightly better. The index of total correlation R1.234 for the semi-rural areas is .63, which makes the index of total determination R21234 = .40. In other words, 40 per cent of the observed change in the number of physicians in the semi-rural areas over the nine years is "explained" by popula- tion change, the initial income level, and the change in income. In greater metro- politan areas 45 per cent of the change is explained by these variables. However, in adjacent, lesser metropolitan, and the 19 standard metropolitan areas, the corresponding percentages are 66, 86, and 90, respectively. Among the variables examined, population change is the one most closely correlated with change in the number of physicians. Total correlation is generally PAGENO="0473" COMPETITIVE PROBLEMS IN THE DRUG INDUSTRY 2023 not greatly affected by dropping either the initial level of income or the change in income as independent variables. In the case of the adjacent, the lesser metro- politan, and the 19 standard metropolitan areas, total correlation is not greatly affected even if both of these variables are dropped. Only in lesser metropolitan areas is there a high simple correlation between change in income and change in physicians. In these same areas the combination of income level and income change also has the highest correlation with changes in the number of physicians. The level of physician income at the beginning of `the period has only a very low correlation with the subsequent increase in physicians. Finally, it should be noted that, on the whole, the change in physicians is more closely correlated with changes in per capita income than with the initial level of per capita income. This is especially the case in the lesser and the greater metropolitan areas. SUMMARY AND CONCLUSIONS It has been shown that the two most important factors affecting the change in location of physicians over time are the regional degree of urbanization and the increase in population. The extent to which an area is urban or rural, or adjacent to an urban area, influences its attractiveness to physicians aside from what is happening to its population or to its level of income. An increase in population has an effective drawing power in urban but not in rural areas. In the isolated rural areas as a whole both population and physicians decreased, but there was no meaningful correlation between these two changes. Rather surprisingly, the level of per capita income at the beginning of `the period turned out to have little effect on physician location. Nor were physicians induced to locate on the basis of higher average physician incomes. Increase in per capita income seems to have been a, factor in attracting more physicians but only in lesser and greater metro- politan areas. The fact that the isolated rural areas had the highest rate of increase in per capita income did not prevent a decrease in the number of physicians. Even when initial per capita income and increase in income are combined, their effect is strong only in lesser and greater metropolitan areas. This indicates that on the whole the income of `the population is a less significant factor than we had originally suspected. However, in evaluating the meaning of these findings it is important to bear in mind that they refer to all physicians, rather than only to those engaged in private practice. The Lorenz curves in Chart I show that there is a substantial difference between the distribution of all physicians and that of private physicians. It is possible therefore that if a comparison over time were made of physicians in private practice, the income of the population would be a more important factor. At least this was suggested by our analysis of the distribution of private physicians in 1959.8 It is quite clear that a more disaggregated analysis is necessary to obtain more precise and more reliable results. For instance, it might be worth while to break down the increase in population into increase through migration and natural increase. It might be profitable also to analyze changes in the number of physicians See Rimlinger and Steele, op. cit. PAGENO="0474" 2024 COMPETITIVE PROBLEMS IN THE DRUG INDUSTRY by examining both entries and exits in a given area. The age structure of the physicians would be important in this respect, since it affects both their income and their exit from the profession. Additional insight would be gained by studying separately the location trends of specialists and general practitioners. Table 5 shows that the degree of urbanization is a much stronger factor in the location of specialists than in the location of general practitioners. TABLE 5 PRIVATE NON-HOSPITAL PHYSICIANS PER 100,000 POPULATION IN THE UNITED STATES, 1959 County group General practitioners Full-time specialists Isolated rural Isolated semi-rural Adjacent Lesser metropolitan Greater metropolitan 42.8 46.5 46.4 42.5 51.6 2.6 22.9 20.0 55.4 64.4 Sou1~cE: Health Manpower Source Book, section 10, p. 12. A final comment may be in order with regard to the general problem of physician distribution. It was shown that the relative inequality of distribution based on regional per capita income has decreased while the urban-rural inequal- ity has increased. This is consistent with the findings of the physician location trends. Since the pull of population increase is stronger than the attraction of per capita income, a redistribution in favor of lower income areas is possible. The change in distribution, however, seems to be dominated by the movement of population rather than the movement of physicians. Population tends to move to higher income areas and physicians follow but at a slower rate. This has the long-run tendency of reducing the concentration of physicians in higher income areas, while at the same time it is not necessarily inconsistent with a continuation of great urban-rural inequalities. But it must be remembered that an increasingly smaller percentage of the total population will be in areas with low physician- population ratios. Complete elimination of higher physician-population ratios in higher income areas is not likely to occur, nor is it necessarily desirable from an economic point of view. How fast the equalization of distribution will proceed depends in part on how the population redistributes itself and how fast physicians follow the movement of the population. Obviously, if population increases in higher income areas did not attract more physicians, the relative concentration of physicians in these areas would be more quickly eliminated. PAGENO="0475" Reprinted from SOUTHERN ECONOMIC JOURNAL Vol. XXX, No. 1, July, 1983 Printed in U.S.A. COMPETITIVE PROBLEMS IN THE DRUG INDUSTRY I. INTRODUCTION The number of physicians in relation to population varies widely between localities and regions in the United States. This pat- tern of distribution is related to differences in regional per capita income levels and raises questions not only from a public health but also from an economic point of view. To what extent and in what manner do economic factors influence the distribu- tion of physicians? What kind of economic behavior of physicians is suggested by their spatial distribution? Do the economic forces at work within the given institutional ar- rangement tend toward a distribution that is consistent with efficient use of health manpower resources? The purpose of this article is to seek an answer to these questions.' It begins with a description of the empirical relationship between the ratios of physicians to popula- tion and regional per capita incomes. Several hypotheses are then developed to explore the economic variables that may account for the observed relationship. Finally, the * We are indebted to Professor Edgar 0. Ed- wards and Dr. George J. Benston for their helpful comments on an earlier draft and to Mr. Vincent Tarascio for his assistance in compiling statistical information. Responsibility for the article is solely our own. 1 Similar questions may be raised with regard to distribution between specialties. None of these questions has been adequately explored by econ- omists. See, however, Rashi Fein, "Studies on Physician Supply and Distribution," American Journal of Public Health, May 1954, pp. 615-624; and M. Friedman and S. Kuznets, Income from Independent Profe8sional Practice (New York: National Bureau of Economic Research, 1954). 2025 Rice University hypotheses are subjected to empirical tests, insofar as is possible with the data available on physician incomes, physician visits, and medical expenditures. Although relevant data in this area are limited, the article does bring together in a concise form a large amount of statistical information from both public and private sources. II. THE DISTRIBUTION OF PHYSICIANS Any analysis of the spatial distribution of physicians poses the problem of defining the geographic area served by physicians. Following the practice of the Health Man- power Survey,' we shall assume that the "county group" is the relevant economic area. A county group combines counties with similar population concentrations within a state; these counties are often but not always geographically contiguous.' The 2 The Health Manpower Surveys used in this study are in M. Y. Pennell and M. E. Altenderfer, Health Manpower Source Book, Section 4, U. S. Public Health Service (Washington, D. C.: U. S. Government Printing Office, 1954); M. Y. Pennell and M. E. Altenderfer, Health Manpower Source Book, Section 5, U. S. Public Health Service (Washington, D. C.: U. S. Government Printing Office, 1954); and in W. H. Stewart and M. Y. Pennell, Health Manpower Source Book, Section 10, U. S. Public Health Service (Washington, D. C.: U. S. Government Printing Office, 1960). `Five groups are used and they are designated as follows: (1) greater metropolitan, where the populations exceed one million; (2) lesser metro- politan, where populations range from 50,000 to one million; (3) adjacent counties, those which are contiguous to metropolitan areas; (4) isolated semirural areas, for counties with at least one city of 2,500 people; and (5) isolated rural areas, for the remaining rural counties. Residents of ad- jacent areas are no doubt served on occasion by AN ECONOMIC INTERPRETATION OF THE SPATIAL DISTRIBUTION OF PHYSICIANS IN THE U. S.~ GASTON V. RIMLINGER AND HENRY B. STEELE PAGENO="0476" 2026 COMPETITIVE PROBLEMS IN THE DRuG INDUSTRY 600 IOPO 1500 2000 2500 3000 CHART I PER CAPITA EFFECTIVE BUYING INCOME 1959 data presented in Chart I thus are based on the number of physicians, the size of the population, and the per capita income of the population in each county group for the entire country in 1959.~ Each observation is for a county group, of which there are one to five in each state. The distribution is stated in terms of privately employed physicians, both general practitioners and specialists, per one hundred thousand persons in a county group. This measure excludes all federal physicians, all retired or inactive physicians, and all full- time hospital, teaching, and research physi- cians. In an analysis of the role of economic variables in the spatial distribution of physi- cians these exclusions seem to be clearly desirable. The income measure to which the physician-population ratio is related in Chart I is the county group per capita net effective buying income.6 metropolitan physicians, which tends to under- state the effective physician-population ratio for adjacent areas and to overstate it for metropolitan areas. Isolated areas are more likely to be self- sufficient. `The data for Chart I are from Health Man- power Source Book, Section 10. `Net effective buying income is identical with "disposable personal income" as computed by the Department of Commerce. See Sales Manage- ment, May 1961, pp. 54-56. The physician-popula- tion ratios and the per capita incomes are weighted averages of all counties of a given county group. The most significant aspect of the distri- bution presented in Chart I is the tendency of the physician-population ratio to rise with increases in regional per capita income.6 For instance, at the $1,000 per capita income level the average ratio is about 45 physicians per 100,000 population, but at the $2,000 level it is around 90 physicians per 100,000 population. This doubling in the number of physicians in relation to population is not due solely to changes in income. Low income county groups are almost exclusively rural while high income county groups are metro- politan areas. Income and environment change together, which makes it impossible to separate their impacts. There are valid reasons for believing that both higher in- comes and urban environment attract more doctors, but we are concerned primarily with the question of how larger incomes may account for higher physician-population ratios. III. THEORETICAL INTERPRETATION This section develops hypotheses to ex- plain the rise of the physician-population ratio with increases in per capita income. The relationship is mildly curvilinear as in- dicated by the equation p = 21.24(2.058)' for the regression line fitted by least squares. 160 ~15C 0 a ~I2C 0 09C 0 6c Sc PAGENO="0477" COMPETITIVE PROBLEMS A. Linear Model Let us begin with the simple case in which it is assumed that all physicians are inter- ested in maximizing money income. We may further assume either that all physicians are of equal ability, or that patients have no way of determining physician ability; in either event, there are no quasi-rents at- tributable to ability or reputation. It is also assumed that all physicians are in the habit of charging what the traffic will bear, and that this amount is a fixed per cent of the patient's income.7 Finally, it is assumed that there is no price competition among physi- cians and that the number of patient visits per physician is a function of the physician- population ratio alone, which rules out any relationship between per capita income and the number of patient visits per time period. On the basis of this set of highly simplified assumptions we shall set up a theoretical model using the following symbols: a = proportion of income spent on the average visit, which is constant b = number of visits per person per unit of time, which is also constant Y = national per capita income Yr = regional per capita income N = total number of physicians P = population R = P/N, national physician-population ratio Rr = regional physician-population ratio y = income of physicians, which in equilibrium is the same for the entire country The equilibrium income of physicians is equal to the aggregate expenditure on physi- cian services divided by the total number of physicians. - (abY)P abY NR In any given region we have The percentage will vary among different types of visits, but it is not necessary to introduce this complication. IN THE DRUG INDUSTRY 2027 (2) abYr y=~ and (3) Rr=~Yr y By; thus determining the relationship be- tween Rr and Yr we have defined the spatial distribution of physicians in relation to population and regional per capita income. This distribution pattern is illustrated in Figure 1 for various total numbers of physicians in the country. The number of physicians on the graph increases from N0 to~ N3 while the corresponding physician income decreases from ys to y~. The physi- cian-population ratio in any area is a linear function of regional per capita income and is represented by a straight line starting from the origin. The larger the total ninnber of physicians (N), other things being equal, the smaller the average physician income (y) and the greater the slope (ably) of the line representing the distribution of physicians. Several characteristics of this distribution may be noted. (1) If physicians and patients behave as the model assumes they do, the physician distribution will be a direct conse- quence of the regional distribution of income. (2) The absolute difference in the number Fin. 1.-Linear Model of Physician Distribu- tion. Regional Physician-Population Ratio (Rr) varies directly with Regional Per Capita Income (Y~). The total number of Physicians (N) deter- mines the level of physician income (y). PAGENO="0478" 2028 COMPETITIVE PROBLEMS IN THE DRUG INDUSTRY of physicians in relation to population be- tween any two areas depends on the differ- ence in income, the proportion of income, a, spent on each visit, the number of visits per person per unit of time, b, and the total number of physicians, N. However, (3) the relative distribution of physicians, as shown by a Lorenz curve, is independent on the magnitude of either a, b, or N. The effect on distribution of an increase in N can easily be demonstrated graphically. In Figure 1 an increase in the number of physicians from N1 to N3 raises the physi- cian-population ratio in area Yr, from R1 to R3 and in area Yr, from R2 to R4. The absolute increase in the ratio is greater in area Y~, than in area Y,! , since R4 - R2 > R3 - R1, but the relative increase is the same in both areas, since R2/R1 - R4/R3. The percentage increase in the physician- population ratio in any area will be the same as the percentage increase in the total num- ber of physicians.8 The distribution of physicians in this situation could therefore not be made more equal by simply increasing the number of physicians; it would in fact become less equal in an absolute sense. Nor would a willingness on the part of physicians to reduce their fees in relation to income result in a more equal relative distribution, unless fees were reduced to zero. A change in fees or in the number of visits per person would 8 This can be demonstrated mathematically by showing that the elasticity of the regional physi- cian-population ratio with respect to changes in the number of physicians equals one: dR? /dN ERr,N1/ ~ (abY)P By substituting -~- for y, we get have the same effect on physician distribu- tion as a change in the number of physicians, since they all affect the slope of the distribu- tion line. B. Nonlinear Models It is not consistent with known facts to assume that the number of visits per person remains constant as per capita income in- creases. As a rule higher incomes are corre- lated with larger numbers of visits, although the absolute increase is not likely to be very large. We can easily take this relationship into account by expressing the number of visits as a function of income. Let b'Y = number of visits per person per unit of time in the country as a whole, and = / regional number of visits per person per unit of time. The regional physician income then becomes ab'Y~2 (4) and the regional physician-population ratio ab'Y2 (5) = -f- The relationship between regional income and the physician-population ratio is now nonlinear, and the distribution lines are curves convex to the origin. Other things remaining equal, the intro- duction of a positive relationship between visits and income makes the physician distribution more unequal, in the sense of pushing the Lorenz curve to the right. How- ever, as before, an increase in the number of physicians would still leave the relative distribution unaffected and increase the physician-population ratio everywhere by the same percentage.9 As in the previous case, the elasticity of the physician-population ratio with respect to changes in number of physicians equals one. From equation (5) we get Y~ N }7~ p y dR~ Y~ and We can now show that: dR~/dN Y~N YrN 1 -1 R~ / ~ - YP* R~ YP YTN/YP - PAGENO="0479" COMPETITIVE PROBLEMS IN There may also be some question as to whether the fee remains a constant propor- tion of income as income changes. An in- creasing proportion would have the same effect as that described above for an increas- ing number of visits. A decreasing propor- tion, which is more likely, would tend to bend the curves in the opposite direction. We can also get a nonlinear relationship between physician-population ratios and income, and hence a distribution that tends to be more unequal than that of our first model, if we abandon the assumption that physicians try to maximize only income. For the sake of simplicity, let us return to our initial assumptions of constant fees in relation to income and a fixed number of visits per person regardless of income. In equilibrium, all doctors have identical in- comes, but those in low income areas work very long hours, as indicated by the low physician-population ratio, while those in high income areas have considerable leisure. Since there is freedom of movement between areas, it seems likely that some doctors will be tempted to leave the low per capita in- come areas and move to higher per capita income areas, where the same physician incomes are being earned with less work. This would result in a spatial redistribution of physicians and physician incomes. At the new equilibrium, each doctor would maximize the utility derived from both leisure and income. Some doctors may actually move from higher to lower income areas as the influx into high income areas lowers income there and increases leisure. But if doctors in general are willing to work Hence d& 1~?~ dN harder only if they get a higher income,'° there will be at the new equilibrium fewer doctors, earning higher incomes, in low per capita income areas and more doctors, earn- ing lower incomes, in high per capita income areas. Somewhere in the intermediate per capita income range the physician-popula- tion ratio and the physician income would remain unchanged. The extent of the re- distribution depends on the physicians' income-leisure preference pattern. Figure 2 illustrates the impact of income- leisure substitutions for three different num- bers of physicians in the country and initial levels of physician income. The convex curves represent the new equilibrium physi- cian-population ratios in relation to per capita income of the population. In this case, the high income areas have increased the ratios and the low income areas have reduced them. How a change in the number of physi- cians, in fees, or in the number of visits would affect the distribution of physicians under these circumstances cannot be deter- mined a priori without specifying the physi- cians' income-leisure preference pattern. 10 Technically, their marginal rate of substitu- tion of income for leisure is less than zero and leisure is not an inferior good. THE DRUG INDUSTRY 2029 Fie. 2.-Linear and Nonlinear Models of Phy- sician Distribution contrasted. Regional Physi- cian-Population Ratio (R?) increases non-linearly with increasing Regional Per Capita Income (Y?), for various assumed total numbers of Physicians (N). Thus dR? /dN dR~ N ER?.N=~~~/ ~;~i~i; ~2 1 Y2P Y?2/Y2P - 1 PAGENO="0480" 2030 COMPETITIVE PROBLEMS IN THE DRUG INDUSTRY C. Role of Mobility The analysis thus far indicates that, as long as the fee rises with income and the use of physician services per person is fixed, physicians in low income areas will tend to have either less leisure or less income than physicians in high income areas, but not both. This follows in part from the assump- tion of mobility and equal ability. If these assumptions are relaxed the result no longer necessarily holds. Absence of mobility from low to high income areas may leave doctors in these areas with both less income and less leisure than in high income areas. On the other hand, a preference by doctors for high income areas has the same effect as a substi- tution of leisure for income; it tends to in- crease leisure and reduce income in high income areas and cause the opposite in low income areas. The relaxation of the assumption of equal ability may also affect the results obtained. Specialists, who repzesent a group with higher than average qualifications, are likely to draw higher fees than general practi- tioners. Since they choose to locate mainly in high income metropolitan areas, they may raise both the leisure and the average physi- cian income in these areas. D. Summary and Conclusion The main determinants of physician distribution, given the regional distribution of income, were found to be (1) the relation of fees to patient income, (2) the relation of demand for physician services to patient income, and (3) the behavior of physicians with respect to price competition, income maximization, desire for leisure, and geo- graphic mobility. The degree of inequality of physician distribution tends to be greater the greater the increase in fees and in de- mand for services for a given increase in interregional income and the greater the inclination of physicians to substitute leisure for income. An interesting insight provided by the theoretical analysis is that so long as ex- penditures on physician services rise with increases in income, and physicians do not engage in price competition but try to maximize income, changes in the total number of physicians, or in the general level of fees, or in the level of demand for services, do not affect the relative distribution of physicians. However, increases in the num- ber of physicians, in level of fees, and in level of demand would increase the absolute physician-population ratio differentials be- tween high and low income areas and in that sense would make the distribution more unequal. IV. EMPIRICAL ANALYSIS OF DISTRI- BUTION DETERMINANTS We shall now examine to what extent available evidence on physician behavior, fees, and demand for services supports the preceding theoreti~cal explanation of physi- cian distribution. A. Income Maximization, Leisure, and Mo- bility In investigating the behavior pattern of physicians we shall make the assumption, which will be substantiated later, that medical expenditure per person (fees times visits) rises with income. This will tend to give us an uneven distribution of physicians in favor of high income are~ss, unless physi- cians ignoi~e income. There is no evidence of such behavior. Nor are there data avail- able that would allow a positive test of the income maximization hypothesis. The only feasible approach is to test the behavior pattern for its consistency with income maximization and to find out in this manner whether substitution of leisure for income and lack of mobility should be considered significant factors in the observed distribu- tion of physicians. Lack of mobility is in- tended to cover all reasons for not moving other than income and leisure. If leisure is valued by physicians, and if they are fairly mobile, we should find higher physician incomes where leisure is low and PAGENO="0481" COMPETITIVE PROBLEMS lower incomes where leisure is high. This pattern seems to emerge in a test of thirty- two major cities." The test, however, is not based on a direct measure of leisure but assumes that the higher the physician- population ratio the greater the amount of leisure available to physicians. For the cities tested, high physician incomes are correlated with low physician-population ratios (low leisure) and lower physician incomes with higher ratios (more leisure). If, on the other hand, leisure does not increase with higher physician-population ratios, low incomes are not compensated by more leisure. This would most likely indicate a lack of concern with leisure and would make lack of mobility a significant factor in the distributive pattern. In this instance lack of mobility makes the distribution more uneven than it would have been under in- come maximization. If physicians tried to maximize income, there would be a redistri- bution from the low income (high ratio) areas in favor of the areas with higher in- comes and fewer doctors. An investigation of data on physician visits in eight major cities supports the hypothesis that lack of mobility rather than desire for more leisure is the major factor in the distribution of physicians among cities.1' 11 This test uses 1949 income data for physicians and 1950 data on physician distribution. The regression on physician income (y) of the physi- cian-population ratio (x) can be expressed by the equation: y = 22.106 - .556 x, where y is expressed in terms of thousands of dollars per year and x is the number of physicians per hundred thousand population. The coefficient of correlation is .808, and is significant at the .05 level. Income data are from W. Weinfeld, "Income of Physicians, 1929- 1949," Survey of Current Business, July 1951, pp. 9-26. Physician data are from Health Manpower Source Book, Section 4. 12 The test is concerned only with physician visits outside of hospitals, for a period extending from July 1957 to June 1959. If we omit Boston (which is very unusual in that its ratio of physi- cians not in private practice to total physicians is by far the highest of the eight cities), the relation- ship between per capita visits (x) and the physician-population ratio (y) may be expressed by the regression equation: y = -4.57 + 28.83x. The coefficient of correlation is .963. Data on visits are from U. S. Public Health Service, Health IN THE DRUG INDUSTRY 2031 Physicians in those cities having the higher physician-population ratios do not appear to have more leisure than physicians in other cities. Hence, in the case of the 32 cities considered above, it was probably not a desire for greater leisure but some obstacle to mobility that kept physicians from mov- ing to towns where there were fewer doctors. A careful examination of the available data leaves a clear impression that desire for leisure is not a strong motivating force in the physicians' choice of location. Their low valuation of leisure is reflected in the fact that they tend not to reduce the length of their workweek as their income increases. The percentage of doctors working 49 hours or more per week is about the same at all income levels and ranges between 70 and 85." It might be, of course, that physicians at higher income levels take longer and more frequent vacations. When we compare physician incomes among communities of greatly differing sizes it becomes extremely difficult to deduce a rational behavior pattern. We find the lowest physician incomes in the smallest and largest size communities, those with under 2,500 and over 1,000,000 population. These are the areas where physician-popula- tion ratios tend to be the lowest and the highest respectively. We must invoke special kinds of immobility to account for the fact that physicians in very small communities tend to have less income and probably also less leisure than their colleagues elsewhere. Perhaps those in the profession who are least qualified tend to locate there. The high concentration of physicians in very large cities, despite lower incomes, may be partly explainable by cultural and professional Statistics, Series C, No.6, p. 25. The high apparent correlation may be partly spurious in view of the presence of the population factor in the denomi- nators of both the dependent and the independent variables. However, the statistical issue raised by this kind of ratio correlation remains unsettled at present. See J. R. Meyer and E. Kub, "Correlation and Regression Estimates when the Data are Ratios," Econometrica, October 1955, pp. 400-416. 13 Health Manpower Source Book, Section 5, pp. 79-81. 81-280 0-68--pt. 5-31 PAGENO="0482" 2032 COMPETITIVE PROBLEMS IN THE DRUG INDUSTRY advantages and by opportunities to practice certain specialties not available elsewhere, which constitute another form of immobility. In our estimation, improved mobility, in the direction indicated by income maximi- zation, would result in a more even distribu- tion of physicians among large cities, but might very well result in a more uneven distribution between large cities and smaller communities. Whether either redistribution would be desirable in terms of improving the relative availability of medical services depends, of course, on the structure of the demand for services. B. Demand for Phyrfcian Services The demand for physician services is relevant both as a factor affecting the dis- tribution of physicians and as a measure of the utilization of physician services. Under perfectly competitive conditions, assuming that doctors are of equal ability and wish to maximize income, there would be regional differences in physician-population ratios so long as there are regional differences in demand for services. Under these conditions, however, differences in ratios would corre- spond exactly to differencesin use of services. In equilibrium physicians everywhere would have the same income and provide the same amount of services. No region would suffer from a relative shortage of physician serv- ices, if use of services could be taken as an index of need. On the other hand, if there are substantial regional differences in use of services per physician, without correspond- ing differences in physician income, we may have an indication of maldistribution of physicians. We shall examine first the factors deter- mining the use of physician services and then the regional use of services in relation to regional physician-population ratios. The major determinants of the use of physician services are income, age, sex, race, education, place of residence, and possession of health insurance. We need not bother with the use of services as related to sex in a regional analysis. All the other factors tend to be correlated with income. The effect of a person's family income on the number of physician visits per year is an increase of 24 per cent from families with annual incomes under $2,000 to families with incomes of $7,000 and over.'4 In both the "under $2,000" and the "$2,000-$3,999" income groups, the number of visits per person per year is the same, namely 4.6; in the "$4,000- $6,999" income group it is 5.1, and in the "$7,000 and over" group, 5.7. These are averages for entire income groups; the differences are greatest if only children and young persons in different income groups are compared, and smallest in comparisons of only middle-aged persons. The tendency toward larger number of visits as income rises was detected also in a comparison of cities with different levels of per capita income. The relationship between annual per capita physician visits and per capita effective buying income was studied for eight standard metropolitan areas. The coefficient of correlation wa~ found to be .619 and the coefficient of determination .383. This correlation is significant at the .1 level. The reasons why persons with high in- comes visit physicians more frequently than those with low incomes are no doubt more complex than merely greater ability to pay. The visit differentials should be related to the economic value of good health, percep- tion of illness, and easy access to physicians. These factors are in turn correlated with economic opportunity, education, and place of residence. In 1958-59, whites visited their physicians 35 per cent more often than non- whites. For families whose head had a college education, the number of visits per person per year was 56 per cent higher than for families whose head had under five years of education and 42 per cent higher than for families whose head had five to eight years of education.1' For the country as a whole `~ Health Statistics, Series B, No. 19, p. 20. "Ibid., p.13. PAGENO="0483" COMPETITIVE PROBLEMS the number of visits per capita in urban areas (towns with 2,500 inhabitants or more) was about 39 per cent higher than in rural farm areas, and about eight per cent higher than in remaining rural areas. Another factor correlated with income is the possession of health insurance,'6 which tends to increase the demand for medical services. People who possess health insurance tend to have a higher number of physician visits, higher hospital admission rates and higher surgical rates than people without insurance at comparable income levels.'7 This is true not only for services that are covered by insurance but also for those that are not covered. People seem to be more inclined to see a physician if at least some of the potential expense is prepaid. It has been established that both the physician-population ratio and the number of visits per person increase with income. The question now is whether the demand for services and the relative supply of phy- sicians increase in the same proportion, which would mean an uneven distribution of physicians but an even distribution of work loads. The available data indicate that regional differences in use of physician services are of a lower order of magnitude than differences in physician-population ratios. We computed the number of non- hospital visits per private non-hospital phy- sician by multiplying the regional average number of visits per person by total regional population and dividing by the regional number of physicians. This measure can be taken as a reliable index of work load, for the physicians concerned, unless there are 16 Possession of any kind of health insurance is about three times more frequent in families with: incomes of $7,000 and over than in families with incomes of under $2,000. See U. S. Department of Health, Education, and Welfare, Medical Care Financing and Utilization, Health Economics Series No. 1 (Washington, D. C.: U. S. Govern- ment Printing Office, 1962), Table 87, p. 99. 17 0. W. Anderson and J. J. Feldman, Family Medical Cost and Voluntary Health Insurance: A Nationwide Survey (New York: McGraw-Hill, 1960), pp. 73, 183, 194. IN THE DRUG INDUSTRY 2033 important regional differences in the amount of: inhospital visits per physician. The re- sults, presented in Table I, show a large work load differential between the region with the highest and the lowest number of visits. Physicians in the West South Central area (Arkansas, Louisiana, Oklahoma, Texas) have 74 per cent more visits of all kinds outside of hospitals than those in New England and 37 per cent more than those in the Middle Atlantic region. Although New England may be a special case, there is a general tendency for the number of visits to increase with decreases in the physi- cian-population ratio. Apparently, doctors in areas with low physician-population ratios either work longer hours or work more intensively. Contrary to what seemed to be indicated in the earlier comparison of metro- politan centers, the census area comparison indicates a tendency toward a possible redundancy of physicians in high ratio areas and toward a shortage in low ratio areas. However, census divisions are much less homogeneous than metropolitan areas, and TABLE I REGIONAL PsIYSIcIAN WoRE LOADS IN THE U. 5., 1959 Region Non- hospital physicl'an In~ne capita Physi- ci~s0~r popula- Average income ~ eraiphy- 1949 New England Middle Atlantic Pacific East North Cen- tral West North Cen- tral Mountain EastSouthCentral South Atlantic West South Cen- tral 3,817 4,847 5,284 5,694 5,986 6,042 6,286 6,499 6,630 2,396 2,543 2,502 2,337 1,978 1,994 1,424 1,804 1,764 113 120 112 83 79 84 65 74 75 $ 9,442 9,574 12,782 12,158 11,961 11,214 11,325 11,131 11,794 Sources: Statistical Abstract of the U.S., 1960 Table 6; 1961, Table 419. Health Manpower Source Booh, Section 10, Table 2; Health Statistics, Series C, No. 6, Table ii; Survey of Current Business, July 1951, Table 12, p.19. PAGENO="0484" 2034 COMPETITIVE PROBLEMS IN THE DRUG INDUSTRY 10 TABLE II PITY5ICIAN HOSPITAL WORK LOADS: 1959 Area . ~ population Hospital discharges ph~ian Patient hospital Phjsic~n Northeast West North Central South 149.0 128.2 103.4 91.4 71.3 94.7 113.6 127.5 728 729 966 930 Sources: Health Manpower Source Book, Sec. 10, pp. 5-6; Health Statieticz, Series B-No. 32, pp. 18-19. there is at least the possibility that in higher income areas physicians spent more time taking care of patients in hospitals than in low income areas, which would tend to offset the disparities noted above. But the opposite is more likely to be the case; that is, in low income regions doctors have probably higher hospital work loads than in high income regions. It was possible to compute the relationships between the number of physicians and the number of hospital discharges and the number of pa- tient hospital days for the four major census divisions of the country. The physicians used for this purpose include all active nonfederal physicians in both private prac- tice and in the hospital service. The results of the computations are presented in Table II, which shows the tendency of higher hospital work loads in areas with low physi- cian-population ratios. The regional uneven- ness in work loads observed above thus tends to be accentuated when hospital work is taken into account. We may also conclude that while differ- ences in physical demand for physician services may account in part for the observed regional differences in physician-population ratios, they are not a sufficient explanation of existing disparities in the supply of physi- cians. C. Fees for Physician Services It was shown theoretically that high physician-population ratios in high income areas may be due in part to the fact that fees are based on per capita income rather than on the supply and demand for physician services. It is not necessary, of course, that there be a complete absence of price competi- tion, as was assumed in the theoretical model. A modest amount of price competi- tion may well exist, which would keep fees from rising as much with increases in regional income as they would otherwise. Some indication that fees are lower in low income areas is implicit in the data presented above in Table I, unless the relationships between regional physician incomes changed considerably between 1949 and 1959. Table I implies that the average income per visit in New England and the Pacific was between 35 and 40 per cent higher than in the south- ern areas. The East North Central area income per visit was about 20 per cent higher than in the South. There are no relevant data available on physician fees,18 but we can remedy this deficiency in part by using another indirect method. The available data on spending and on number of visits make it possible to compute an index of fee differentials between income groups and regions. For the United States as a whole we have for a twelve- month period in 1957-58 an estimated $25 spent on physician services per capita by families earning under $2,000 annually;19 this income group is reported to have had 4.6 visits per capita during a twelve-month period in 1957-59, which indicates an ex- penditure of 85.43 per visit. By the same process we obtain an expenditure of $6.93 per visit for the 87,500 and over income 18 The Bureau of Labor Statistics publishes data on certain types of fees paid by comparable wage and salary earner families in twenty major cities. These data, unfortunately, are not suitable for our purposes, since we are interested in com- parisons between income groups. 19 One third of $75 spent on all health services. See 0. W. Anderson, P. Collette, and J. J. Feld- man, Family Expenditure Patterns for Personal Health Service, 1953 and 1958, Health Information Foundation Research Series No. 14 (New York: Health Information Foundation, 1960), Table 4a, p.S. PAGENO="0485" class. The highest income class paid on the average 28 per cent more per visit than the lowest. This is not a measure of price dis- crimination in the ordinary sense, since it does not refer to a single service but to a composite bundle of services bought by different groups. There may be differences in kind as well as quality between bundles. There may be more homogeneity i,n a comparison of expenditures of people living in similar environments. A calculation for farm families only gave the following ex- penditure differentials: under $2,000 family income, $4.01 per physician visit; $2,000- 3,999 income, $4.35 per visit; $4,000-6,999 income, $4.74 per visit; $7,000 and over income, $5.71 per visit.'5 In this case, the highest income group paid 42 per cent more per visit than the lowest. It is noteworthy that the heaviest discrimination occurs be- tween the highest and the next highest in- come group. A computation by region shows farm families in the South paying $4.13 per visit against $4.82 in the North Central area. In spite of the necessarily approximate nature of our statistical results, all indica- tions are that prevailing pricing practices contribute to the observed differences in physician-population ratios between areas with different income levels and between urban and rural areas. v. CONCLUSIONS It has been shown that high income areas have substantially larger numbers of pbysi- cians in relation to population than low income areas. If we assume that in a wealthy country the entire population should have reasonable access to physician services, the implications of this i~ncome-related differ- ential deserve careful consideration. Although there are other factors making high income areas attractive to physicians, 20 Expenditure data from J. L. Pennock, "Farm Medical Care Expenditures," Public Health Re- ports, April 1958, p. 290; visits are adjusted for rural level. COMPETITIVE PROBLEMS IN THE DRUG INDUSTRY 2035 there are definite economic incentives work- ing in the same direction. Persons with higher incomes tend to visit physicians more often and have more expensive visits. The larger number of visits per capita in higher income areas tends to justify a certain unevenness in the distribution of physicians in relation to population, but high income areas tend to have more physicians than would be called for by this difference in use of services. The ability to charge higher fees to more well-to-do clients makes possible a high concentration of physicians in high per capita income areas without commensurate loss in physician income. In this sense, the widely defended practice of charging accord- ing to ability to pay may have undesirable consequences in terms of the spatial avail- ability of physician services.2' The avail- ability of services should be considered not only in terms of access to a physician but also in terms of the amount of the physi- clan's time available to the patient on each visit. It was shown that where the physician- population ratio is low, physicians tend to have very large numbers of visits which are almost necessarily brief. The economic implications of the observed wide regional differentials in physician work loads depend on how health is valued. If we value health from a strictly economic point of view, it is in all probability more advisable to spend a lot of physician time on high income executives, who presumably have a high productivity, rather than on low income unskilled workers. In this case, the observed spatial distribution of physicians may well be optimal, or perhaps there should still be more physicians in high income areas and fewer in low income areas. This approach, however, becomes very awkward when we extend it to the children of various income groups. To the extent that the children have comparable productive potentials, economic 21 Within a given area, of course, this practice may make medical care available to some low income groups who would otherwise be deprived of it. PAGENO="0486" 2036 COMPETITIVE PROBLEMS IN THE DRUG INDUSTRY logic suggests that they should have com- parable access to medical care. If we value health from a social point of view and put the same value on the health of all individuals, the economic problem becomes essentially that of adjusting the available supply of physicians as efficiently as possible to the existing need. In the absence of any data on actual need, let us use expressed need, as represented by the number of visits. On these assumptions it can be argued that the existing pattern of physician distribution involves an ineffi- cient allocation of medical manpower re- sources. To the extent that wide disparities in regional work loads involve substantial differences in the amount of time a physician can devote to each patient, there is a tend- ency toward inefficiency. The assumed social values make the social utility of a compara- ble visit the same for all individuals. From the medical point of view, it is probably reasonable to assume that the effectiveness of a visit does not increase proportionally with the length of the visit beyond a fairly short minimum period. Physician resources made available for long visits are therefore not employed most efficiently. This does not even take into account the fact that where long visits are possible, the visits themselves may be less urgent. Further, where physi- cians are overloaded, the pressure of time to accommodate many patients may impair a doctor's ability to function with optimal efficiency. Finally, there may be economic inefficiency involved with regard to patients, who may have to waste a great deal of time in the waiting rooms of overloaded doctors. An efficient spatial allocation of physician manpower would therefore be one which tends to equalize the average allowable time per visit between all areas. A few brief observations are in order also with regard to lack of mobility, which is clearly an important factor in the existing distribution pattern. It is rather doubtful that the existing distribution would be more equal if physicians were more mobile, in the sense that they would be more inclined to maximize either income or some combina- tion of income and leisure. Many of the poorer areas of the country would undoubt- edly be even worse off if all physicians made their location decisions with strict economic rationality. Economic incentives and the lack of mo- bility also have implications for the improve- ment cf the geographic distribution of physi- cians. It is unlikely, at least in the short run, that the distribution can be made significantly more even for the country as a whole by simply increasing the manpower output of our medical schools. It is very doubtful whether an increase, perhaps even a substantial increase, in the supply of physi- cians would automatically solve the distribu- tion problem, as some medical authorities seem to think.22 Such a measure might result in an increase in the number of physicians in areas where there is a shortage, but this would tend to be accompanied by an even greater increase in areas where there is already a relative excess supply of doctors. This does not promise to be an economical approach to the problem of distribution. On the other hand, there is good reason to believe that the unevenness in the distribu- tion of physicians will diminish automati- cally, in the long run, along with the tend- ency toward regional equalization of per capita income. 22 See for instance the statement by Dr. P. R. Hawley in Committee on Medical Care Teaching, Readings in Medical Care (Chapel Hill, N. C.: University ~f North Carolina Press, 1958), p. 27. PAGENO="0487" COMPETITIVE PROBLEMS IN THE DRUG INDUSTRY 2037 Senator NELSON. I would like to submit for the record at this point a letter from E. B. Anderson, Hoffman-La Roche Co. (The letter referred to follows:) HOFFMANN-LA ROCHE INC., Nutley, N.J., January 2, 1968. Hon. GAYLORD NELSON, Chairman, Senate Monopoly Subcommittee, U.S. Senate, Washington, D.C. DEAR SENATOR NELSON: I want to thank you for the opportunity which you have given us to meet with your staff concerning recent statements which have been issued on the prices paid by certain government authorities for two Roche products. It was most helpful to review together the data which had been accumulated. Hoffmann-La Roche Inc. shares your belief in the desirability of uniform pricing policies under which government agencies purchasing equal quantities of the same drug product would receive the same price. For some time Roche has had in effect a pricing policy designed to meet this objective, under which any Federal, State or local government which purchases from Roche is charged the same price for the same quantity of the same product. On December 19, 1967, you issued a public statement and a price table which referred to the prices of several drug products, two of which are sold by Hoffmann-La Roche Inc., chiordiazepoxide hydrochloride and sulfisoxazole. Since several figures contained in the statement and in the survey data released at the same time are incorrect, I am writing, to you to provide the correct informa- tion, as I am sure you want the record to be as accurate as possible. We also believe that some explanation of other items in the statement and the price table would permit a more meaningful comparison of prices for these two products. In addition to the statement in the hearing, there was an additional state- ment by your office released on Decrembér 24, 1967, relating to the prices paid by various government purchasers for LIBRITJM capsules (chlordiazepoxide hydrochloride). This later statement, as reported in the New York Times of December 25, 1967, and several other papers, also contains some incorrect figures and conclusions which we are sure you would like to see corrected. In several instances the statements and the table have confused prices for 25 my. capsules of chlordiazepoxide hydrochloride with prices for 10 mg. capsules of the same drug. As a result, the prices for different dosages have been com- pared, which is obviously misleading. A. Chiordiazepoccide hydrochloride We have included hereafter the items in the statement and the price table of December 19, and the statement of December 24, relating to chlordiazepoxide hydrochloride which are not in accord with the facts, or which we believe leave an erroneous impression which should be corrected. 1. In regard to the City of Des Moines, the table states that the city pays the highest price of $45.00. While it does not state that Roche sold the product to Des Moines at this price, some persons may have so inferred. Actually, there are no direct sales of chiordiazepoxide hydrochloride to the City of Des Moines by Hoffmann-La Roche Inc. Since Roche has had no direct dealings with Des- Moines for this product, the city's purchase of this product would have been from a wholesaler or retailer. As in the case of Des Moines, the cities of Cleve- land and Paterson do not buy directly from Roche. Thus, the prices paid by those cities for Librium are presumably paid to wholesale or retail distributors. Roche has made some direct sales to county hospitals where these cities are located, but the prices in the table and statement are not the prices which we have charged for these sales. Broadlawns Polk County Hospital (Des Moines) and Highland View Hospital in CuyahOga County (Cleveland) have purchased Librium 25 mg. capsules in bottles of 500 for $89.60, which is our price to direct government purchasers which do not buy the 5,000-capsule bottle or in quanti- ties large enough .to support a contract. Preakness Hospital in Passaic County (Paterson New Jersey) has purchased Libriurn 25 mg. capsules in bottles of 50 capsules at a price of $4.40 per bottle which is our price for that size bottle. The District of Columbia, we understand, purchases its requirements of Librium from a Veterans Administration depot. PAGENO="0488" 2038 COMPETITIVE PROBLEMS IN THE DRUG INDUSTRY I am sure you understand that the manufacturers of many different types of products cannot sell directly to all end-users, and, as a result, the prices paid by end-users will vary. In addition to direct sales to large-volume users, it is essential that Roche sell through drug wholesalers and retailers in order to serve adequately all of the small-volume purchasers. We, of course, have no control over the prices charged by other persons who distribute our products after we sell to them. 2. The information you received from the City of Philadelphia stated that Philadelphia paid $18.50 for 500 25 mg. chlordiazepoxide hydrochloride capsules. Our contract with Philadelphia signed on June 12, 1967, in fact specifies a price of $28.50 per 500 25 mg. capsules for a quantity in excess of 100,000 capsules. The price quoted in the statement, namely $18.50, refers to bottles of 500 10 mg. capsules for a quantity in excess of 250,000 capsules. In addition, it was stated that New York City paid $28.50 per 50025 mg. capsules of chlordiazepoxide hydro- chloride. That was our previous price to New York City. The current contract, which was effective October 1, 1967, calls for $25.50 per 500 25 mg. capsules, based on a quantity in excess of 500,000 capsules. 3. The statement noted that the Defense Supply Agency paid $13.20 per 500 25 mg. capsules of chlordia.zepoxide hydrochloride. Actually, the price of $13.20 per 500 capsules is for 10 fig, capsules under a contract for over 20 million capsules. Our contract with the Defense Supply Agency for capsules of the 25 mg. strength calls for a price of $24.00 per 500 capsules in a total purchase of over 3,000,000 capsules. 4. Your information that Philadelphia got a better price than the Veterans Administration for 25 mg. capsules of chlordiazepoxide hydrochloride is incor- rect. As noted above, Philadelphia paid Roche $28.50 per 500 25 rng. capsules, based on the purchase of over 100,000 capsules. The Veterans Administration paid $21.75 per 500 25 mg. capsules based on a contract purchase of over 5,000,000 capsules. As can be seen, the Veterans Administration pays a lower unit price for capsules of the equal strength because it buys in greater volume. 5. The article also states that "Los Angeles paid a premium price even though it bought three times the quantity Philadelphia bought." This statement is incorrect. As the following table shows, for 10 mg. capsules Los Angeles paid Roche a lower price based on a quantity purchase of over 1,000,000 capsules than Philadelphia paid based on a quantity of over 250,000; and for 25 mg. capsules Philadelphia paid Roche $28.50 based on a quantity of 100,000 and Los Angeles paid $25.50 based on a quantity of 500,000: Philadelphia 10 milligrams $1850 25 milligrams 2850 Los Angeles 10 milligrams 16.50 25 milligrams 25. 50 6. The prices liSted in the article for the cities not covered above are also con- sistent with Roche's policy of equal prices for government buyers which purchase equal quantities of the same drug from Roche. The cities of Wichita, Kansas; Newark, New Jersey; and Grand Rapids, Michigan, all have paid $39.60 per 500 capsules of the 25 mg. strength of LIBRI1JM. This is the Roche price for a bottle of 500 25 mg. capsules, to direct government purchasers who do not buy large sizes, such as the bottle of 5,000 capsules, or annual quantities large enough to justify a contract purchase. B. $ulfjsoa,azole The statement and price table of December 19 also contain information on sulfisoxazole which we believe may lead to some erroneous inferences and there- fore further explanation. 1. In regard to the City of Des Moines, the table states that the city paid $25.30 per bottle of 1,000 0.5 grain sulfisoxazole tablets. Roche does not sell this product directly to Des Moines. Purchases of Roche's sulfisoxazole by the City of Des Moines would have been from a wholesaler or retailer, or Des Moines may have purchased sulfisoxazole from another company. 2. The table quotes a price of $6.90 for a bottle of 1,000 0.5 gram sulfisoxazole tablets to the City of Philadelphia. The current Roche contract for sulfisoxazole with Philadelphia, dated~ June 12, 1967, calls for a price of $9.00 per bottle of 1,000 tablets based on a quantity of 1,000,000 tablets. PAGENO="0489" COMPETITIVE PROBLEMS IN THE DRUG INDUSTRY 2039 3. The prices cited in the table of $8.35 and $8.50 per 1,000 0.5 gram sulfisoxazole tablets paid to Roche by the Defense Supply Agency and the Veterans Adminis- tration, respectively, are below that paid *by Philadelphia because they are supplied under contracts for very much larger quantities. As you can see in the above items, much of the information upon which you relied regarding the prices for Roche products is inaccurate. We believe that, left uncorrected, these inaccuracies reflect adversely on the reputation and poli- cies of Roche, and upon your Subcommittee. Accordingly, we respectfully request that the records of your Subcommittee be corrected to reflect the facts regarding the Roche prices and policies as stated herein. We appreciate the cooperation which both you and your staff have given us in reviewing this data. Very truly yours, E. B. ANDERSON, Vice President. Senator NELSON. The committee will adjourn until next Thursday, January 25, in room 457. (Whereupon, at 1 :15 p.m., the hearing in the above-entitled matter was adjourned, to reconvene on Thursday, January 25, 1968, in room 457.) PAGENO="0490" PAGENO="0491" COMPETITIVE PROBLEMS IN THE DRUG INDUSTRY THURSDAY, JANUAB,Y 25, 1968 U.S. SENAIT, MONOPOLY SUBCOMMIITEE OF THE SELECT COMMrITEE ON SMALL BUSINESS, Wathi'ngton, D.C. The subcommittee met, pursuant to recess, at 10:10 a.m., in room 457, Old Senate Office Building, Senator Gaylord P. Nelson (chairman of the subcommittee) presiding. Present: Senator Nelson. Also present: Benjamin Gordon, staff economist; James H. Gross- man, minority counsel; Susan H. Hewman, research assistant; and William B. Cherkasky, legislative director, staff of Senator Nelson. Senator NELSON. I now open: the hearings of the Monopoly Subcommittee. Our witness today is Dr. William Comanor. Dr. Comanor, we appreciate very much your taking the time to come here today to present your testimony to the committee. I have read through your testimony and I think it represents a very thought- ful, valuable contribution to the record of the hearings that we are conducting here on the pricing structure in the drug industry. I notice that you have a biographical sketch. That will be printed in full in the record preceding your testimony. If you would like to just state extemporaneously your background and your present position, then you may proceed with your statement. I trust that you have no objection to interruptions for questions? STATEMENT OP WILLIAM S. COMANOR, PH. D., PROFESSOR, DEPART- MENT OF ECONOMICS, HARVARD UNIVERSITY, CAMBRIDGE, MASS. Dr. COMANOR. Not at all (The biographical sketch follows:) (JIIERIODIUM VITAE Name: William S. Comanor. Address: Residence-10 Fernald Drive, Cambridge, Massachusetts, 02138. Office-1746 Cambridge Street, Cambridge, Massachusetts, 02138. Date of birth: May 11, 1937. Birthplace: Philadelphia, Pennsylvania. Marital status: Married, no children. Nationality: United States. 2041 PAGENO="0492" 2042 COMPETITIVE PROBLEMS IN THE DRUG INDUSTRY Education: 1951-1955 Central High School, Philadelphia, Pennsylvania. 1955-1957 Williams College, Williamstown, Massachusetts. 1957-1959 Haverford College, Haverford, Pa., A.B. in Economics. 1959-1963 Harvard University, Cambridge, Mass., Ph. D. in Economics. 1963-1964 London School of Economics, London, England. Thesis title: "The Economics of Research and Development in the Pharmaceuti- cal Industry." Fields of interest: Industrial Organization and Public Policy. Economics of Science and Research. Statistics and Econometrics. Economic Theory. Honors and Fellowships: Sophomore Honors, 1957. High Honors in Economics, 1959. Phi Beta Kappa, 1959. Harvard University Fellowship, 1959. National Science Foundation Fellowship, 1963. Sloan Foundation Research Grant, 1966. Professional experience: Teaching Fellow, Harvard University, 1961-1963. Resident Tutor, Quincy House, Harvard University, 1961-1963, 1964-1965. Economist, Caterpillar Tractor Co., Peoria, Ill., Summer 1960. Economist, United Research, Inc., Cambridge, Mass., Summer 1961. Research Assistant, Professor Richard E. Oaves, Harvard University, Sum- mer 1963. Instructor in Economics, Harvard University, 1964-1965. Special Economic Assistant to the Assistant Attorney General, Antitrust Division, U.S. Department of Justice, Washington, D.C., 1965-1966. Assistant Professor of Economics, Harvard University, 1966. Consultant, Booz, Allen, and Hamilton, Inc., 1966-1967. Consultant, Antitrust Division, U.S. Department of Justice, 1966. Publications: "Research and Competitive Product Differentiation in the Pharmaceutical Industry in the United States," Econornica, November 1964. "Review and Technical Change in the Pharmaceutical Industry," Review of Eccrnomics and Statistics, May 1965. "The Drug Industry and 1~Iedical Research: The Economics of the Kefauver Committee Investigation," Journal of Bnsiness, January 1966. "Competition and the Performance of the Midwestern Coal Industry," Jour- nal of Indnstrial Economics, July 1966. "Vertical Mergers, Market Power, and the Antitrust Laws," American Eco- nomic Review, May 1967. "Advertising Market Structure and Performance," Review of Economics and ~tatisti~s, November 1967 (with Thomas A. Wilson). "Market Structure, Product Differentiation, and Industrial Research," Quarterly Journal of Economics, November 1967. Senator NELSON. Go ahead, Doctor. Dr. COMANOR. My name is William S. Comanor. I am presently as- * sistant professor for economics at Harvard University. I received my A.B. from Haverford College in 1959 and my Ph. D. from Harvard in 1964. During the academic year 1965-66, I served with the Department of Justice as Special Economic Assistant to the Assistant Attorney General in charge of the Antitrust Division. I currently serve as a consultant to the Antitrust Division of the Department of Justice. My field of special interest within economics is industrial organiza- tion, problems of competition and monopoly, and government policy toward business. For sometime now, I have been especially concerned with the economics of the pharmaceutical industry. My Ph. D. thesis was entitled "The Economics of Research and Development in the Pharmaceutical Industry." From my research, I have published three PAGENO="0493" COMPETITIVE PROBLEMS IN THE DRUG INDUSTRY 2043 articles dealing with the economics of the drug industry. The first is entitled "Research and Competitive Product Differentiation in the Pharrnaceut1cal Industry in the United States."1 It appeared in the 1964 issue of Economica. The second is entitled "Research and Tech- nical Change in the Pharmaceutical Industry," 2 and appeared in the May 1965, issue of the Review of Economics and Statistics. The third is entitled "The Drug Industry and Medical Research: The Economics of the Kefauver Committee Investigation,"3 and appeared in the Jan- uary 1966, issue of the Journal of Business, published by the Univer- sity of Chicago. I have copies of all of these articles. You may wish to include t.hem in the record. Senator NELSON. The articles you have mentioned will be printed in the record at the conclusion of your testimony, along with the supple- mental study you have furnished the committee, entitled "Advertising Market Structure and Performance," ~ which I understand is to be published. Dr. COMANOR. That is right. I will mention that. Senator NELSON. That study will also be printed at the conclusion of your testimony. Dr. COMANOR. I would like to report this morning on some findings from a study which Prof. Thomas A. Wilson, of the University of Toronto, and I have been carrying out on the relationship between advertising and profit rates in the American economy, and then I want to comment on some of the implications of the study for competi- tion and public policy toward the pharmaceutical industry. The statistical results are contained in a paper entitled "Advertising Market Structure and Performance," which will be published in the November, 1967, issue of the Review of Economics and Statistics. I am afraid that although economists advocate efficiency for others, they are frequently as lax in their own operations as anyone else, and the November, 1967, issue of this review is not scheduled for publication until February and March of t.he current year. THE STATISTICAL FINDINGS The core of this paper is a statistical analysis of the relationship between advertising and profit rates across a wide range of consumer goods industries. Our analysis was limited to this sector of the economy for it is only here that advertising seems to have a major competitive impact. Even among consumer goods industries, advertising seems far more important in some industries than others. In most industries, adver- tising expenditures account for a relatively low proportion of total revenues while in a handful of others, they reach very high propor- tions. Thus, in the 41 consumer goods industries examined in our study, 25 had advertising-sales ratios below 3 percent, eight had ratios be- tween 3 percent and 6 percent, and only six industries had ratios which exceeded 6 percent. 1 See article, p. 2069, infra. 2 See article, p. 2078, infra. ~ See article, p. 2086, infra. ~ See article, p. 2092, Infra. PAGENO="0494" 2044 COMPETITIVE PROBLEMS IN THE DRUG INDUSTRY In the latter group, perfumes had an advertising-sales ratio of 15 percent, cereals and drugs were approximately 10 percent each, soap 9 percent, beer 7 percent, and soft drinks slightly more than 6 percent. At lower levels, cigarettes and wines had advertising-sales ratios of about 5 percent. These findings suggest that the economic effects of advertising expenditures are not likely to be felt generally throughout the economy but rather are limited to a small subset of industries. Senator NELSON. Doctor, I notice that you comment on this later on in your testimony, but the figure that you use for expenditures for advertising of drugs is 10 percent of sales. What you are saying is that the total expenditure for advertising for drugs equals 10 percent of the total sales of drugs, is that correct? Dr. COMANOR. That is right. Senator NELSON. And in this figure, you are using both prescription and nonprescription; that is, proprietary drugs. Dr. COMANOR. That is right. Senator NELSON. What is the total amount spent on prescription and proprietary drugs? Dr. COMANOR. I have not looked at the most recent figures, it is something in the order of $2.5 billion, $3 billion, but this could be checked. Senator NELSON. This figure confuses me. My memory is that we have had testimony that somewhere around $4 to $4.5 billion is spent on prescription drugs alone in this country. Dr. COMANOR. It has been some time since I looked at the specific figure that you are asking. I may be completely wrong. I would like to check the numbers before I answer t.he question. Senator NELSON. Do you have those figures in your study? Dr. COMANOR. The figures which you ask are n'ot contained in the printed paper which I am presenting to you, but I have them in Cambridge. I can certainly send them to you. Senator NELsoN. I would appreciate your getting these figures for the committee, because it seems to me quite a critical statistic. If I recall the testimony correctly, $4.5 billion is now spent on prescrip- tion drugs; that is one figure. I would like to know what the figure is on proprietary drugs. Then the next aspect of the question is, it seems to me, the critical aspect of advertising in the drug industry; that is, advertising for prescription drugs, which is directed only to the medical profession. I would like to know what percentage the advertising cost as a percentage of sales for prescription drugs. Again the figure that we have here is that $800 million is spent on advertising of prescription drugs, not counting detail men; $8(X) million as a percent of $4.5 billion is a much higher advertising figure than the figure you use of 10 percent. I may be wrong in my recollection on the statistics. Dr. COMANOR. The 10-percent figure applies both to proprietary as well as prescription drugs. As I note later in my statement, it applies to the average values for the years 1q54 through 1957. So we are dealing with a period over 10 years ago. PAGENO="0495" COMPETITIVE PROBLEMS IN THE DRUG INDUSTRY 2045 Senator NELSON. Your statistics apply to what period? Dr. cOMANOR. They apply to average values for the years 1954 through 1957. I state later in my testimony precisely why this period was chosen. They are not the most recent figures that one could find on this particular percentage in the pharmaceutical industry. Senator NELSON. The reason it seems to me that it would be valu- able to have the statistics separated out with a separate set of statistics for proprietary and prescription drugs is that they present two very different problems in terms of the difficulty in entering the market- place that you talk about. It seems to me that it is one kind of problem to get into the marketplace and compete in the sale of a prescription drug when all of the advertising of that drug is aimed solely at the medical profession and quite another problem to get into the market- place with a proprietary drug where all the advertising, or almost all of it, is aimed at the general public. Dr. COMANOR. I agree with you completely. I think these two sepa- rate industries create very different problems for public policy. This figure was used because the data was obtained from the Statistical Division of the Internal Revenue Service and the IRS publishes data for the two industries combined. The only tax return data that are available to the general public for this particular industry combine both proprietary and prescription drugs. So that although for public policy purposes, I agree with you completely, for the purposes of our study, we were forced to use the combined industry. Senator NELSON. That is because the profit figures- Dr. COMANOR. The profit figures also apply to the combined indus- try. This is the way the IRS publishes tax return data. Senator NELSON. But the statistics are available, are they not, for the total amount of proprietary drugs sold and the total amount of prescription drugs sold? Dr. COMANOR. That is right, but we wanted to undertake a study dealing with consumer goods industries generally throughout the economy. Senator NELSON. I realize your study was not directed at precisely all of the points that are being raised here. But for purposes of the record, if the statistics are easily available, I would appreciate it if you would submit the most current statistics that you can get on total sales of prescription drugs and what percent of sales is represented by advertising for prescription drugs and the same for proprietary drugs. I think it will probably show that the percentage of advertis- ing to sales of prescription drugs is much higher than this average of 10 percent that you have given us for all drugs. Dr. COMANOR. I will try to get the statistics for you. Senator NELSON. -If it is easily available to you, if you would submit that for the record in the next week or 10 days so we could print it.at this stage of the testimony, I would like that. Dr. COMANOR. Fine. PAGENO="0496" 2046 COMPETITIVE PROBLEMS LN THE DRUG INDUSTRY (The material referred to, subsequently received, follows:) HARVARD UNIvimSITY, DEPARTMENT OF ECONOMICS, Cambridge, Mass., February 8, 1968. Senator GAYLORD NELSON, - Chairman, Monopoly Subcommittee, Select Committee on Small Business, U.S. Senate, Washington, D.C. DEAR SENATOR NELSON: During my testimony before the Monopoly Subcom- mittee on January 25, 1968, you requested that I submit some additional infor- mation to be included in the record. In the1963 Census of Manufacturers, Pharmaceutical Preparations (Standard Industrial Classification 2834) were divided into two classes: ethical and pro- prietary. Domestic shipments of ethical preparations were $2,054,897,000, while domestic shipments of proprietary preparations were $836,166,000. I am afraid, however, that total advertising expenditures in these two industries are not available from public sources. For the same year, Internal Revenue Service statistics report that total business receipts in the drug industry (IRS Minor Industry 2830) were $4,505,870,000 while total advertising expenditures for the same firms were $405,007,000, or 9.0% of sales. For the following year, 1964, the IRS listed a smaller number of firms in this industry, and business receipts for these firms were $4,175,331,000, while total advertising outlays amounted to $461,648,000, or 11.1% of sales. The drug industry, as defined in IRS statistics, include sales of both ethical and proprietary products. Furthermore, the IRS statistics include advertising outlays and revenues for entire firms even though some firms may have con- siderable operations which are not related to the productions and distribution of drugs. Census statistics, on the other hand, are based on data gathered at the establishment level, and therefore include only business operations which deal specifically with a particular industry. As I noted in my testimony, the Statistics Division of the Internal Revenue Service accepts the definition of advertising which is made by individual firms in their tax returns. Different firms, moreover, appear to list different categories of expense under different headings, so that gifts for doctors and medical students are included in advertising expense by some firms while others list these outlays under promotional expense. Furthermore, I understand that contrary to my statement on January 25, the cost of samples are sometimes included in adver- tising expense. Thus, it appears that reported outlays on advertising understate the true volume of advertising and promotional effort in the drug industry even apart from the salaries paid to detail men. Sincerely, WILLIAM OOMANOR. Dr. COMANOR. In most industries, even in the consumer goods sec- tor, advertising probably plays a relatively minor role, although it appears to play a major role in a few others. One further feature of these statistics is that the subset of indus- tries with high advertising-sale~s ratios is comprised entirely by those which produce consumer nondurables rather than consumer durables. Industries which produce electric appliances or radio and television sets spend relatively little on advertising as a proportion of total sales. And yet, it is precisely these industries which are generally considered to produce highly complex products about which consumers might need considerable information. In the statistical analysis, we examine the joint effect of advertising, together with a number of other market structure variables, on profit rates. The additional variables are the degree of market concentration, the rate of growth of demand, and estimates of two entry restricting factors: the extent to which production economies of scale exist in the industry, and the total amount of capital required for entry. These additional variables were introduced into the analysis although our PAGENO="0497" Profit rate Advertising sales ratio 1. Soft drink 10.0 6.2 2. Malt liquors 7.2 6. 8 3. Wines 7. 3 5. 2 4. Distilled liquors 5. 0 2. 1 5. Meat 4.6 .6 6. Dairy 7.9 2.2 7. Canning 6.4 2.9 8. Grain mill products 7. 0 1.9 9. Cereals 14.8 10.3 10. Bakery products 9.3 2.9 11.Sugar 5.8 .2 12. Confectionery 10. 6 3. 5 13. Cigars 5.3 2.6 14. Cigarettes 11. 5 4. 8 15. Knit goods 3. 8 1. 3 16. Carpets 4. ~ 2. 0 17. Hats 1.6 2.2 18. Men's clothing 5. 9 1. 2 19. Women's clothing 6. 1 1. 8 20. Millinery -1. 3 . 8 21. Furs 5.7 1.0 22. Furniture 9.7 1.5 23. Screens and venetian blinds 9. 3 1. 6 24. Periodicals 11.7 .2 25. Books 10. 1 2.4 26. Drugs 14. 0 9. 9 27. Soaps 11.7 9.2 28. Paints 9.9 1.5 29. Perfumes 13. 5 15. 3 30. Tires and tubes 10. 2 1. 4 31. Footwear 7.6 1.5 32. Handtools 11. 4 4. 2 33. Household and service machinery (not electrical) 7. 3 1. 9 34. Electrical appliances 10. 3 3. 5 35. Radio, TV, and phonograph 8.8 2.2 36: Motorcycles and bicyles 5. 2 1. 1 37. Motor vehicles 15.5 .6 38. Instruments 12.0 2.0 39. Clocks and watches 1. 9 5. 6 40. Jewelry (precious metal) 5. 3 3. 2 41. Costume jewelry 1. 4 4. 0 Dr. COMANOR. A significant correlation exists between these two variables. While high profit rates may be and frequently are asso- ciated with relatively low advertising-sales ratios, such as instruments which include cameras, industries with high advertising-sales ratios COMPETITIVE PROBLEMS IN THE DRUG INDUSTRY 2047 primary attention was directed at the impact of advertising outlays because they are generally considered to influence profit rates and we wished to examine the net effect of advertising on profits after the influence of these other factors had been accounted for. The statistical analysis is founded largely on tax return data which are compiled by the statistical division of the Internal Revenue Serv- ice. It is based on industries as defined by the IRS and refers to aver- age values for the years 1954 through 1957. Averages were taken so that the results would not be influenced by the particular business conditions which happened to exist in a given year, and the years chosen refer to a complete business cycle. The statistics on profit rates and advertising-sales ratios which are used in the empirical analysis are presented in table A-i attached to my statement. (The statistics referred to follow:) TABLE A-i-AVERAGE PROFIT RATES AND ADVERTISING-SALES RATIOS IN 41 CONSUMER GOODS INDUSTRIES, 1954-57 [In percent] 81-280 O-68-'---pt. 5-32 PAGENO="0498" 2048 COMPETITIVE PROBLEMS IN THE DRUG INDUSTRY almost without exception have relatively high profit rates. And this refers specifically to the six industries that have advertising-sales ratios which exceed 6 percent. Thus, the average profit rate for those six industries was 11.9 percent which is 65 percent greater than the average return for the 35 remain- ing industries of 7.~ percent. Senator NELSON. When you say "average profit rates," are you talking about after-tax profits? Dr. COMANOR. I am talking about after-tax profits. While motor vehicles has the highest profit rate of any industry in the sample, and at the same time has a relatively low advertising-sales ratio, the next three most profitable industries-cereals, drugs, and perfumes-are all those with very high levels of advertising expenditures. Senator NELSON. Does the automobile industry fit into the picture you describe later of an industry that has some very difficult entrance barriers-that is, the vast investment it takes to get in? Is advertising a factor there? Dr. COMANOR. I think most economists would agree that the auto- mobile industry has very high entry barriers, but that these are not accounted for by advertising expenditures as much as by the important scale economies which exist in the industry, as well as by the very high absolute capital requirements. In addition, the current franchise sys- tem does tend to restrict entry. While important entry barriers exist, I think most economists would not stress the effect of advertising in this particular industry, although firms in the industry spend very large absolute sums on advertising, since after all, it is one of our largest industries. But these large absolute sums account for a relatively small percentage of sales. Senator NELSON. Would you agree that there is a distinction be- tween the automobile and the prescription drug industry on at least a couple of grounds relative to the advertising question, the automo- bile industry having a relatively low advertising ratio to sales, with the drug industry having a high advertising ratio to sales? Would you agree that one of the reasons for the distinction there is that there is a multiplicity of drugs and the person who makes the determination about what drugs will be used and what brand will be used is the doc- tor, whereas in the automobile industry, there are only four major companies in this country and the consumer, by looking at the product, makes the decision about what he will buy? Dr. COMANOR. While this factor is certainly important, I would em- phasize a somewhat different factor, which is that in the motor vehicle industry, factors such as high absolute capital requirements inhibit entry so that the auto firms do not feel compelled to allocate further sums to restrict the entry of new firms. In addition, while advertising promotes product differentiation in the automobile industry, there are various other factors besides ad- vertising which are at least as important, if not more important, in creating product differentiation, such as the annual model changes. In the pharmaceutical industry, where products which embody the same chemical compound but which are produced by different firms are now similar to one another, advertismg plays a much more im- PAGENO="0499" COMPETITIVE PROBLEMS IN THE DRUG INDUSTRY 2049 portant role. It is in this context that it is interesting to note that high advertising outlays generally are concentrated in the nondurables field, where product differences are probably much less. Advertising is very high, for example, in the detergent industry, although deter- gents are probably less different from one another than automobiles. In addition, we know that in the proprietary drug field, there are very high advertising outlays for products such as aspirin, although aspirin is aspirin. Therefore, what we find is that advertising expenditures tend to be highest where real product differences are not pronounced. And ad- vertising tends to be relatively low where real product differences in fact exist. One might conclude that if there are real product differences, ad- vertising does not play the role that jt does play when these differences are absent. It is in this way that I think we can interpret the statistics which indicate that advertising-sales ratios are higher in the con- sumer nondurable sector than in the consumer-durable sector. Senator NELSoN. I realize there are a lot of factors involved in this, but there are several differences. There is one fundamental difference between the automobile and the drug industry; that is, that the con- sumer in the one case, automobiles, is qualified to make a judgment of his own to decide whether he likes the style of the car, what horse- power he wants, how many miles per gallon it will travel, what size automobile he wants, and then based upon his experience in buying cars, he is qualified to make his own judgment. However, he has no qualifications for making any judgment whatsoever about a prescrip- tion drug. So they are fundamentally different problems, are they not? Dr. COMANOR. I agree, they are quite fundamentally different prob- lems. At the same time, we can and do find high advertising sales ratios in industries where, in fact, the consumer does make the purchasing decision-such as cereals and perfumes and soaps. So I would not stress this particular factor as the primary reason why advertising outlays are high in this industry, although I think that the fact that the doctor is the purchasing agent for the final consumer plays a very crucial role on the type of advertising which is undertaken and the effectiveness of these advertising outlays. Senator NELSON. I was not suggesting that this factor that I men- tioned was the most important, or one of the most important, but that it is a factor which differentiates the two products. Please proceed. Dr. COMANOR. These preliminary observations were corroborated by our statistical analysis where the influence of the additional mar- ket structure variables was also considered. We found that advertising has a statistically significant impact on industry profit rates and that this effect is stronger than that of any of the other variables examined. Furthermore, the magnitude of the effect is surprisingly high. In- dustries with high advertising outlays earned, on average, profit rates which exceeded those in other industries, after correcting for the other variables, by nearly 4 percentage points. This difference repre- sents, moreover, a 50-percent increase in profit rates. Since profits rep- resent the difference between prices and costs, and since advertising outlays is a cost to the firm making the outlay, these findings suggest PAGENO="0500" 2050 COMPETITIVE PROBLEMS IN THE DRUG INDUSTRY that high advertising expenditures have a double effect on price levels. Not only do they represent higher costs, but also they are associated with higher price-cost margins. ADVERTISING AS A BARRIER TO ENTRY The empirical evidence described above indicates that industries which spend substantial proportions of their revenues on advertising tend at the same time to earn high rates of return, and I want now to examine some economic factors which might explain this result. In a free market economy, high profit rates, which are not accounted for by differences in rates of growth of demand or by the degree of risk, will not persist for a prolonged period of time in the absence of factors which restrict the entry of new firms. If high rates of return in an in- dustry, in firms will enter the market, and this will generally have the effect of driving prices down to more competitive levels. High prices and profits suggest, thereby that specific factors are present which re- strict the entry of new firms. Senator NELSON. The economists who have testified in behalf of the Pharmaceutical Manufacturers Association have argued that the high profits of the drug industry were the result of high risks. Do you have any observations to make about that? Dr. COMANOR. Yes; I do. I have read parts of the testimony pre- sented by these economists. It seems to me that to consider this, one would have to ask what we mean by risk and in how we can mensure it. Risk involves some notion of unexpected changes in the value of the firm or unexpected changes in profitability. But this cannot be measured simply by the variability in profit rates across firms in an industry. After all, there are generally differences among profit rates. They may measure accustomed and ex- pected and traditional differences among firms. They may measure systematic as opposed to unexpected differences. Different firms in an industry may have different profit rates for a number of reasons: be- cause of economies of scale, because of other differences in efficiency. because they operate and sell in different segments of the market, or because of differences in consumer acceptance or product differentia- tion. It seems therefore that the variability of profit rates across firms in an industry cannot be used to measure risk because these changes may be accustomed and traditional rather than unexpected and the essence of risk is the inability to predict what will happen in the future. In addition, risk, however measured, is likely to be very different for firms of different sizes, and the simple variability across firms does not account for these differences. This is another reason why simple variability across firms is not an adequate measure of the degree of risk in an industry. It seems to me that a much better measure might be the intertem- poral difference and the variability of profit rates, across time, espe- cially after correcting for size of firm. I note from testimony given recently that when risk is measured in this manner, the degree of risk for the drag industry tends to be very low. PAGENO="0501" COMPETITIVE PROBLEMS IN THE DRUG INDUSTRY 2051 These statistics were presented to Dr. Mueller's testimony. There- fore, I think it is difficult to argue, and I certainly would not, that the drug industry is a high risk. Senator NELSON. Did you read Dr. Mueller's testimony? Dr. Coi~rANoR. Yes, I did. Senator NEr4soN. Do you agree with him on the point he makes on this question? Dr. COi~rANOR. Yes, I do. In the consumer goods industries where advertising is relatively important, we frequently observe unbranded products which sell at. prices substantially below those of highly advertised products even though there may be little real difference between them. Because prod- ucts which are little advertised must sell at far lower prices than those of their established rivals, the latter can raise prices above costs and earn high profits witihout fear of enticing the entry of new firms and of the resulting effect on price levels. High advertising outlays create effective entry barriers through a number of routes. In the first place, high current levels of advertising expenditures create additional costs for new entrants which will gen- erally exceed those for established firms. Because of buyer inertia and brand loyalty, more advertising messages per prospective customer must be supplied to induce brand switching as compared with repeat buying. And therefore, new entrants must be prepared to supply more advertising messages per prospective customer than do their estab- lisheci rivals. A further disadvantage faced by new entrants is that they must spend nearly as much in total advertising and other forms of promotion as existing firms if their products are to compete with established and well-known products. Since new entrants are generally small, this total expense must be spread over fewer units of output, and therefore the cost per unit may be quite high. Larger established firms, on the other hand, have the advantage of being able. to spread this cost further and often spend considerably less per unit sold. Thus, where advertising is important, new entrants are frequently found to bear higher unit ad- vertising costs as compared with their larger rivals.5 And this factor again enables established firms tO set relatively high prices without fear of the effective competition of new entrants. Furthermore, there is evidence that the cost per advertising message declines as the number of messages purchased increases.6 New entrants, thus, face higher unit costs which contribute further to their relative disadvantage. Finally, the need to spend considerable funds on advertising will raise the amount of capital required for entry into the market. Since new firms frequently find it difficult indeed to raise large amounts of In the automobile industry, for example, the two smaller firms during the 1950's were found to spend more than twice as much on advertising per car sold as did either Ford or General Motors; Between 1954 and 1957 Studebaker and American Motors spent annually on nntional advertising approximately $64.04 and $57.89 per automobile sold while General Motors spent $26.56 per unit and Ford spent $27.22 per unit. Chrysler assumed an inter- mediate position, spending $47.76 per unit. Leonard W. Weiss, "Economics and American Industry," p. 342. o The extent of discounts given to large advertisers is documented in Federal Trade Corn- mossion v. The Procter 4 Gamble Co., brief for the Federal Trade Commission in the Su- preme Court of the United States, December 1966, pp. 12-13. PAGENO="0502" 2052 COMPETITIVE PROBLEMS IN THE DRUG INDUSTRY capital, high advertising expenditures by existing firms, and thereby high advertising requirements for new entrants, will further restrict the rate of entry. For all of these reasons, heavy advertising expenditures serve to create substantial entry barriers. They act as an important restriction on competition and permit established firms to charge high prices and earn high profit rates without fear of the competitive consequences. And the statistical findings presented earlier merely describe the extent of the effect of advertising on competition. THE PROBLEM OF CASUALITY Throughout the analysis, we have assumed that the direction of casuality runs from advertising to high profit rates, and it is necessary to consider whether the reverse could be the case. This is especially so since a plausible argument could be made that, since advertising re- flects the discretionary behavior of firms, high profits could lead to high advertising. In other words, high profits could be "spent" on advertising. There are a number of factors, however, which suggest that the appropriate direction of casuality is in fact from advertising to profits. The first is that if profits were "spent" on advertising, which is after all an expense to the firm, higher measured profit rates would be as- sociated with lower advertising expenditures, and lower measured profit rates, with higher outlays. But we have observed a positive rather than a negative relationship between these two variables. A second and even more important reason stems from the fact that the amount of advertising expenditures depends on many factors besides the whims of individual managers. Product and market characteristics make advertising a more profit- able activity in some industries than in others, and there are few reasons for believing that managers in some industries are better equipped to take advantage of their opportunities than those in others. In fact, if we assume that managers make their decisions on adver- tising burgets so as to maximize profits, then the differences between industries which we observe reflect not the discretionary behavior of individuals but rather the particular product and market characteris- tics of member firms in the industry. The figures presented above, which deal with the average values across a 4-year period, may be interpreted in fact as describing the op- timal levels of advertising expenditures in the particular industry. And there is no reason to believe that firms with higher profit rates will have higher optimum advertising expenditures than firms with lower profit rates but similar market conditions. This latter consideration is an important one for interpreting the statistical results presented earlier. Since the analysis focuses on inter- industry differences in advertising outlays, and since these differences are more likely to reflect the product and market characteristics in the industry rather than the peculiarity of individual managers, these results have few behavioral connotations. They do not imply that an industry can earn higher profit rates simply by spending more on PAGENO="0503" COMPETITIVE PROBLEMS IN THE DRUG INDUSTRY 2053 advertising. They do not suggest that firms which currently spend little on advertising should `double their advertising budgets if their objective is to o'btain higher returns. But rather, they describe the fact that firms and industries with higher optimum advertising ex- penditures will and do earn higher rates of return than firms in less advantageous market positions. THE POSITION OF THE DRUG INDUSTRY Although this study refers to the entire consumer goods sector of the economy, we can note that the drug industry is one of those in the sample with both high profit rates and high advertising-sales ratios. During the period between 1954 and 1957, this industry stood in third place out of 41 in both respects. As defined in this analysis, however, the drug industry refers to both prescription and proprietary products. In the prescription drug industry, advertising accounts for only a portion of the firm's total selling and promotional budget. Detail men and other forms of direct selling also are important, and therefore figures on advertising alone understate the total effort in persuasion which is carried on by the drug companies. Senator NELSON. In your figures on advertising, you did not include any of the costs of the detail men? Dr. COMANOR. That is correct. Senator NELSON. What did you include? Dr. COMANOR. We included expenditures on advertising as reported in tax returns. Senator NELSON. You reported what the industry claimed as adver- tising and what the IRS accepted as advertising? Dr. COMANOR. That is correct. Senator NELSON. Do you know what that included? Dr. COMANOR. That includes expenditures on printed advertise- ments, direct mail advertisements,, and also advertisements in medical journals. It does not include salaries for detail men, for samples, or things of this sort. Senator NELSON. It did not include samples? Dr. COMANOR. It did not include samples. Senator NELSON. Even though samples had some advertising ma- terial with it? Dr. COMANOR. I think that is correct. Senator NELSON. Did you include other expenditures such as ma- terial furnished to medical students, stethoscopes, that sort of thing? Dr. COMANOR. I am not absolutely certain, but I think not. I would have to check with the IRS before I could answer these questions completely. Senator NELSON. I do not know what these expenditures amount to, but we have had testimony that medical students get a considerable amount of material, some advertising and promotional, some of them stethoseopes and so forth. Do the firms deduct this as a business expense? Dr. COMANOR. This is surely a business expense. I am not certain whether it is included in these advertising figures. My wife is a medical PAGENO="0504" 2054 COMPETITIVE PROBLEMS IN THE DRUG ThTDUSTRY student, and I am forced to admit that she has received a stethoscope and a black bag. Something thnt you might be interested to know is that one of the drug companies provides every second-year medical student with a black bag which is plastic. T11e same company in their fourth year pro- vides the medical student with a black bag of real leather. She is now going around with her plastic one. Senator NELSON. They have somewhat less confidence in the pros- pects for the second year student than the senior student? Dr. COMANOR. That must be it. Senator NELSON. Would it be easy for you to check and see whether these items include this advertising expense in the tax column? Dr. OOMANOR. One could check with the IRS on this question. Senator NELSON. If you could, along with the other material you are going to submit, it would be useful to put. it in the record. Dr. COMANOR. I will do this.7 Advertising generally is particularly effective in situations where consumer uncertainty or ignorance about the relative nierits of com- peting products is relatively high. It is not so much that physicians or other consumers consciously accept the blandishments of an adver- tising message as much as they consider tl1e purchase or prescribing of an advertised product as the "safe" course of action. When considerable uncertainty exists, the choice of a highly adver- tised product often represents minimizing the risk to the consumer that the product will do the job for which it is being purchased. And what becomes important is not the content of the advertising message but simply the fact that the product is well advertised. Thus, products are frequently described as "advertised in Life." 8 The effect of advertising on entry barriers in the pharmaceutical industry is strengthened by an important interaction which exists between research and selling expenditures. The relatively high level of research in this industry has led to a rapid rate of new product intro- duction throughout the postwar period. During the decade ending in 1960, over 3800 new products or dosage forms were introduced into U.S. pharmaceutical markets. Of this total, slightly more than 11 per- cent were new single chemical entities. Given the rapid pace of new product introduction, it is not surpris- ing that doctors have been open to the persuasion and influence of the drug companies. At a 1959 medical conference on the evaluation of new drugs, it was reported that "physicians are frightened, confused and puzzled by advertising material which pushes as many as a. thousand new drugs or combinations of drugs every year * * * several prac- titioners at the session said they felt a. void of information about t.he proper use of new drugs." ~ The simple pace of new product introduc- tion has been a major factor which has increased the effectiveness of heavy advertising and promotional expenditures. Furthermore, most new products have received patent protection. Thu~s, in 1961, of the 656 single chemical entities which were used for See letter, p. 2046, supra. ~ view is similar to that taken in Donald F. Turner, "Advertising and Competition," speech delivered before the Briefing Conference on Federal Controls of Advertising and Promotion, June 2, 1966, pp. 3-4. ° Herman Somers and Anne Somers, "Doctors, Patients, and Health Insurance," p. 100. PAGENO="0505" COMPETITIVE PROBLEMS IN THE DRUG INDUSTRY 2055 therapeutic purposes, 377 were sold only by a single firm, which amounted to about 57 percent of the tOtaL 10 The impact of the patent system, however, has not been to create tight monopoly positions, since patented products are often highly substitutable and compete with one another. But ratl1er, it has been to foreclose, to a great extent, rivalry between identical chemical entities or standardized commodities about which price competition might develop. It has strengthened and en- couraged a form of chemical product: differentiation. In the majority of cases, therefore, advertising has been able to exploit and emphasize the chemical differences which do in fact exist among products. When doctors are forced to choose between different chemical com- pounds which purportedly do the same thing, a large measure of un- certainty is certain to exist. Even though a doctor might believe that different compounds have similar therapeutic effects, he is never quite sure, and the prudent course of action is to prescribe the drug which has become well-known. And becoming well-known is, of course, the func- tion among other things of the level of advertising and promotional expenditures. Mr. GORDON. Did you say the purpose of the advertising and pro- motional activities is to prevent price c0mpetiti0~ Dr. COMANOR. After the experience of the drug industry with peni- cillin, which was introduced in the early postwar period, when the price dropped precipitiously in the course of 5 or 8 years-as I recall, the price of penicillin for a standard: dosage dropped from $20 in 1943 to on the order of five cents by 1950- Senator NELSON. $20 for- Dr. COMANOR. I think it was 100,000 units. Senator NELSON. From $20 to what? Dr. COMANOR. The price dropped from $20 for 100,000 units in 1943 to 4.5 cents in 1950. The price of streptomycin behaved similarly. At the prices which finally resulted, profits were clearly quite low. Senator NELSON. What accounted for the drop? Dr. COMANOR. Penicillin is a standardized commodity. A large num- ber of new firms entered into the market, vigorous competition took place, and prices were pushed down to costs. Senator NELSON. Was a profit made on it? You say the prices were pushed down to cost. That would mean no profit. Dr. COMANOR. It is unclear as to just what the profit margin was for penicillin. It may very well have been that profits were low and even nil, No figures are available on this question. This experience, I think, had a :~ sharp impact on the industry. It became clear that profits in the future would depend on introducing specialized products, on restricting the growth of standardized com- modities, and thereby restricting effective price competition. Price competition seems to demand a number of firms who sell `the same chemical compound, for it is only in this setting that the purchaser may focus on price rather than the therapeutic effects of a product. It became clear that some means of differentiating the product was re- 10 p~ Walker, "Price Levels and Market Power in the Ethical Drug Industry," paper presented at the December 1967 meetings of the Econometric Society, Washington, D.C., table 3. PAGENO="0506" 2056 COMPETITIVE PROBLEMS IN THE DRUG INDUSTRY quired. We note that throughout the decade of the 1950's, large num- bers of new products were introduced. Most of these were protected by patents. And also, advertising expenditures grew rapidly. We can un- derstand the incentive for both research and advertising in the drug industry in terms of the aim to create and promote product differenti- ations which would prevent the development of standardized commodi- ties and which would prevent the type of price experience which occurred in penicillin and streptomycin. Senator NELSON. Thank you. Mr. GORDON. The product differentiation you just talked about is the chief barrier to entry into the various therapeutic categories, is that correct? Dr. COMANOR. That is correct. It is the role of both advertising and research in this industry to promote a form of product differentiation which then serves to restrict entry and to restrict the development of price competition. In this manner, the patent system interacts with the high level of advertising and promotion by insuring that the competition which ex- ists in this industry will focus on chemically differentiated products, and this factor increases substantially the effectiveness of the selling efforts in this industry. Were the patent system not to exist, the effectiveness of these efforts is likely to be reduced. If new entrants were free to produce the same chemical compound as established firms, the effectiveness of advertising in constructing barriers to entry would be lessened, and we might expect to find more stringent constraints placed upon the price be- liavior of the leading firms in the industry." Even in the absence of chemical differences among products, how- ever, considerable uncertainty remains as to whether different brands of the same compound are equally good. It is claimed that tradename products which are sold by the larger companies are manufactured under quality controls which produce higher quality products. While there is no readily available evidence on quality control expenditures for pharmaceutical firms in the United States, statistics are available for Canada. Twenty-one large firms, many of which are U.S. subsidi- aries, reported to the Canadian Government that quality control expenditures accounted for only 1.2 percent of total sales or approxi- mately 3.6 percent of the cost of goods sold. Senator NELSON. What is the distinction? 1.2 percent- Dr. COMANOR. Of total sales. Senator NELSON. I understand the 1.2 percent of total sales and approximately 3.6 percent of the cost of goods sold-what precisely do you mean? Dr. COMANOR. Production costs. You can translate cost of goods sold as production costs, of the total production costs of the product, qual- ity control accounts for only 3.6 percent. After these costs are adver- tising, promotion, profits, to make up total revenues. Senator NELSON. So when you are talking about cost of goods sold, you mean the total cost of producing the product? "Restrictive Trade Practices Commission, "Report Concerning the Manufacture, Dis- tribution, and Sale of Drugs," Department of Justice, Ottawa, app. Q, p. 145. PAGENO="0507" COMPETITIVE PROBLEMS IN THE DRUG INDUSTRY 2057 Dr. COMANOR. That is correct. SenatorNELsoN. And then the sales price is different in that you add your profits and your advertising and your promotion, is that what you are saying? Dr. COMANOR. That is what I am saying. These statistics imply, then, that production costs are approxi- mately one-third of the total selling price. Senator N1~soN. Where do you ~et that figure? Dr. COMANOR. We are dealing with the same expenditures on quality control, and these outlays are taken as a fraction of two different quantities. One is total production cOsts and the other is total revenues. One percentage is a third that of the other. Senator NELSON. You have several factors in here that I do not understand. You are talking in one case about quality controls. Quality control is not the same as production costs? Dr. COMANOR. Oh, no; in fact, quality control is only 3.6 percent of total production costs. Senator N1~soN. And 1.2 percentof sales cost? Dr. COMANOR. 1.2 percent of total sales revenues. Senator NELSON. OK. Dr. COMANOR. These statistics suggest that relatively small firms should be able to invest in quality control facilities and as can be seen, these expenditures are considerably less cos~ly to the smaller firm or new entrant than are heavy advertising and promotional efforts. Senator NELsoN. As to these statistics that you got from the U.S. subsidiaries of manufacturers in Canada, does the Canadian Govern- ment require a breakdown of these various costs? Dr. ~OMANOR. I do not think this breakdown is reported every year. A report published by the Restrictive Trade Practices Commission in Canada included these statistics. Senator NELSON. When was that report published? Dr. COMANOR. I am afraid I dO not have the date here. Within the last few years. I have a copy of it. Senator NELSON. Is it a large report? Dr. COMANOR. It is quite a large report. It must be 500 or 600 pages. Senator NELSON. In this report, do they have the statistics on the costs of production of a drug and the cost of quality control and the sales and profits-all of these statistics? Dr. COMANOR. Not per drug, but for the entire company. Senator NELSON. All together? Dr. COMANOR. Or it may not even have been exported for individual companies but rather for the total of these 21 firms, only some of which were U.S. subsidiaries. Senator NELsoN. Did they not have any breakdown company by company? They must have gotten the figures from some place. Dr. COMANOR. They may not have published them. Senator NELSON. In order to get the composite, you have to have the individual figures. Dr. COMANOR. They may not have published them. I am not certain about this point. I could find the answer for you. PAGENO="0508" 2058 COMPETITIVE PROBLEMS IN THE DRUG INDUSTRY Senator NELSON. If you would give use the name of the report, I will have the staff look into it. Dr. COMANOR. Footnote No. 11 provides t.he reference to the report. Senator NELSON. All right. * Dr. COMANOR. While more work needs to be done in gathering and analyzing statistical evidence on this problem, there is no evidence which indicates that only branded drugs produced by large firms are of a high quality. THE CRITICAL ROLE OF INFORMATION The analysis above suggests that advertising and other forms of promotion. have an important effect on competition in `the pharmaceu- tical industry. These outlays permit the leading firms in the industry to enjoy `a considerable freedom from competitive contraints, and they have been translated into very high rates of return. Furthermore, the analysis suggests that advertising and other forms of promotion have been effective to a large extent, because of the high degree of uncertainty about the relative merits of competing products which is present in the minds of t.he physician who serves as the pur- chasing agent for the final consumer. On these grounds, there is much to be said for providing the physi- cian with an impartial source of information which will relieve him of his current reliance on information provided by the drug companies. Senator N~soN. I think this is a very good point. The fact is that a good part of the thrust of the industry's position is that you can't rely upon a drug unless it is produced by an established, well-known brand name company and that you are taking a risk with drugs from other companies. They end up convincing the physician of that.. He has be- come convinced that there is not any such thing as generic equivalency or that it is a very serious problem. Further, there is not any available compendium of drugs that t.he doctor is prepared to trust, and that can assure him that he can prescribe a generic drug and trust its relia- bility. You are suggesting, I take it, that there ought to be a reliable, independent source so that the doctor could select drugs that are as- serted by an independent source to be equivalent so that he can pre- scribe a drug that gives his patient a better price break. Is that what you are saying? Dr. COMANOR. That is correct. It seems to me that advertising is ef- fective primarily because of the uncertainty which exists in the mind of the purchasing agent, which in this case is the physician. And if these expenditures were not so effective, we might expect a lower amount to be spent. What is important is to reduce the effectiveness of heavy advertising outlays by providing the doctor with an independent and reliable source of information. This is not to say that the drug companies do not provide accurate information, although it may sometimes be less than complete, but rather that information from an independent source would have an element of impartiality which salesmen or printed advertisements rarely achieve. Senator NELSON. If it is less than complete, it is not accurate, is it? PAGENO="0509" COMPETITIVE PROBLEMS IN THE DRUG INDUSTRY 2059 Dr. COMANOR. I agree that the drug companies do not always pro- vide full information about the relative merits of competing drugs. Senator NELSON. If it omits certain information to the doctor, then it is not accurate, is it? Dr. COMANOR. That is correct. Senator NELSON. But what you are saying is that they do not neces- sarily say anything incorrect, but if they omit something, the informa- tion is not accurate? Dr. COMANOR. That is right. There are certain types of information which it is very important for the doctor to receive and which he rarely receives under the current system. He rarely receives a statement that two drugs are the same, because the incentive for the drug com- pany is the opposite-to differentiate their product from those of their rivals. One piece of information which it is important for him to re- ceive is that two drugs have the same or similar therapeutic effects, if that in fact `happens to be the case. Senator NELSON. In the majority of cases, no company is going to advertise that several other companies have a drug that is equivalent to its own. Dr. COMANOR. That is right. Senator NELSON. So it has to come from somebody else. Dr. COMANOR. There are certain types of information which the doc- tor simply does not receive from advertisements. Therefore, he is left in the dark. And also, and `this seems to me to be equally important, it would be open to the smaller firm and new entrant as well as the large company. Where various firms produce the same chemical compound, the phy- sician should be informed either that all products are equally good, in terms of purity, strength, or some other characteristic, or that one brand is to be preferred to another. And where different compounds compete, impartial information would lead to improved medical practice as well as reduce the signifi- cance of advertising and promotion. If a small firm or new entrant introduced a beneficial new drug, it would compete on the basis of: therapeutic properties rather than advertising claims. There is no evidence, moreover, that smaller firms are at a disadvantage in research as compared with advertising and promotion.12 On many grounds, then, there seems to be a considerable need to provide the Nation's doctors with greater information about new drugs which is independent of the messages they receive from the drug companies. Mr. GORDON. Dr. Comanor, in the footnote on page 16, you state "The available evidence suggests, moreover, that small firms have relatively more efficient research facilities than their larger rivals." What evidence do you have to show this? Dr. COMANOR. In one of the papers which I have published, I report on a statistical analysis of the relationship between research input and research output. Senator NELSON. Your analysis was comparing the products result- ing from the research? 12 The available evidence suggests, moreover, that small firms have relatively more efficient research facilities than their larger rivals. William S. Comanor, "Research and Technical Change in the Pharmaceutical Industry," Review of Economic8 and Stati8tiCs, May 1965. PAGENO="0510" 2060 COMPETITIVE PROBLEMS IN THE DRUG INDUSTRY Dr. COMANOR. That is right. I carried out a study of a large number of pharmaceutical firms and I measured the relationship between research input and output. I measured the output of a research estab- lishment by the number of new products introduced by the firm since research in the pharmaceutical industry is directed largely toward new products rather than new pr~esses. The number of new products was multiplied by its sales during the first 2 years following intro- duction. To make these computations, I obtained access to the Gosselin data which provides sales by product for all products sold in the coun- try. Research input was measured by the number of research personnel employed by the individual pharmaceutical firm, and the relationship between input and output was then estimated. Senator NELSON. You did not make, I take it, any distinction between the efficiency of the research of one company in developing a new drug versus another company in producing a new drug? You did not make a distinction between that and the research which produced a product differentiation, did you? Dr. COMANOR. No, I did not. Senator NELSON. So whatever they produced, either as a product differentiation or a new drug, is included in the statistics? Dr. COMANOR. I used two measures of new products, but the most important included only new single chemical entities. I was, therefore, dealing with products which were new in the sense that they were different chemical compounds which had previously been introduced in the U.S. market. While it is true, of course, that some of these embodied a greater therapeutic advance than others, and it may be the case that some embodied no or very little therapeutic advance, I did not feel capable of deciding this question. I simply took all new single chemical entities, multiplied by their sales in the first 2 years following intro- duction, to measure the output of research. When I carried out the statistical analysis between research input and research output, which is reported in one of the papers that I have published on the industry, there were a number of findings: First, there was a fairly sustained and direct association between input and output. In other words, research is not undertaken in this industry simply in the hope of some payoff in the future, but rather there is this fairly sustained and systematic relationship between the amount of research undertaken and the number of new products intro- duced by the individual firm. Senator NELSON. When you say you are comparing the relatively small firms versus their larger rivals, what standard did you use to distinguish a relatively small firm and a large firm? Dr. COMANOR. I measured the size of firm by the level of pharma- ceutical sales. Senator NELSON. Yes; what was the cutoff pomt? Dr. COMANOR. There was no cutoff. It was undertaken not on~ a discrete but on a continuous basis. I exammed the entire relationship between small firms and large firms without any arbitrary cutoff point to distinguish large from small. Senator NELSON. So some of of them were relatively small? Dr. COMANOR. Yes, and I had firms of all sizes in my sample. PAGENO="0511" COMPETITIVE PROBLEMS IN THE DRUG INDUSTRY 2061 I found that economies of scale exist in very small firms, but at the same time, diseconomies of scale exist in large firms and even in research carried out in moderate-sized firms. From these statistical results there appear to be no advantages from increasing the size of the firm's research effort beyond 50 to 75 professionals. At the same time, we can note that most of the leading firms in this industry have research labs which are far larger than this size. The cutoff point, which was not determined arbitrarily but resulted from the statistical analysis, was that economies of scale appear to exist up to the level of 50 to 75 professional personnel, but become diseconomies after that level. The position stated above is identical to that taken earlier by the Assistant Attorney General in charge of the Antitrust Division, Mr. Donald F. Turner, who in a recent speech discussed the competitive effects of advertising expenditures generally throughout the economy. Mr. Turner emphasized: It is the extent of uncertainty about the relative merits of competing products which contributes to the large effect of advertising, and this suggests that Govern- ment policies be directed toward neutral vehicles of information which tend to deal directly with the uncertainty. He goes on to mention the Medical Letter, which is an important source of "technical information about the therapeutic value of new drugs", and con- cludes that "there is much to be said for providing Government funds to the organization which publishes the Medical Letter so that its publication may be supplied free to all doctors. In addition, the latter could be expanded to ensure that doctors receive their first information about a new drug from this source rather than from the lips of a detail man." ~ Mr. GORDON. Is it your opinion that this measure could offset the huge expenditures by the drug industry on advertising and promotion? Dr. COMANOR. I think that this type of policy, while it may not offset fully the impact of advertising and promotion on product dif- ferentiation and thereby on price competition, would have an im- portant effect. I would expect that it would reduce significantly the effectiveness of these outlays and think we could expect a higher degree of price competition with this information provided to all doctors than we find currently. I agree fully with Mr. Turner's position that an impartial and in- dependent source of information is, needed-that it would be an im- portant step forward, both in improving the quality and standard of medical practice throughout the country, and in reducing the anti- competitive consequences of heavy advertising. There are a number of alternative routes to achieving this objective. One might be to provide Federal support to the nonprofit organiza- tion which publishes the medical letter. Not only would this permit a more comprehensive study of drugs and drug claims than is possible on the basis of available facilities, hut also it would allow an expanded Medical Letter to be sent free to all phyisicians throughout the country. Mr. GROSSMAN. Dr. Comanor, may I ask a question? Several of the economists who have appeared here have talked about the fact that we should get rid of the brand name, trade name, and just use the generic name. Do you feel that way as well? Dr. COMANOR. There are a number of routes to achieving the same objective. One route, m some sense an easier route, would be to provide ~` Turner, p. 2054, supra. PAGENO="0512" 2062 COMPETITIVE PROBLEMS IN THE DRUG INDUSTRY information to the doctor which would reduce the impact of trade names and heavy advertising outlays. Another route, which is perhaps a stronger step, would be to elinu- nate all trade names altogether. I certainly feel that further informa- tion is needed now regardless of whether or not trade names are eliminated. At the same time, it seems to me that trade names in this industry have become a very important source of limiting the entry of new firms into the relevant therapeutic markets and restricting the development of price competition. Therefore, although I have not emphasized this problem in my statement, I would certainly sub- scribe to this position. Mr. GROSSMAN. If we did away with trade names, would it not be likely that we would have a generic name (Schering), generic name (Upjohn), and the company would promote their name? They would find a way to get around it? Dr. COMANOR. There is no question that the drug companies would aim to achieve just the end which you suggest. One might argue that their ability to differentiate their products on this basis would be less than it is currently, and as a result, we might find a higher degree of price competition. This is not an all-or-nothing situation; it is clearly a matter of degree. While some measure of product differentia- tion would remain, and might even be desirable at a small level, it would not lead to the high prices which currently exist and the very high rates of return. Mr. GROSSMAN. Do you thii± doctors should be given some informa- tion as to variable pricing such as the formulary? Dr. COMANOR. I would like to see the price of the product included in the Medical Letter or any sources of information provided to doctors. Mr. GROSSMAN. Once that is done, the doctor should be able to make the choice? So long as he is given this information, he should be able to make a free choice? Dr. COMANOR. That is a very difficult question. Although price in. formation should be provided to physicians, I do not think I would go so far as to say that in every case, only the doctors should decide among generically equivalent drugs-that the patient should not be permitted to achieve generic equivalency if this were desired. I am not prepared to say that there should be no option to the patient on generic equivalency. Senator NELSON. The patient does not have the option on generic equivalency. Dr. COMANOR. No. Senator NELSON. The doctor does. Dr. COMANOR. The question I think we are really getting at is whether or not the pharmacist should be permitted to substitute among generically equivalent drugs. I find that it is a very difficult problem. I am not prepared to say definitely that providing since in- formation to the physician is sufficient. Mr. GROSSMAN. If we list various drugs under a generic heading, be they trade names or generic names, and we also list the prices next to them, will we have gone far enough? PAGENO="0513" COMPETITIVE PROBLEMS IN THE DRUG INDUSTRY 2063 In other words, we will be giving the doctor quite a bit of informa- tion, more than he has now. Is this going far enough? Dr. COMANOR. I think that is an important step forward. Whether or not that would be sufficient is a question which is very difficult and I am not prepared to say definitely that it is. Mr. GROSSMAN. But it would be a step forward? Dr. COMANOR. I think so. An alternative proposal would be to create a new institute in the National Institutes of Health which would be responsible for evaluat- ing existing information on the therapeutic properties for new drugs, carrying out their own studies where necessary, and also for providing the accumulated information freely to all physicians. In this manner also, an independent source of information about both current and new drugs would, become available. By whatever route this is achieved, however, it seems clear that the relatively small level of expenditures required for these purposes would pay for itself many times in terms of limiting the effects of heavy advertising out- lays and in terms of lower prices for prescription drugs. While a new source of information will not solve all of the com- petitive problems which exist in the pharmaceutical industry, and restrictions on competition will remain because of the current opera- tion of the patent system; because of the practice of trade name pre- scribing; and because of the problem of drug substitution or the lack thereof at the level of the pharmacist, still Federal action in this direction should make an important contribution toward improved medical practice as well as toward lower prices for prescription drugs. Mr. GROSSMAN. Independent economists have previously recom- mended compulsory licensing as one means of making the industry more competitive. Would compulsory licensing lead to lower prices? Dr. COMANOR. The patent system currently operates to provide a form of chemical product differentiation in the sense that competi- tion exists on the basis of different chemical entities. While it is true that different products compete with one another, the patent system increases the extent to which competition is based on different chemical entities. And this always provides an important element of uncer- tainty to the physician. He is never quite sure whether one product is better than another. For this reason, it increases the effectiveness of advertising. It per- mits advertising to have a major impact on prescribing habits, on the entry of new firms, and on price competition. With compulsory licensing, we might expect to find new firms enter the industry, competition take place on the basis of the same chemical compound, and here, advertising would probably be less effective than it is currently, especially if further information were provided to physicians. And we might expect to find a higher degree of price competition, and as a result, lower prices. Mr. GROSSMAN. Now, would that lead to less research by the drug companies? Dr. COMANOR. With compulsory licensing, we would expect to find different firms competing on the basis of the same che~nical com- 81-280-OS-Pt. 5-33 PAGENO="0514" 2064 COMPETITIVE PROBLEMS IN THE DRUG INDUSTRY pound. This would tend to reduce, as I have noted, the effectiveness of heavy advertising. And we might expect to find more vigorous compe- tition and lower prices. An example of the effect of patent protection is the major difference in price behavior between penicillin, where price competition has been vigorous, and the broad spectrum antibiotics, which for a long time were patent protected, when there was little price competition. At the same time, we have to expect that with lower rewards from research, it is likely that less research will be undertaken, and it is important to recognize that there is in fact a tradeoff between indus- trial research and lower prices. However, it seems to me that we pay a high price for industry re- search in terms of the high prices and profits which exist in this indus- try. A major question for public policy concerns what we get for this price. To answer this question, we have to think about the character or payoff which society realizes from industrial research. We have to consider the value of industrial research. Industrial research is carried on within the context of strong product rivalry. This rivalry influences significantly the type of work which is in fact carried out. We can note that there is a large effort to invent around existing patents, to find a new product which has similar therapeutic properties to those already on the market. And this, of course, is what we mean by molecule manipulations, which are pur- ported to have the same therapeutic effect as existing products. Also, if we examine the research efforts and output of the industry, there appears to be a considerable effort to create new drug combina- tions, combining different active ingredients with one another, and to introduce new dosage forms. As a result, we do not expect the major discoveries to be made within the industry precisely because of the character of the research effort which in fact is undertaken. Still, it seems to me that the minor advances which result from the research laboratories within the phar- maceutical industry, the modifications and variations which result, are important elements in the process of developing a new technology. The second major emphasis of industry research is the process by which new knowledge is translated into marketable l)roducts. We know that many problems remain even after a basic discovery has been made. It has been estimated, in fact, that 50 percent of research and develop- ment expenditures within the pharmaceutical industry goes to support the work of synthesizing, purifying, modifying and preparing sub- stances for testing. And we know that considerable testing is required before a drug may be released to the public. In addition, there is also frequently a considerable gap between laboratory synthesis and large-scale manufacturing which is also car- ried out within the research laboratories of the pharmaceutical industry. Although this might be considered development work, it is still quite necessary. And we have only to note that although penicillin was dis- covered originally in 1928, it was not used effectively until some years later, until the Second World War, 1943. This time lag is likely to b~ much shorter because of the current. facilities of the drug industry. PAGENO="0515" COMPETITIVE PROBLEMS IN THE DRUG INDUSTRY 2065 To evaluate industry research and development, however, we have to place it within the context of the tOtal medical research and devel- opment effort in the United States. In 1960, the industry provided about 30 percent of total research and development outlays. At the same time, it provided only 12 percent of the Ph. D.'s involved in medical research and employed only about 5 percent of the M.D.'s in medical research. This indicates, thus, that the character of research in the industry is likely to be far different, at a lower level of scientific~ caliber, than the medical research which is carried on outside of the pharmaceutical industry. It seems to me, therefore, that the industry research effort cannot be evaluated or judged by the same standards as nonindustry research. It seems to me that the industry research effort should be considered as a, complimentary rather than as a competitive or alternative type of activity. There appear to be extensive resources which are devoted to) fairly routine sets of activities. When evaluated on its own terms, on a different basis than might be appropriate for evaluating academic or Government medical research, it seems to me that the industry effort, by and large, does a good job. It has, in fact, been responsible for the large number of modifications and improvements of existing drugs.: It has also been responsible for accelerating the process of testing aild developing new drugs. And I would argue that both of these provide important benefits to the development of medical technology and medical practice in the country. At the same time, the question rern~dns as to whether the "price" that society pays in terms of the opportunity costs for these results is too high. If we examine the proposal of compulsory licensing after a short period of time, such as the Kefauver suggestion of 3 years, it is clear that this proposal would reduce the incentive to undertake research. The industry certainly would have less incentive to undertake purely duplicative research in order to enter a profitable therapeutic market. Senator NELSON. If they had less incentive to do purely duplica- tive research, that would be a net benefit to society; would it not? Dr. COMANOR. It certainly would not be any loss to society. Senator NELSON. And they would tend to do more pure research to develop new products and that would be beneficial; would it not? Dr. COMANOR. I think that is right. At the same time- Senator NELSON. What I am really getting at is that we have had testimony that much of the research is directed solely at product dif- ferentiation, with the production of a product that does the same thing another drug does, and in some ins~.ances not even as good a job as another drug. I understand that the argument the industry will make is that as a result of this kind of research and molecular manipu- lation, we sometimes get an improved product. You recognize that, too. But if it did eliminate just purely duplicative research, the result of which is to produce just another product doing the same thing as ~ drug already on the market, then it would seem to me that it would be beneficial; would it not? Dr. COMANOR. Yes; that is correct.: PAGENO="0516" 2066 COMPETITIVE PROBLEMS IN THE DRUG INDUSTRY With compulsory licensing, it seems that the incentive to under- take purely duplicative research would be clearly much less. At the same time, the incentive to undertake expensive basic research projects would also be much less. Basic research work is likely to be much more costly than applied work, as well as more risky, and if the returns from this work would be lessened through compulsory licens- ing, the industry would probably carry out much less pure funda- mental research than it does now. Senator ~%ELSON. I do not follow the argument, because if they produced a new product that. is useful, then every doctor who has a use for that type of drug is going to prescribe it. They then could make their profit. over some period of time and when they are forced to license other firms they would receive a royalty from what is sold all over the world. Why is that not an incentive? Dr. COMAXOR. That is an incentive, but the important question is not simply whether an incentive exists but how much. And clearly with compulsory licensing, the prospective gains would be less than they are currently, because currently, they can charge high prices throughout. With lower prospective gains, we might expect to find less research undertaken of a very costly nature, and basic research is probably the most costly type of work which is undertaken. At the same time, it seems to me that with regard to basic research, it is difficult to argue that private firms have a comparative advantage relative to Government, foundations, and universities. Most basic re- search in the medical area is performed outside of the indust.ry now, imd even if the industry carried on no basic research, it is not clear that this charge would involve heavy social costs. Basic research, as many commentators have noted, depends on an atmosphere of freedom of inquiry, which may not be best achieved within the confines of a profitmaking organization. It seems to me, then, that the comparative advantages of basic re- search probably lies in universities and Government laboratories. While the magnitude of industry basic research would decline, this is precisely an area where the industry does not seem to have a com- parative advantage. At the same time, there would still be an incentive for the firm to differentiate its products from those of its rivals and to create entry barriers. Furthermore, it would continue to earn high returns during the first 3 years, and receive patent royalties thereafter. The industry would be likely to emphasize the area where imme- diate payoff is greatest. They would emphasize modifications which represent relatively small advancements in medical practice. They would also, it seems to me, be expected to emphasize the devel- opment. and testing of new drugs which were originally discovered outside of industry laboratories. Before a new drug is introduced, not only must the basic discovery be made but also considerable scientific and technical resources are required to test it under many different conditions. This is precisely the work in which the industry appears to have a comparative advantage. This is the sort of work in which the payoff specifically from industry research is likely to be greatest, and we might expect to see it continued even with compulsory licensing. PAGENO="0517" COMPETITIVE PROBLEMS IN THE DRUG INDUSTRY 2067 As you note, high gains would be obtained during the original 3-year period, and in addition, of course, gains from patent royalties would remain. There are important areas where drug companies probably do have a comparative advantage in terms of the type or character of research, but also it seems likely that this type of research effort would continue even with compulsory licensing. Therefore, I would certainly sub- scribe to the view that compulsory licensing, after a period of time at reasonable royalty rates, is a desirable policy change and should be enacted. While it is true that the magnitude of research and development in the pharmaceutical industry would probably decline, it seems likely that projects of limited medical value and lowest industry compara- tive advantage would be eliminated first. Senator NELSON. You are talking about pure research as contrasted, say, with applied research. The industry does not do very much pure research now, does it? Dr. COMANOR. There is, of course, considerable controversy over precisely how much pure research drug firms currently undertake. Senator NELSON. It is product-directed; is it not? Dr. COMANOR. It certainly is. Senator NELSON. Is it not true, by most definitions, that there is not very much pure research done by industry as a whole in this country as compared to applied research? Dr. COMANOR. I think the correct statistic, throughout the entire economy is something on the order of 10 percent, but I would have to check to be completely accurate. But however much basic research is done now, I would expect that less would be done with compulsory licensing. At the same time, I am not unhappy with that result, because as I have noted, this is not an area where the industry appears to enjoy a comparative advantage relative to nonindustry research. Therefore, I would suspect that with compulsory licensing, the mag- nitude of pure research would decline., but research with the highest industry comparative advantage would continue. I think it iiiappro- priate to argue that the more research the. better in this type of profit- oriented context, and it may well be that the present research effort carried out in the industry is sufficient or even exceeds that which is needed for its major research functions and responsibilities; Senator NELSON. What percentage of sales does the research effort represent now in the industry? Dr. COMANOR. Different sources provide different estimates. From the statistics I have seen, it is on the order of 6 to 8 percent. Senator NELSON. Of sales? Dr. COMANOR. Of sales. Therefore, the lower research which would result from compulsory licensing seems to me to be a small price to pay for a more competitive determination of prices and profits in this industry. Mr. GORDON. Dr. Comanor, the relevant market of the drug indus- try, as I understand it, is not the industry as a whole but each mdi- vidual, therapeutic category. Would you say that is correct? PAGENO="0518" 2068 COMPETITIVE PROBLEMS IN THE DRUG INDUSTRY Dr. COMANOR. When economists define the limits of a market, their aim is to determine the area, whether in terms of geographic limits or product class limits, in which prices are set, and this generally depends on the area in which different products may be considered as highly substitutable. And if we examine the question of substitutability in the pharmaceutical industry, it is clear that a diuretic, which is used for one type of ailment, cannot be substituted for an antibiotic, which has a very different use. Given the very low degree of substitutability among therapeutic classes, I think most economists would agree that the relevant markets here, which define the limits within which prices are determined, is the therapeutic class rather than the industry as a whole. Mr. GORDON. What do you have to say about the concentration of output in each of these therapeutic categories? Dr. COMANOR. If one examines therapeutic markets, the concentra- tion ratios are very high. This is precisely because effective entry bar- riers have been created. If advertising and research had not been able to create effective product differentiation, we would expect high prices to attract new firms into the market. That this has not occurred is due to the factors which I have discussed. Mr. Go~oN. You have stated that price competition is virtually absent from this industry. What do you have to say about the drive to reduce costs and pro- duce drugs more efficiently? Dr. COMANOR. Since production costs account for such a small pro- portion of total costs in this industry, on the order of a quarter to a third, there is clearly relatively little to be gained by the firm from putting a considerable effort i~ producing drugs more efficiently. More- over, lower production costs would have little competitive effect. With little to be gained, it is not surprising to find a relatively small effort devoted to reduce production costs. While I am afraid that no hard statistical evidence is available on this question, it is likely that the drive to reduce costs is much less than would be the case if price com- petition were more effective and more vigorous. Mr. GolmoN. So actually, we have a situation, do we not, where you have not only product differentiation but also a highly concentrated market? What would be the effect of both of these on the possibility of price competition? Dr. COMANOR. We observe that there is relatively little price com- petition among highly differentiated products, although there does seem to be some degree of price competition among those products which are not effectively differentiated. For example., we still see effec- tive price competition for penicillin and streptomycin. But, as we can tell from the high overall profit rates, these must ac- count for a relatively small share of the total number of products pro- clucecl. During the period of the Kefauver committee investigations, it was reported that-I think this was for 1957-the profit ra.te in the pharmaceutical industry was double or more than double the average of all manufacturing. Mr. Goimox. Even if von did not have product differentiation, the high concentration itself would tend to bar the possibility of competi- tion in this industry? PAGENO="0519" COMPETITIVE PROBLEMS IN THE DRUG INDUSTRY 2069 Dr. COMANOR. That is true., but high concentration and entry bar- riers are related to one another. The primary reason for high concen- tration in relevant therapeutic markets is precisely because of the entry barriers created by product differentiation. So it is difficult to distinguish the competitive effects of one from the other. Mr. GORDON. Now, given the type of research conducted by the in- dustry, and I want to highlight this point which you have already brought up, would it be fair to say that the industry is heavily de- pendent on new scientific knowledge developed outside the industry? Dr. COMANOR. I think it. is clear that much of a research effort in the industry makes use of new scientific advancements which come about in nonindustry laboratories. At the same time creating new drugs out of new scientific knowledge is precisely what might be con- sidered the major research responsibility for the industry. So I certainly would not be critical of this fact.. It seems to me this is precisely what society should expect and de- mand from the industry, that it will utilize the research efforts from outside, where the majority of medical research is in fact undertaken, and translate important discoveries into marketable products. So while much of the industry research effort is devoted to this type of work, I feel that this is precisely what is desirable or useful from the point of view of society. At the same time, this type of work would probably be carried on even under a regime of compulsory licensing after 3 years, and there- fore, I think we have much to be gained from adopting this policy. Senator NELSON. Thank you very~ much, Dr. Comanor. Your testi- mony was a very useful and valuable contribution to the record. We appreciate your taking the time to come here and testify today. Dr. COMANOR. Thank you very much. (The supplemental information submitted by Dr. Comanor follows:) [From "Economica," November 19641 RESEARCH AND COMPETITIVE PRODUCT DIFFERENTIATION IN THE PHARMACEUTICAL INDUSTRY IN THE UNITED STATES (By William S. Comanor) During the past two decades expenditures in the United States on research and development have expanded greatly. From outlays of the order of a half- billion dollars at the start of the Second World War, they have grown to well over $10 billion. The rapid growth of an "industry of discovery"1 has; given rise to a large number of questions concerned with the rOle and function of industrial research. Why do firms spend large amounts on research and development? What is the relationship between research and market structure? What has been the impact of public policy, in the form of the patent system, on these expenditures? In dealing with questions of this sort it is necessary to stress the pattern of relationships among firms within which research and development is undertaken. In an activity characterized by high degrees of uncertainty, we should not expect 1 This article is based on the author's unpublished Ph.D. dissertation, The Economics of Research and Development in the Pharmaceutical Industry, Department of Economics, Harvard University, June 1963. I would like to acknowledge the assistance of J. W. Mark- ham and R. B. Heflebower, who acted as supervisors of the original study, and also of E. T. Penrose for a number of helpful suggestions in the writing of this article. The study was undertaken with the aid of a research grant from the Science and Public Policy Pro- gramme at Harvard, and I am indebted to the Programme for this support. 2 John T. Dunlop, "Introduction: Problems and Potentials", in John T. Dunlop, ed., Automation and Technological Change, p. 2. PAGENO="0520" 2070 COMPETITIVE PROBLEMS IN THE DRUG INDUSTRY firms to invest so heavily in the absence of specific pressures. These pressures arise most frequently from oligopolistic market structures in which the need for product differentiation has been recognized and where rivalry has been limited to variables other than price. This article puts forward an analytical framework within which these questions may be considered, and also attempts to provide some answers dealing with the experience of the pharmaceutical industry in the Tjnited States. THE GROWTH OF RESEARCH AND DEVELOPMENT The spectacular growth of the pharmaceutical industry in the postwar period has been founded upon a number of important medical advances and has been associated with some radical changes in the structural and behavioral character- istics of the industry.2 Prior to 1937 and the introduction of the early suipha drugs the industry was composed largely of long established firms producing relatively standardized commodities. Unlike as at present, a large proportion of medical prescriptions were compounded by the pharmacist.4 Barriers to entry were low and a high degree of competition prevailed. With the introduction of penicillin and streptomycin during and immediately after the Second World War, the nature of the industry changed. As neither of the products was protected by patents,5 the rapid growth of demand resulting from their introduction was accompanied by the entry of many new suppliers and by the development of active price competition. The price of a standard form of penicillin dropped from $20 for 100.000 units in 1943 to 4'h cents in 1050 0; the price of streptomycin behaved similarly.7 Excess capacity was created, and it may well be that prices fell to the vicinity of short-run marginal costs. Following this experience, it was clear to the leading firms that their profits in the future would depend on the development of more protected market posi- tions to be achieved through some form of product differentiation. This would provide the producer with substantial control over the prices of his products as well as act as a significant barrier to entry into the relevant therapeutic market. With the trend in the new technology towards the compounding of prescriptions by the manufacturer rather than the pharmacist, this development seemed all the more promising. Although increased selling expenditures would be necessary, these were not likely to be sufficient to establish effective differentiation, for time industry faced a relatively informed consuming public in the medical profession, and it would be more difficult to differentiate between products consisting of identical chemical substances. Only through the extensive introduction of new products could significant differentiation be achieved. This would enable indi- vidual firms to emphasize in their selling activities the improved quality of new drugs. Moreover, if each of the leading firms followed this course of action, coin- petition in the industry would largely cease to be founded on the sale of stand- ardized commodities.8 The accelerated growth of pharmaceutical research and development expendi- tures during the post-war years w-as due largely to the desire to promote the In this study the pharmaceutical industry is defined to include those firms which pro- duce ethical drugs, as opposed to proprietaries, and which distribute these products in dosage-form. Pharmaceuticals are, thereby, marketed and sold only through the medical profession and require, for the most part, a written medical prescription. In 1961 prescriptions compounded by pharmacists accounted for only 3.6 per cent of all prescriptions in the United States. While a patent was issued to the inventor of streptomycin, Dr. Selman A. Waksman, it was assigned to the Rutgers Research and Endowment Foundation, and was licensed on a relatively liberal basis. 0 Federal Trade Commission, Economic Report on Antibiotics Manufacture, 1958, p. 230. See United States Senate, Subcommittee on Antitrust and Monopoly, Report of tlze ,S'tudy on Administered Prices in time Drug Industry, 87th Congress, 1st Session, 1961, pp. 81-8. 8 In this context it is instructive to note the statement made by an industry spokesman in 1950. When considering the industry's future, he declared: "From a profit point of view . . . the only realistic solution of this problem lies in the development of new and exclusive antibiotic specialties. This . . . Is an exceedingly costly and vigorous alternative; nevertheless, it is the avenue of approach being most extensively explored by certain antibiotic houses today. This is the approach being followed by Pfizer". Statement by John McKean, President of Chas. Pfizer & Co., quoted in Senate Report, op. cit., pp. 130-31 (italics added). Between 1951 and 1960, research and development expenditures in the pharmaceutical Industry grew from $60.4 million to $206.5 million, an increase of over 200 per cent. (Pharmaceutical Manufacturers' Association, Yearbook, 1961-62, p. 168.) These figures are measured in current dollars and are derived from surveys of P.M.A. member firms. PAGENO="0521" COMPETITIVE PROBLEMS IN THE DRUG INDUSTRY 2071 introduction of new products and the accompanying extension of product differ- entiation. Research laboratories may, in fact, be regarded as a cost of differentia- tion. Facilities were designed for the discovery, development, and testing of new drugs, and as a result, new products were introduced at a rapid pace. While completely new chemical entities were the most desirable type of new product, since these would provide the highest degree of differentiation, other forms were also useful for this purpose, and in addition frequently had the advantage of requiring less research input per product. Among the latter group were new combinations of drugs already on the market, and new dosage-forms of previously introduced products. During the decade ending in 1960, over 3,800 new products and dosage-forn~s were introduced into U.S. pharmaceutical markets, or an aver- age of more than 380 per year. Of these slightly more than 11 per cent. of the total were new single chemical entities.1° THE IMPACT OF COMPETITIVE PRE55IIIIE5 In examining t.he impact of competition on the growth of research, it is useful to consider the rOles played by certain structural characteristics of the industry. The most significant of these is the co-existence of a number of fairly large firms with a sw-arm of smaller ones, which, however, account for a substantial propor- tion of the total output of the industry. Relevant data are presented in Table 1. As may be observed, the industry overall is moderately concentrated, with a significant fringe of smaller firms.1' Some further important characteristics con- cern the relatively low proportion of variable costs in total costs-in the larger firms the proportion tends to be less than half `2-and a generally low price elasticity of demand for pharmaceuticals.13 In a siaution of this sort, not only are competitive price declines likely to be large but also firms could easily be com- pelled to price below average total costs. As a result, there would be heavy pres- sures on profits if rivalry were allowed to take the form of price competition. TABLE 1.-PHARMACEUTICAL INDUSTRY: CONCENTRATI ON RAIl OS, 1958 Sales Amount (millions) Percent Largest 3 companies Largest 7 companies Largest 15 companies All companies ~384 800 1, 190 1,945 20 41 61 100 Sources: Pharmaceutical Manufacturers' Association, "Domestic Human Ethical Drug Business Survey," conducted by Arthur Andersen & Co., 1959, mimeographed. Pharmaceutical Manufacturers' Association, "Yearbook, 1961-62," p. 170. The data for the largest 15 firms are derived from a survey of sales in 1958 of 50 pharmaceutical firms, all members of the Pharmaceutical Manufacturers' Association. Given the nature of the survey, it can be assumed that the 50 com- panies included the largevt 15 in the industry. The figure for industry sales refers only to P.M.A. membership; it has been maintained by the association that member firms account for over 95 porcent of total industry output. The interaction of these two sets of factors has been significant in determining industry behaviour and performance. Where smaller firms are relatively im portant, one might expect to find a high degree of price competition. That this result has not occurred can be attributed to the post-war strategy of research and product differentiation pursued competitively by the leading firms in the industry. Rivalry has been restricted largely to areas other than price. A high degree of price stability on existing products has been maintained, and new prod- ucts have been priced, for the most part,: to compete with older ones in the same 10 Paul tie Haen, New Product Survey, 1060. 11 The pharmaceutical industry falls easily into Kaysen and Turner's Type II oligopoly. According to their definition, a Type I oligopoly is one in which the largest eight firms account for at least 50 per cent of industry output while the largest twenty comprise at least 75 per cent. A Type II oligopoly, on the other hand, is one in which the largest eight firms account for at least 33 per cent. of industry output. See Carl Kaysen and Donald F. Turner, Antitrust Policy, p. 27. 12 In a sample of twenty-two largest pharmaceutical firms, and with regard to 1958 drug operations only, costs of goods were only 43 per cent. of total costs. See Senate Report, op. cit., p. 31. `~ See The Economics of Research and Development in the Pharmaceutical Industry, pp. 58-60. A similar conclusion was reached: in the Senate Report, op. cit., p. 3. PAGENO="0522" 2072 COMPETITIVE PROBLEMS IN THE DRUG INDUSTRY therapeutic class,'4 At the same time, however, a vigorous form of rivalry has developed, which has taken the form of what Dr. Penrose has termed "competitioox in creativity".'5 A policy of competitive innovation seems to have been formulated by the individual firms, and research laboratories have been established as a vehicle of this policy. Profits have come to depend primarily on the firm's position in the innovative race, and the cost of an organized and efficient research effort may be regarded as a necessary expense of entering the race. Once some firms had taken this position, moreover, others were forced to follow in order to maintain their market positions relative to their rivals. In this manner the growth of pharmaceutical research establishments has been stimulated to a considerable extent by competitive pressures. Not only has the latent price competition, arising from these pressures, impressed upon the larger firms the need to achieve effective product differentiation, but also a highly competitive form of research and development has been promoted. Competi- tive factors have influenced the direction and scope as well as the size of research establishments. PRODUCT COMPETITION WITHIN THERAPEUTIC MARKETS The expansion of pharmaceutical research facilities and the concomitant in- crease in the introduction of new products have given rise to a rapid rate of product obsolescence. New products have competed with older ones, and the latter have been rapidly replaced by the former. By 19G0 only 31 per cent of sales accrued from products introduced before 1951.16 As may be seen in Table 2, the sales of products introduced in a given year normally reached their peak (as a proportion of total sales) in the second year following their introduction; after that they declined rapidly in relative importance. New products were intro- duced, reached their peak, and declined; and all within a relatively short period of time." TABLE 2.-SALES OF PHARMACEUTICAL PRODUCTS BY AGE OF PRODUCTS Year of introduction of products 1951 1952 Percent of sales 1953 1954 1955 - 1956 1957 1958 1959 1960 1960 4.6 1959 7.1 10.5 1958 4.7 7.2 7.3 1957 10.1 16.8 16.7 15.1 1956 5.8 9.0 8.4 7.7 6.6 1955 9.0 18.7 15.9 11.4 10.2 8.2 1954 7.4 13.7 12.8 10.8 10.2 8.8 8.1 1953 5.4 16.0 14.8 12.1 9.8 7.2 5.8 4.7 1952 4.6 9.5 7.7 6.1 4.8 4.4 4.0 3.1 2.9 1951 1.1 1.7 1.6 1.1 .8 .7 .6 .6 .6 .6 Before 1951 -- - 98.9 93.7 83.5 67.9 55.6 45. 1 39. 5 36. 7 32.8 31.4 Total - 100.0 100.0 100.0 100.0 100.0 100.0 100.0 100.0 100.0 100.0 Source: Arthur D. Little, Inc., "A Report on the Aspects of Concentration and Product Obsolescence in the Pharma- ceutical Industry in the United States," p. 33. The table is based on a sample of 33 firms, comprised, for the most part, of the larger firms in the industry. These firms account for approximately 80 percent of industry output. 14 A similar view was expressed by the Senate Subcommittee, which reported: "When a new product is put on the market, the customary procedure is to introduce it at or very near the price charged for an existing drug used to treat the same general type of ailment". Senate Report, op. cit., p. 98. "E. T. Penrose. The Theory of the Growth of the Firm, p. 106. "In the United States a kind of `competition in creativity' has become a dominant motif In the pattern of competitive behavior in many industries, where consumers and producers alike are caught up in an almost compulsive obsession for that which is `new', In the extreme case the inch- vidual firm Is forced constantly to remould its products-to create the `new' and `improved' either in performance or design. To a large extent the new products are superior in per- formance: to a considerable extent they are merely new and can be sold only if the consumer can be convinced that the `newest' is the `best'." 16 Arthur D. Little, Inc., A Repos-t on the Aspects of Concentration and Product 00- solescence in the Pharmaceutical Industry in the United States, p. 33. 17 Somers and Somers report that the average life-span of a new drug is said to be be- tween two and five years. Herman 1~I. Somers and Anne R. Somers, Doctors, Patients and Health Insurance, p. 96. PAGENO="0523" COMPETITIVE PROBLEMS IN THE DRUG INDUSTRY 2073 Because of the segmented nature of demand, it is misleading to consider the market for pharmaceuticals as a single entity; the relevant market does not span the entire industry but, rather, is limited to individual therapeutic classifications.18 And it is within these therapeutic classifications that the rapid rate of product introduction has developed into vigorous product competition. New products have replaced older ones, and in this process the ranking of the leading firms in particular therapeutic markets has changed frequently. In only nine out of twenty such markets did the same firm persist as the leading firm between 19.51 and 1960; and in most cases the same five firms did not retain completely the five leading positions.1° While past position tends to be an advantage in determining present market shares, it is not of crucial significance, and the measure of market control derived from a single product appears to be limited to a relatively short period of time.21 NEW PRODUCTS OR NEW PROCESSES Research laboratories in the pharmaceutical industry have been concerned largely with the introduction of new products rather than with the development of new processes n for old products.22 Given the nature of competition, this is what we should expect. Where product rivalry is high and the effective life of in- dividual products is correspondingly short, firms are unwilling to invest large amounts towards reducing production costs. By the time new techniques have been developed, it is quite likely that demand for the product concerned will have dropped to a relatively low level. Even with regard to products which have fairly long lives, moreover, or to processes used to produce many products, it is unlikely that a large research effort is undertaken in the direction of reducing production costs. To the extent that rivalry is based on price, the introduction of new processes will place' the firm at a' competitive advantage by enabling it to price below its rivals without reducing its profit margins. When, however, price behaviour is largely non-competitive, a reduction of costs will increase profit margins without immediately affecting the firm's position relative to those of its rivals. While this may enable the firm to undertake large'r selling or re- search expenditures, or to' raise funds more advantageously in the capital mar- ket, the competitive effect will not be immediate or direct. The introduction of new products affects the competitive position of the firm in a quite different manner. To the extent that rivalry takes the form of coin- petition between products which are priced at the same or similar levels, the number and character of new products which are introduced directly affect the demand for the firm's output. If the firm introduces new products which do the job "better", then its output and total profits may be higher even if the cos'ts of the new products are greater than those of their predecessors and profit mar- gins correspondingly reduced. The firm that falls behind in the race to intro- 18 Relatively high cross-elasticities of demand will be found when competing products treat the same or similar ailments or have similar therapeutic effects. Between therapeutic classifications, on the other hand, these elasticities will approach zero. While cases will remain in which a specific product is required to treat a certain condition, the gap in the chain of substitutes will appear at the boundaries of therapeutic classes. Although high cross-elasticities of supply will exist across these boundaries, we still find it appropriate for our purposes to define the relevant markets in terms of therapeutic classifications. 19 Data describing the market shares of the leading five firms within twenty them- peutic markets for the years 1951 and L960 are presented in Arthur D. Little, Inc., op. cit., pp. 11-30. 21 This conclusion is corroborated by McKie who reports the following: "In all cases it is clear that the period of `dominance' of any one product is short-four or five years at most-and that a firm which fails to bring out improvements or new substitutes will find its share of the market rapidly passing to others". James W. McKie, "An Economic Analysis of the Position of American Home' Products Corporation in the Ethical Drug Industry", in United States Senate, Administered Prices in the Drug Industry, Hearings before the Subcommittee on Antitrust and Monopoly, 86th Congress, 2nd Session, 1960, Part 17, p. 9957. The latter document will be cited as Administered Price Hearings. 21 `The distinction between products and processes cannot be based on. characteristics of the good in question. Where intermediate products are concerned, a product for one' firm may be a process for another. The distinction is actually relevant only within the firm. A product is a good which is offered for sale in the normal course of business. A process, on the other hand, is only infrequently offered for sale, and is used within the firm to facilitate production of other commodities. 22 This emphasis was sufficiently great for a survey of resea.iich budgets of pharmaceutical firms to allocate total outlays among four.' categories without providing one for new process research. The categories were: new products, product improvements, new appli- cations, and basic research. See Chemical and Engineering News, March 17, 1958, p. 52. PAGENO="0524" 2074 COMPETITIVE PROBLEMS IN THE DRUG INDUSTRY duce new products may find its demand and profits lower than its rivals', even if it should succeed in reducing the costs of producing its older products. This will be the case especially if the price elasticities of demand for it~ products are relatively low. This explanation involves the question of profit maximization under con- ditions of oligopoly. Where profit rates are minimal, the pressures to reduce costs are associated with staying in business. Where profit rates are higher, however, the pressures to reduce costs via the introduction of new techniques are likely to be substantially less. But the pressures associated with the intro- duction of new- products may be similar to those existing in the competitive model. Here the introduction of competitive new products by a firm's rivals may result in a large, and in some cases fatal, decline in the demand for the firm's output. While the conventional theory of profit maxhuization requires that the "car- rot" be as effective as the "stick" in eliciting certain kinds of behaviour, it ap- pears likely that the prevention of declines of profit is more important than the making of gains, that the maintenance of existing market shares through the introduction of new products is more important than the reduction of cost, and that a firm n-ill work harder for the former purpose than for the latter. This means, merely, that in an uncertain world firms operate under some kind of minimax strategy. On this basis, firms will emphasize new products rather than new processes in their research efforts. THE ROLE OF THE PATENT SYSTEM United States law provides that patents on pharmaceuticals may be granted for new products as well as for new processes. Process patents, however, are a considerably weaker form of protection. As in most chemical industries, it is comparatively easy to modify a process somewhat and thereby evade the patent. Product patents, on the other band, are a relatively strong vehicle of protection from competive suppliers. Once a product patent has been granted, a rival firm cannot supply the identical compound without fear of legal proceedings. As a result, nearly 80 per cent of medical patents are issued for new products rather for new processes.23 Patent protection in this industry is, moreover, especially sig- nificant since a large proportion of individual products are covered, A survey of the industry estimated that over two-thirds of all prescription sales are for patented drugs.2~ Although patent protection in the pharmaceutical industry takes the form primarily of product patents. it does tend to be limited to specific chemical sub- stances. In most cases patents cover only a single compound or a small number of compounds. Especially when the advancement in knowledge is small, the scope of the patent may be rather limited. In addition, there is the question of patent specifications and the problem of anticipating all possible variants of the product.n With regard to pharmaceuticals, it is frequently possi,ble to invent around existing patents; to find a variant which has not been specified, obtain a patent for it, and introduce it as a competing product. Since a large proportion of pharmaceuticals have some degree of patent pro- tection, entry into a specific therapeutic market requires, in most cases, some form of scientific or chemical product differentiation. In a world of competing monopo- lists, rivalry requires the ability to acquire a monoply position. The importance of pharmaceutical patents, however, can easily be overstated. While monopoly positions are conferred, patent protection does not normally confer the power to monopolize any of the therapeutic markets. This is borne out by the high turn- over among leading firms and the vigorous product competition within these markets. The impact of the patent system has not been to create monopoly posi- tions which remain active throughout the seventeen-year life of the patent, but rather it has been to foreclose to a great extent rivalry between identical chemi- in 1957 the industry's rate of return on investment after taxes encuallecl 21.4 per cent, which placed it second on the F.T.C. list of thirty-nine industries. The comparable rate reported for All Manufacturing was 11.0 per cent. See Federal Trade Commission, Report on Rates of Return for Identical Companies in ~Selected Manufacturing Industries, lDJiO, 1947-1957, pp. 34-49. and Senate Report, op. cit., pp. 53-55. 24 In a sample of the twenty-two major pharmaceutical firms, for 1958 drug operations only, selling expenditures reached nearly 25 per cent of total sales. Research expenditures, on the other hand, equalled only slightly more than 6 per cent. Senate Report, op. cit., p. 31. Arthur P. Little. Inc.. op. cit. pp. li-so. While these ra~os are probably over-stated, the extent of the over-statement is not likely to be by much more than one-fifth. PAGENO="0525" COMPETITIVE PROBLEMS IN THE DRUG INDUSTRY 2075 cal entities or standardized commodities about which price competition might develop. It has strengthened and encouraged the high degree of chemical product differentiation which is the primary form taken by technical change within the pharmaceutical industry. THE ACHIEVEMENT OF MARKET POWER The growth of effective product differentiation has led to an appreciable in- crease in industry profits and their maintenance at a relatively high level. In fact, the pharmaceutical industry has `become one of the most profitable in the American economy.26 The high level of market power, of which these profit rates are indicative, has evolved as part of the' pattern of inter-relationship among the major firms, and is founded upon the aChievement of product differentiation. The crucial significance of product differentiation is that it provides the pri- mary barrier to entry into the relevant therapeutic markets. Since effective entry normally requires some form of technical advance, the cost and risk of research coniprise an important part of this barrier. Joined with research and develop- ruent, moreover, `are the extremely high selling expenditures undertaken by the larger firms.27 Not only do these outlays accentuate the degree of differentiation among older products, but they also raise considerably the costs associated with launching a new product, and thereby provide a further barrier. Entry barriers, created in this fashion, have resulted in fairly high levels of concentration within therapeutic markets. In a group of twenty such markets, the proportion of output accounted for by ~he leading five firms ranged from 56 per cent to 98 per cent.28 It is within these markets that decisions on prices are made, and given such concentration ratios, `we should not expect individual firms to disregard their own impact on market parameters. It is on this basis that market power has ~ieen achieved. In this manner the leading five firms in the industry have maintained control of a large number of marke'ts, and have' created conditions within which the rivalry among themselves will proceed on a non-price basis. What is equally important, the achievement of product differentiation has succeeded in impeding the entry into these markets of the large number of smaller firms which would be likely to introduce some measure of price:' competition. Smaller firms have been forced to rely largely on standardized and non-patented products which are frequently non-competitive with the highly: differentiated products which lead in most markets. To consider further the role played by differentiation, it is instructive to examine the behaviour of the industry in p'atent licensing. Although pre-emption of the entire demand for a product seems most desirable for the firm, there are a number of factors which have increased the scope of licensing agreements. Not only may smaller firms wish to profit from the dis:tribution facilities of their larger rivals,22 lint cross-licensing agreements may also be required to produce drug combinations and to settle "interference" proceedings.2° In all of these cases, however, there appears to be a definite reluctance on the part of the major firms in the industry to license their smaller rivals even when licenses are granted to 23 See the table compiled by the United States Patent Office of all patents relating to medicine issued during 1901. This table appears in United: States Senate, Drug Industry Antitrust Act, Hearings before the Subcommittee on Antitrust and Monopoly, 87th Con- gress, 1st Session, 1961, Part 3, p. 1261. This document will be cited as Antitrust Act Hearings. 2~1bid., Part 5, p. 2621. 2S See the statement by George E. Frost, a member of the patent bar, in Antitrust Act Hearings, Part 4, p. 2119. Frost maintained with regard specifically to pharmaceutical patents that "the patent applicaat is rarely able to anticipate and include all variants of his inventive concept in this document". ~ The classic example is the case of meprobamate. The patent here is assigned to Carter Products, Inc., which markets the product under the name of Miltown. Within a month after the issue of the original patent,' Carter licensed American Home Products Corp., one of the industry's largest firms with extensive selling and distribution facilities, to market the product under its own name, Equanil. Although sales of the latter quickly exceeded their own, Carter benefited through extensive royalties from the expansion of demand stimulated by the larger firm. ~° An "interference" is declared by the Patent Office when a number of patent applications lay claim to the same invention. While the normal procedures in this case involve the Patent Officein an administrative hearing to determine the true inventor, the private settlement of' claims is widely used in the pharmaceutical industry. This commonly re- sults in the withdrawal of all but one of the' original applications and the licensing of all parties'when the patent is finally issued. See Senate Report, op. cit., pp. 152-54. PAGENO="0526" 2076 COMPETITIVE PROBLEMS IN THE DRUG INDUSTRY other large firms. The motive is likely to have been the fear of instigating some form of price competition. THE DIRECTION OF RESEARCH The pharmaceutical industry during the post-war period has undertaken a heavy research effort which has been founded no a number of factors. The most important of these was an industrial structure characterized by both competitive and oligopolistic elements. The fact of vigorous price competition on the ~asis of standardized commodities led to a search for an effective method of achieving product differentiation. In addition, the development of a new technology, based originally on scientific advances made outside the industry, favoured the move towards a chemical form of differentiation. And finally, the existence of a strong patent system increased substantially the degree of differentiation which could be gained through the introduction of new products, and thereby strengthened considerably the incentive for research. These factors, however, have not only determined the extent of research in the pharmaceutical industry, but also they have influenced the type of work which is pursued and the framework within which it is undertaken. Pharmaceutical research is conducted within a context of strong product rivalry. As a result a substantial amount of work is done which is designed pri- marily to invent around existing patents. While a number of important improve- ments have been made in this fashion, this work has led also to the development of a great many products which have pharmacological effects that are quite similar to those of drugs already on the market.3' Moreover, a good deal of work has gone into developing new combinations of existing drugs and new dosage- forms. While research of this type is unlikely to result in major scientific ad- vances, and frequently duplicates what has already been done, the minor advances and improvements w-hich do arise are, nevertheless, important elements in the process of developing a new technology, and are essential if its full benefits are to be realized through the inevitable plethora of modifications and variations. Product rivalry and the drive to strengthen product differentiation have led also in a second direction, w-hich is concerned with the process by which new scientific knowledge is translated into marketable products. Competitive pres- sures have forced research directors in the industry to pay constant attention to new knowledge arising from external sources and have ensured that promising leads w-ill rapidly be acted upon. This form of product development is vitally important because the major share of new scientific knowledge is likely to be produced in laboratories outside the industry. Not only do the research expendi- tures of the industry account for only about 30 per cent of total resources devoted to medical research within the United States,'2 but also considerable scientific activity is pursued abroad. Even where the basic discoveries are made elsewhere, however, a good deal of work remains to be done." The first problem concerns isolating or otherwise obtaining compounds which may have promising properties. It has been esti- mated. in fact. that approximately one-half of total pharmaceutical research and development expenditures is used to support the w-ork of synthesizing, purifying, modifying and preparing suitable substances for subsequent physio- logical tests.'4 Moreover, quite extensive biological and clinical testing is required 3' With regard to products in the latter group, the term "molecule manipulation" is commonly used because of the similarity in molecular structures which is frequently evi- dent. An example of this feature of pharmaceutical research is provided by the case of three competing antibiotics: tetracycline, aureomycin, and terramycin. Although there may be some differences in their medical activity, these are not likely to be great, and the products may be considered as substitutes. Their molecular structures, although highly complex. are all rather similar. These structures are presented in Federal Trade Commission, Economic Report on Antibiotics Manufacture, 1958, p. 249. 3' In 1960 the pharmaceutical industry allocated $215 million to research and devel- opment out of total national expenditures of $715 million. Government provided $400 million while other private support amounted to S100 million. These figures refer to the U.S. fiscal year 1960. Antitrust Act Hearings, Part 3, p. 1705. 3' Despite the original discovery of penicillin by Alexander Fleming in 1928, thorough testing of the substance did not begin until 1939, and the development work, even under the exigencies of war. was not accomplished until 1943. It was only after the Second World War that penicillin became available on a significant scale to the civilian population. For a survey of the early discovery and development of antibiotics, see Federal Trade Com- mission, op. cit., pp. 34-45. 3' Statement by Austin Smith. President of the Pharmaceutical Manufacturers' Associ- ation, in Administered Price Hearings, Part 19, p. 10725. PAGENO="0527" COMPETITIVE PROBLEMS IN THE DRUG INDUSTRY 2077 before a new drug may be introduced. Research and development activities of this sort are extremely costly, although there is a reasonable degree of assur- anice of commercial viability. They demand extensive resources devoted to scien- tific efforts which, however, may be fairly routine. Thus, a major share of phar- maceutical research should be considered as complementary, rather than competi- tive, to research activities outside the industry. TABLE 3.-NUMBER AND PERCENT DISTRIBUTION OF SCIENTIFIC AND PROFESSIONAL MANPOWER IN MEDICAL AND RELATED RESEARCH, 1960 A. DISTRIBUTION BY SECTOR ENGAGED Levelof training Total M.D. and Ph. D. Other D.D.S. Number 39,700 11,400 18,000 Percent: 10,300 Government 19. 7 17. 3 12. 6 34. 5 Industry 18.1 4.8 12.2 Universities and research institutes 62. 2 77. 9 75. 2 43.2 22. 3 Total 100.0 100.0 100.0 100.0 B. DISTRIBUTION BY LEVEL OF TRAINING Percent Number Total M.D. and Ph. D. Other D.D.S. Gnvernment 7, 800 100. 0 25. 3 29. 2 Industry 7,200 100.0 7.6 30.6 Universities and research institutes 24, 700 100. 0 35. 9 54. 8 45. 5 61.8 9. 3 Total 39, 700 100. 0 28. 7 45. 3 26. 0 Source: National Institutes of Health, `Manpower for Medical Research-Requirements and Resources, 1965-70," a report for the Committee on Appropriations, U.S. House of Representatives, February 1962, p. 24. To consider this poin.t further, it is useful to examine the data on educational levels of professional personnel engaged in medical research within and outside the pharmaceutical industry. The relevant information is presented in Table 3. As may be observed, 38 per cent of research professionals in the industry held doctoral degrees while 55 per cent of government researchers and 91 per cent of professionals within universities and research institutes had attained this level. Moreover, although the pharmaceutical industry provided approximately 30 per cent of total expenditures on medical research in 1960, it utilized only 5 per cent of total personnel with the M.D. degree and 12 percent of those with the Ph. D. degree. It appears, thus, that industry resources competed with non-industry resources for scarce, highly trained personnel to a lesser extent than would be in- dicated by the total size of the industry effort, and also that these resources w'ere used primarily for alternative, non-competitive purposes. The growth of competitive product differentiation has been associated w-ith sub- stantial outlays on research and development; but equally significantly, it has been associated with a specific emphasis in its direction. Research is a generic term which covers a broad spectrum of activities, and thus the type or character of the research activities undertaken as well as their volume are important fac- tors in an analysis of the relationships among mnaricet structure, research and technical change. PAGENO="0528" 2078 COMPETITIVE PROBLEMS IN THE DRUG INDUSTRY [From The Review of Economics and Statistics, Vol. XLVII, No. 2, May 1965] RESEARCH AND TECHNICAL CHANGE IN THE PHARMACEUTICAL INDUsTRY: (William S. Comanor) In recent years there has been a good deal of discussion concerning the relation- ships among market structure, research and development, and the rate of tech- nical change. Much of this discussion has focused on the question of whether large firm size is a necessary condition before firms w-ill engage in research, and whether research and development (R and D) is likely to grow more or less thali in proportion to increases in firm size. A further set of questions deals with the relationship between research and the rate of technical change experienced by the firm. Can variation in the latter be explained largely by differences among firms in the size and character of their research programs? Are economies of scale in R and D likely to be present? What is the effect of firm size on the pro- ductivity of a research establishment? This paper provides an empirical analysis, concerned with these questions, of the experience of the United States pharma- ceutical industry during the period between 1955 and 1960. SOME MEASURES OF RESEARCh AND TECHNICAL CHANGE The empirical investigation in this paper utilizes multiple regression techniques on a cross-sectional basis at the firm level. It examines the impact on technical change of a number of variables associated w-ith the character of a firm's research and development effort. At this point, we will define the measures used to repre- sent technical change and R and D, and describe the sample upon w-hich the analysis is based. Research and development establishments in the pharmaceutical industry 1 are concerned to a very large extent with the introduction of new- products, and only meagerly with the development of new processes.1 This is due to the fact that the primary mothating force behind R and D outlays in this industry lies speci- fically in the attempt to achieve scientific or chemical product differentation.3 ucts at reasonable levels of cost, these are properly included under new product research, for we are concerned here not merely w-ith the discover or invention of the product, but with the total R and D outlays, asseciated with its introduc- tion. As a result, research output, in this paper, is measured entirely in terms of new proJucts and we assume that the reesarch effort designed to produce new processes is small and can be safely ignored. A concern with research output in terms of new products requires considera- tion of all products introduced w-hich are new to the firm regardless of whether they have been introduced previously by competitors. This includes those prod- ucts which are innovations or imitations and those which lie in an intermediate position, embodying varying degrees of differentiation relative to products already on the market. This last category, which in the pharmaceutical industry in- cludes the great majority of new products, covers the entire spectrum of which pure innovations and pure imitations are the extreme values. Our measures of research output are, thus designed to represent the rate of new product techni- cal change experienced by the firm.4 *This paper is based on the author's unpublished Ph.D. dissertation: "The Economics of Research and Development in the Pharmaceutical Industry," (Department of Economics, Harvard University, June 1963). I should like to express my appreciation to R. B. Hefie- bower, J. W. Markham, and L. D. Taylor for their counsel and advice during the course of my work on this project. I am also indebted to R. E. Caves, P. Dhrymes, W. E. Gustafson. J. Schmookler, and T. A. Wilson for a number of helpful suggestions, and to P. dellaen and H. A. Gosselin who generously provided the data used in this paper. The study was supported financially by the Science and Public Policy program at Harvard. and the coin- putations were carried out under grants of subsidized time on I.B.M. 7090 computers at the Harvard Computation Center and at the M.I.T. Computation Center. 1 In this study, we shall distinguish ethical drugs from proprietaries, and shall consider the pharmaceutical industry in terms of the former. Pharmaceuticals are marketed and sold only through the medical professions, and require, for the most I)art, a written medical prescription. So great is this emphasis that a 1958 trade journal survey of research bucl~ets of pharmaceutical firms divided expenditures into four categories without providing a cate- gory for new process research. See Chemical and Engineering NCws (March 17, 1958). 52. Although funds are expended to develop techniques for mass producing new prod- 1 See W. S. Comanor, op. cit., 65-85. Technical change is considered to encompass Imitation as well as innovation and thereby deals with the entire process by which new technology is placed into actual practice. PAGENO="0529" COMPETITIVE PROBLEMS IN THE DRUG INDUSTRY 2079 New pharmaceutical products may be divided into a number of classes con- cerned with the innovative character of a new drug. In the first category are the single chemical entities which have not been introduced previously into United States markets. While these are largely innovative, they do include salts or derivatives of older products. Other classes deal with new combinations of active ingredients and new dosage forms of previously introduced products. These groups contain imitative as well as innovative characteristics. At the same time, many new products are purely imitative and merely duplicate products already on the market. In this study, we shall measure technical change by means of two variables. The first of these deals only with the introduction of new chemical entities, while the second includes a large imitative component and encompasses new products of all types. New products vary not only in terms of the degree of innovative character which may be associated with them, but also in terms of their economic impact. To deal with this factor, we have weighed each new product on the basis of its sales during the first two calendar years following introduction.5 We shall designate Y1 to equal total sales, in their first two years, of all new chemical en- tities introduced by the firm during the period between 1955 and 1960, and Y2 to equal a comparable sum when "new products" assume the broader definition.6 The scale of research and development has also been measured in two ways, both dealing with the number of persons employed in research establishments The first variable is the average number of professional B and D personnel employed in 1955 and 1960. The second deals with the average number of total personnel employed in R and D facilities at these two points in time. This latter value equals the sum of professioal and supporting personnel. The sample used in this study includes 57 pharmaceutical firms,8 and accounts for nearly 80% of total pharmaceutical prescription and hospital sales during the period between 1955 and 1900. Moreover, the sample is not dominated entirely by large companies, but rather covers the entire range of the size distribution of pharmaceutical firms.9 The distribution of the sample is J-shaped with one- Sales data by product were obtained from a marketing research firm, R. A. Gosselin & Co., Inc., which supplied information for the years 1955 through 1961. These figures are derived from a representative sample of the nation's drug stores and hospitals. One short- coming of these figures is that they do not represent total output but only that portion which is sold to consumers via prescriptions and to hospitals, and thereby do not include sales to government, corporations, or those made in physicians' offices. Prescription and hospital sales, however, account for approximately 80% of total pharmaceutical sales. See Lucy Kramer, "Drugs and Medicines" in Public Health Reports (Oct. 1938). 932. We are forced to assume that our conclusions would not be substantially altered if the remaining portion were included. Information on new products was obtained from Paul deHaen, a consultant to the pharmaceutical industry. In addition to providing informa- tion on year of introduction, these surveys noted the therapeutic classification of each new product and whether the product at introduction was a new chemical entity, duplicate, new compounded product, or new dosage form. For a more detailed discussion of the value and limitations of these data, see W. S. Comanor, op. cit., 99-106. 6 The data on pharmaceutical sales, used in the creation of these variables, have not been deflated for price changes but rather have been expressed in current dollars. This step was taken because of the high degree of price stability among pharmaceuticals during the period under consideration. See the price index computed by John M. Firestone in United States House of Representatives, Drug Industry Antitrust Act, Hearings before the Antitrust Subcommittee of the Committee on the Judiciary, 87th Congress, 2nd Session (1962), 608. Information on this subject for the years 1955 and 1960 has been published in Na- tional Academy of Sciences-National Research Council, Industrial Research Laboratories in the United $tates, tenth and eleventh editions. These surveys distinguish between pro- fessional and supporting staff within R and D establishments. In addition, for those companies in which the research effort was divided among a number of separate labora- tories, the size and direction of research of each laboratory were, in most cases, listed individually. Since in many large diversified firms, the pharmaceutical R and D labora- tories were distinguished from other research facilities, it was possible to limit our data for these firms to pharmaceutical research personnel. In the inevitable instances when we came across laboratories in which work was done on non-pharmaceutical as well as pharmaceutical projects, the facilities were included when it appeared from the sum- maries provided that the major part of the effort was concerned with pharmaceutical endeavors. 8 The sample was chosen on the basis of data availability but with the condition that no substantial merger took place with other pharmaceutical firms between 1955 and 1960. Exceptions were macla for those mergers which occurred during 1960 and in which the 1960 R and D survey reported separate and distinct research facilities. It was as- sumed, in these cases, that the pre-merger firms behaved and acted independently through- out the period. Size, in this study. is measured by the mean value of annual prescription and hospital sales betweea 1955 and 1960. 81-2S0-68-pt. 5-34 PAGENO="0530" 2080 COMPETITIVE PROBLEMS IN THE DRUG INDUSTRY third of the observations falling in the smallest size class. This sample is essentially non-random, and strictly speaking, our empirical con- clusions apply only to it and not to any larger population. Since, however, the sample constitutes such a large proportion of industry output, and has a size distribution similar to one which might be expected from the entire industry, we may presume that conclusions based on an analysis of this group of firms prob- ably can be applied to the entire industry. THE MAJOR EMPIRICAL FINDING5 In the analysis which follows, a quadratic term for research and development is introduced as well as a variable designed to represent the interaction between size of firm and scale of research establishment.10 Some hypotheses concerned with the question of economies of scale in R and D distinguish between the productivities at a given level of research activity according to firm size. For this reason, a variable was computed which equated the product of scale of research and size of firm. Since R and D is measured on two bases, there are also two measures of the interaction between research and firm size. We may designate I~ as the interaction variable where R and D is the number of pro- fessional, and 12 where the number of total personnel is used. In addition, we include firm size as an explanatory variable and also a measure of output di- versification.11 The latter step enables us to test the relationship between this factor and the productivity of research. In the regression analysis, the measures of technical change and R and D are deflated by size of flrm.'~ As may be seen in table 1, the model fits the data far better when research is measured by professional rather than total R and D personnel. While we shall have more to say later about the question of supporting personnel, it appears that technical change is primarily associated w-ith the number of professional investigators. It may also be observed that diversification is negatively associated with our measures of technical change. When new products are defined in terms of new chemical entities, the coefficients are significant at the 99% level. With the broader definition of new products, however, the coefficients remain negative although there is some doubt as to their statistical significance.13 It thus appears that diversification is more closely related to the introduction of new entities than to new products in general. In addition, the negative dimension of the parameters suggests that for given research and development, higher rates of technical change will be achieved if attention is concentrated towards a few product areas. To the extent that diversification of output denotes the scope of research ac- tivities, it may be that inefficiencies result from R and D undertakings w-hich are "spread too thin" and that in the context of pharmaceutical research, it is better to work exhaustively with a limited number of problems.14 A further point is that Y1 appears more closely associated with research and development than Y2. We would expect this to be the case because F1 deals only with new chemical entities w-hose introduction is likely to have a relatively high degree of research input. Since F2 includes new chemical entities as well as other new products, the correlations observed with this variable may denote largely the influence of new entities. ~° I am grateful to Lester D. Taylor for originally suggesting the use of an interaction variable. 11 Diversification, in this analysis, deals only with the division of output among the various pharmaceutical markets and not with the division between pharmaceutical and non-pharmaceutical markets. On the basis of apparent medical usage, 40 therapeutic markets are defined. Our first measure of diversification, Di, is the number of markets serviced b-i- the firm which account for at least 2% of total pharmaceutical sales. The second measure, D2, deals with the proportion of sales outside the firm's primary market, and equals one minus the ratio of sales in the firm's largest market to total pharmaceutical sales. The third measure, D3, is the composite of the two previous ones. It is defined as the mean value of the products of Di and D2, calculated on an annual basis. These meas- ures are taken from Michael Gort. Diversification and Integration in American Industry, 8-li and 23-24. In these regression equations, D3 Is the variable which is introduced. 12 This step is taken to increase the likelihood that the assumption of homosceclasticity is satisfied. ii We should note that Di and D2 were introduced into the analysis and provided quite similar results to those obtained from D3. See W. S. Comanor, op. cit., Table XXXI, p. 140. 14 An alternative explanation of the negative sign of this coefficient concerns the possi- bility that there may be selling efficiencies which result from concentrating sales in a small number of therapeutic markets. PAGENO="0531" COMPETITIVE PROBLEMS IN THE DRUG INDUSTRY 2081 The relationship between research effort and technical change appears to be curvilinear and the quadratic term is always highly significant. The linear term, however, is not significantly different from zero and has an alternating sign when Y2 is used to measure technical change, but is significantly negative when 11 is used. The existence of a negative cofficient for research and development effort in these cases does not result from the nature of the research process and the experience of the pharmaceutical industry, but rather is due to some statisti- cal problems associated with the data concerned with the introduction of new chemical entities. These problems shall be discussed in the following section. TABLE 1.-Regression and correlation analysis-Summary findings A. RESEARCH AND DEVELOPMENT MEASURED BY PROFESSIONAL RESEARCH PERSONNEL . Intercept RD RD2 Size Ii D ~R2 (1) (2) y1 ~ b 0.422 (0. 136) b 0. 873 (0. 222) b -4.671 (1. 285) -0. 060 (2. 098) b 0.547 (0. 107) b 0. 557 (0. 175) b 0.0000344 (0. 0000083) 0. 0000289 (0. 0000136) b -0.000000128 (0. 000000031) c -0. ooooooioo (0.000000051) b 0.130 (0. 040) c -0. 111 (0. 066) b 0.40 C 0. 22 B. RESEARCH AND DEVELOPMENT MEASURED BY TOTAL RESEARCH PERSONNEL Intercept RD RD~ Size 12 D R2 (3) (4) 3( Y2 b 0. 471 (0. 150) b0.932 (0. 232) c -1. 989 (0. 981) 1.381 (1. 513) b 0. 112 (0. 031) bO.118 (0.047) b 0.0000300 (0. 0000102) 0.0000194 (0.0000157) b -0.000000051 (0. 000000017) -0.000000034 0. 000000027) b -0. 120 (0. 044) -0.100 (0. 068) b 0. 28 0.18 Throughout this study, statistical significance for the regression coefficients is determined by one tailed tests, and the significance of R2 by the F ratio test. b Indicates statistical significance at the 99-percent level. Indicates statistical significance at the 95- percent level. THE ZERO VALUE PROBLEM The data with which we are concerned cover a very wide range of variation in the scale of research facilities, and deal with the six years between 1955 and 1960 inclusive. Where new products include all introductions of whatever type, this period was sufficiently long to enable each firm to introduce at least one new product. In the case of new chemical entities, however, 17 firms, all with rela- tively small research establishments, failed to make even one introduction. As a result, the value taken by Y1 for each of these 17 firms is zero. Although the zero values are clustered at the lower end of the distribution of research, when measured absolutely, they appear -throughout the distribution when R and D is deflated by size of firm. If data for a longer time period had been available the number of zeros would most likely have been considerably reduced. The existence of 17 zero values among the dependent variables raises a number of questions concerning the interpretation of the statistical results presented above. Although the usual least-squares properties apply to our estimates, some doubt is cast on the assumption that the disturbances are distributed normally, on which our tests of significance are based. Even more important, however, is the possibility of distortion resulting from the discontinuity which is introduced by a substantial cluster of values at zero. The zero problem occurs only when Yiis used as the dependent variable. In no cases did Y2 take the value of zero. Because of this fact, it is useful to compare regression coefficients between equations based on each variable. As may be noted in table 1, the coefficients in all cases but one are not widely different and are frequently quite similar. Only with regard to the first degree term for R and I) do the coefficients diverge substantially. To test for distortion in the equations concerned with Y1, two techniques are used. The first is simply to eliminate the 17 zero values and compare the coeffi- cients based on 40 observations with those based on 57. The second approach is concerned with examining regression equations estimated from grouped data. PAGENO="0532" 2082 COMPETITIVE PROBLEMS IN THE DRUG INDUSTRY Firms are aggregated in groups on the basis of research input, and an analysis is undertaken on group means.'5 In this manner, we are able to examine the rela- tionship between research and technical change for firms even of the smallest size category without encountering the problems created by zero values in the dependent variable.'8 Table 2 presents the results of a regression analysis in which the same relation- ship has been examined in four formats.1' As may be observed, the equations in which the zero values are included and excluded present rather similar results. With the exception of the linear term for R and D, the regression coefficients do not take widely different values, although the standard errors of the coefficients are larger with the exclusion of the zero values. These parameters are highly significant. It does appear, however, that the significance and magnitude of the first degree R and D coefficient is peculiarly dependent upon the zero values. The grouped regressions present greater difficulties of interpretation. Although the units were relatively homogeneous with regard to absolute values of R and D, there may have been considerable variation within group boundaries in terms of the explanatory variables used in the model. Grouping A, which has the effect of consolidating only the smallest research establishments, provides coefficients which are reasonable approximations of the ungrouped values. The estimates, with the exception again of the linear coefficient for R and D, remain highly significant; the intercept remains significant although not highly so. The second grouping format influences the relationship to a larger extent than the first. The coefficient of diversification becomes a good deal more negative and the standard errors of the parameters are increased substantially. As a result, only two esti- mates remain statistically significant. With the exception of the first degree expression of research and development, it does not appear that our estimated regression coefficients have been distorted substantially by the inclusion of the zero values in the measure of technical change. It should be recognized, however, that we have not presented any theoretical explanation concerning the influence of these values.'8 TABLE 2-Regression analysis-Examination of the zero problem Intercept R. & D.~ R. & D.' Size . Ii D Number of obser- vations (1) Zero values included.... b 0. 422 (0. 136) b -4* 671 (1. 285) b 0. 547 (0. 107) b 0. 0000344 b 0. 0000083 b -0. 000000128 (0. 000000031) b -0. 130 (0. 57 (2) Zero values excluded_ - (3) Grouping A b 0. 690 (0. 227) 0. 424 (0. 219) -13. 584 (10.284) -3. 604 (3. 522) b 0. 545 (0. 144) b 0. 627 (0. 153) b 0. 0000306 (0. 0000114) b 0. 0000422 (0. 0000128) b -0. 000000118 (0. 000000041) b -0. 000000152 (0. 000000045) 040) b -0. 163 (0. 055) b -0. 169 (0. 066) 40 29 (4) Grouping B 0. 545 (0. 695) -2. 249 (6. 625) `0. 538 (0. 268) C 0. 0000430 (0. 0000225) -0. 000000114 (0. 000000088) -0. 237 (0. 194) 16 R. & D. is measured by the number of professional research personnel. Indicates statistical significance at the 99-percent level. Indicates statistical significance at the 95-percent level. 15 Grouping was undertaken in two formats on the basis of non-deflated values of professional research personnel. The first format was to lay out the entire distribution of firms from sniallest to largest in terms of the grouping variable and mark off boundaries so as to insure that at least 50 persons were included in each unit; grouping was begun at the small valise end of the distribution. Twenty-nine groups were defined in this fashion of which all but ten consisted of single firms. In the second format, six groups were defined between the values of zero and 50, and intervals of 25 marked off through- out the remainder of the distribution. This scheme provided 16 groups. IC If we assume that the disturbances associated with individual firms have constant variances, then groups of varying size necessarily introduce heteroscedasticity. To retain the efficiency of our estimates, it has been necessary to weight all observations by the square root of the number of firms in the group; this also includes the column of ones used to determine the intercept. 17 Since a grouping procedure in itself influences the degree of variation "explained" by a regression analysis on grouped observations, multiple correlation coefficients with group- ing cannot be compared with those obtained from the original data. For this reasosi, the R's associated with these equations have not been presented. `~ While a more sophisticated technique exists for dealing with this matter, the corn- putational problems involved were too great to warrant its use. See James Tobin, "Estima- tion of Relationships for Limited Dependent Variables," in Econometrica (Jan., 1958), 24-30. PAGENO="0533" COMPETITIVE PROBLEMS IN, THE DRUG INDUSTRY 2083 THE INFLUENCE ON TECHNICAL CHANCE OF SCALE OF RESEARCH AND SIZE OF FIRM In the model presented above, an interaction variable is introduced to test whether firm size affects the shape as well as the position of the relationship between research and technical change. Since it is necessary to deflate by size of firm in order to deal with the problem of heteroscedasticity, this approach is used to reintroduce a scale factor anti thereby examine differences in the form of the relationship as between small firms and large. In addition, an addi- tive expression denoting firm size is used tO consider the impact of this variable on the equation's position. As may be observed in table 1, the linear expression of firm size is positively associated with research output. Since new products have been weighted by their sales during their first two years after introduction, we should expect our measures of technical change to be influenced by such factors as distribution facilities, selling effort, and firm reputation. To the extent that these factors are correlated with firm size, the size variable introduced into the equations will represent their influence, and the coefficient will be positive. Firm size appears to have acted to increase the gains resulting from new product introduction, and these gains increased more than proportionately with size of firm. Our findings imply further that the relationship between research and devel- opment and the rate of technical change is substantially influenced by size of firm. Not only is the relationship shifted in total by differences in this variable, but also the functional form of the relationship is dependent on these values. Towards the lower end of the size distribution of firms, economies of scale in research seem to be present, while diseconomies are likely when firm size becomes large. It appears, thus, that we cannot deal with the question of economies of scale in R and D without considering the size of the firm within which the re- search is undertaken.1° These findings may be considered further by dealing with our primary rela- tionship in a different format. If both sides of equation 1 in table 1 are multi- plied by 2, the relationship may be expressed in the form: Y=a+bR+cR2 (1) where b=-4.671-0.000000128 $2, (2) a is a function of S and D, and c cquals 0.547. Thus ~=b+2cR. (3) Since b is inversely related to 8, is larger for smaller values of S. For given values of R, declines with the square of S. These results imply that for the pharmaceutical industry, the marginal productivity of professional research per- sonnel is inversely related to size of firm. Using the functional form of the relationship presented in (1), the elasticity of Y1 with respect to R was calculated to three points in the distribution of ~ 19 One additional point concerning our measures of technical change should be mentioned. Since these variables have been deflated by size of firm, they may be taken to represent the proportion of sales accounted for by new products. If larger firms sell relatively more standardized or older products than smaller firms, then we should expect the deflated ratios denoting technical change to be inversely related to firm size, and this fact would have little to do with the productivity of research establishments. It appears, however, that new products account for as large a proportion of total sales in large firms as in small ones. The proportion of total sales accounted for by products introduced within the preceding five years was computed for each year between 1955 and 1960 for each of the 57 firms included in the sample. The simple correlation coefficient (r) between the mean proportion for each firm and the corresponding measure of firm size was + 0.18. The simple correlation coefficients between deflated values of l'i and F2 and firm size were also computed. The coefficients were +0.14 and +0.08 respectively. Since it does not appear that larger firms sell, on the average, proportionately less new products than smaller firms, we should not expect this consideration to distort our findings. SYR 20 The elasticity was calculated as -h.. PAGENO="0534" 2084 COMPETITIVE PROBLEMS IN THE DRUG INDUSTRY Since this elasticity is defined as the percentage change in research output for a unit percentage change in R and D, holding S constant, it will describe the find- ings on economies of scale in research which are implicit in the equation results. Elasticities were computed at values of S equal to 1,000, 10,000, and 50,000 (phar- maceutical sales in thousands of dollars) ~ When S equaled 1,000, the calculated elasticity was 1.39. At S equal to 10,000 and 50,000, the corresponding figures were 0.61 and 0.54. While there are likely to be increasing returns to scale in R and D at low values of 5, decreasing returns seem to be the case when S becomes moderately large.n 23 SOME ALTERNATIVE HYPOTHESES Up to this point we have proceeded on the assumption that the direction of causation runs from research and development to `technical change, and have implied that the degree of diversification affects the rate of technical change. It is necessary, however, to consider some alternative hypotheses. It may be that successful new product introduction influences the extent of research activi- ties rather than or in addition to the opposite alternative. A rapid rate of tech- nical change may stimulate the firm to expand the scale of R and D facilities; accelerated technical change may result in higher profit rates which enable the firm to increase its investment in research and development. Similarly, it may he hypothesized that the rate of new product introduction is the vehicle by which diversification proceeds. If these alternative hypotheses are correct, we should expect that technical change at the start of the period would be better correlated with R and D and diversification at the end, than if the lead-lag relationship were reversed. Our original hypotheses would postulate that R and D and diversification should lead the variables denoting tecimical change. To investi- gates these questions, we shall consider our regression equations with alternative lead-lag structures. The empirical findings concerned with this matter are presented in `table 3. The research and development variables denote the number of professional research personnel employed in either 1955 or 1960, while the technical change and diversification variables span the period which is designated. The size vari- able remains the same as in the previous analyses. The results are striking. With a lead-lag structure consistent with our original hypotheses, all coefficients are significant at the 99% leveL With the opposite lag structure, the R2's are not significant, the parameters of diversification are smaller `than their standard errors and alternate in sign, and the coefficients for size, interaction, and the quadratic expression for research are not significant. There is, however, some positive linear association between technical change in 1955-1957 and research in 1960 when the variable Y2 is used, but this re- lationship may be relatively weak for it fails to appear in the regression equa- tions deallng with the entire six-year period. We may conclude, `therefore, that the apparent lead-lag structure is not inconsistent with our original hypotheses. 21 Corresponding values of R were determined by taking the mean value of that variable for firms in the neighborhood of the appropriate value of S. For S equal to 1,000, firms were included between S equal to 500 and 1,500; for S equal to 10,000, the corresponding figures were 5,000 and 15,000; while for 50,000, tile firms with Sbetween 40,000 and 60,000 were included. The three values of R w-hich were determined in this fashion were 13.1, 59.2, and 353.3, respectively. D was computed at its mean value throughout. 22 One qualification of the above results should be mentioned. It may be that large firms undertake more "basic research" than smaller firms. Although it is assumed that the returns from this activity are obtained through its advancement of the rate of new product technical change, the time horizon in this case may be substantially longer than that of "applied research and development" so that the six-year period under examination may not be long enough to measure accurately the returns. As a result, the level of research output for larger firms would be understated relative to smaller firms. It does appear that larger firms spend somewhat larger proportions of their R. & D. budgets for "basic research." For 1959, National Science Foundation data for industry group "Drugs end Medicines (SIC 283) disclose that firms with total em- ployment over 5,000 undertook 50% of total industry funds for the performance of R. & D. while these firms accounted for nearly 63% of industry "basic research." Tile difference between tile elasticities is, therefore, likely to be somewhat overstated. 23 One further implication of the functional form of (1) should be mentioned. Since b is negative and c is positive, the function, at given values of 5, originally declines, reaches a minimum, and then increases. This minimum is reached at values of R equal to ~ In addition; since b is an inverse function of 5, the values of R at which the minimum is reached increases with S. At values of S equal to 1,000, 10,000. and 50,000 the minimum is reached at values of R equal to 4.4, 16.0 and 298. Actual values of R will, in nearly all cases, be greater than these values. At the extreme upper tail of tile distribution of 5, however, the marginal productivity of research estimated l)y equation (1) in table 1 will be negative. PAGENO="0535" COMPETITIVE PROBLEMS IN THE DRUG INDUSTRY 2085 TABLE 3.-Regression and correlation analysis-Various lead-lag structures Intercept RD RD2 Size Ii D R2 1955 1955 1955 1955-1957 (1) Fi (1958-1960)_... (2) 1"i (1955-1957)_U bO.357 (0. 100) 0.125 (0. 076) b-3.360 (1. 025) 1900 -1.193 (0. 751) 1955 bO.351 (0. 083) 1960, 0.0861 (0. 0519) 1955 0.0000159 (0. 0000049) 0.00000649 (0. 00000485) b_0.0000000693 (0. 0000000212) 1960 -0.0000000172 (0. 0000000152) 1955 b--0.109 b031 (0. 032) 1958-1960 -0.0164 0.13 (0. 0205) 1955-1957 (3) F'2 (1958-1960) - - - - b 0. 594 (0. 122) b -3. 898 (1. 252) 1960 b 0. 465 (0. 102) 1960 b 0. 0000171 (0. 0000060) b -0. 0000000810 (0. 0000000259) 1960 ~-0. 118 b 0. 32 (0. 039) 1958-1960 (4) F'2 (1955-1957)~.. `0.311 (0. 140) b4343 (1. 387) 0.0053 (0. 0959) 0. 00000309 (0.00000895) 0. 0000000012 (0. 0000000280) 0.0131 0.18 (0. 0378) R and D is measured by the number of professional research personnel. Indicates statistical significance at the 99 percent level. Indicates statistical significance at the 95 percent level. THE ROLE OF SUPPORTING PERSONNEL At this point we shall examine the question of whether a large ratio of sup- porting personnel to professional staff substantially increases the efficiency of a research facility. The relevant variable is defined a's the mean ratio in 1955 and 1960, and the analysis is carried out by introducing it into our regression equations. The cofficients of the ratio of supporting to professional personnel are always negative and in no instances are they statistically significant.24 Thus, we may conclude that supporting personnel play only a minor role in determining the magnitude of research output. High ratios of auxiliary staff to professionals do not seem to increase the productivity of the professional researchers. These results corroborate our earlier observations that the number of professionals seems to be a better measure of research input than the total number of It and ID personnel. THE EFFECT OF GROWTH IN RESEARCH FACILITIES In this section, we examine the influence on technical change of rapid expan- sion of R and D activities. Some observers maintain that research efficiency is likely to be impaired if the rate of expansion is large.n To consider this position, we define two variables and introduce them separately into the regression equations. The first variable, which is designated G~, is the ratio of total research personnel in 1960 `to the corresponding value in 1955. The second variable G2, is the ratio of 1960 to 1955 levels of professional research personnel. The coefficients of G~ and G-~ are not statistically significant although they are negative and thereby do take the hypothesized sign.2° As a result, the hy- pothesis was restated and examined in a different format. It may be that rapid expansion of research facilities does not affect the rate of technical change continuously. Inefficiences may be created only when growth is pushed above some threshold rate. To investigate this position, three dummy variables are defined according to whether the research establishment showed no growth, moderate growth, or rapid growth, between 1955 and 1960, and the first two dummies are introduced into the regression equations.27 24 When the ratio of technicians to professionals is introduced into equations 1 through 4 of table 1, the estimated coefficients and standard errors are: (1) -0.181, 0.144; (2) -0.149, 0.220; (3) -0.086, 0.139; and (4) -0.161, 0.211. (In equations 2 and 4, the non-significant linear RD variable was omitted when these estimates were obtained.) ~ Mansfield has reported: "There are considerable costs involved in a very rapid ex- pansion of a firm's R and D department, the importance of which was stressed in inter- views with various executives." Edwin Mansfield, The Ewpenditures of the Firm on Research and Development (mimeographed), 4. 26 When G~, is introduced into our original regression equations, the estimated coefficients and standard errors are: -0.124, 0.126 in the case where technical change is measured by Fi, and -0.168, 0.185 when F2 is used. The corresponding estimates for G2 are: -0.0229, 0.0958 and -0.061, 0.153. In these calculations, RD is measured by the number of pro- fessional research personnel. ~ The dummy variables are defined on the following basis: If, during the period, the laboratory remained stable or declined in size, the firm is listed in the first category; If the facility increased in size but by less than 100%, the moderate growth grouping is designated; while if the research establishment at least doubled in size, it is considered to have experienced rapid growth. The three categories are defined in terms of increases in both total and professional research personnel. PAGENO="0536" 2086 COMPETITIVE PROBLEMS IN THE DRUG INDUSTRY An F test is used to deal with the question of whether the dummy variables significantly affect the rate of technical change. While the dummies based on increases in professional personnel do not lead to significant values of F even at the 50% level, those based on the expansion of total research personnel resu~1t in test values which are significant at the 75% level when F1 is used to measure technical change. In addition, the estimated regression coefficients founded on the latter set of dumniies take the expected sign and relative order of rnagnitude.IS These findings point out that any inefficiencies resulting from rapid expansion of research facilities are more likely to be associated with increases in the total size of the establishment rather than particularly related to increased hiring of professional personnel. It should be emphasized, however, that empirical support for this hypothesis exists only if we are willing to accept statistical findings as the 75% level of significance. As a result, this conclusion is highly tentative. CONCLUDING STATEMENTS In this paper, we have investigated the relationship between research and development and the rate of new product technical change in the pharmaceutical industry. From the empirical evidence, there appears to be a fairly sustained association between research input and new product output. Within this industry, research expenditures are not undertaken merely with the hope of some distant but unknown returns, but rather with the expectation that profitable gains will accrue within a reasonable period of time. Our analysis also provides some evidence that in the pharmaceutical industry, there are substantial diseconomies of scale in R and D which are associated with large firm size; and that these disadvantages are encountered even by mod- erately sized firms. One implication of this finding is that an actively enforced pro-competitive policy in this sector is not likely to dampen the rate of technical change and may well stimulate it. While little is known about the extent `to which this result is applicable to the economy at large, it does appear that there are grounds for considerable doubt as to the position that large firm size is always a necessary condition for rapid teelmical advance. [From The Journal of Business of the University of Chicago, Vol. XXXIX, No. 1, Part I, January 1966] THE DRUG INDUSTRY AND MEDICAL RESEARCH THE EcoNoMics OF THE KEFAUVER COMMITTEE INvE5TIGATI0N5* (By William S. Comanort) In the course of recent American politics, few confrontations have been more lively and more prolonged than that w-hich was waged between the pharmaceuti- cal industry and the Senate Subcommittee on Antitrust and Monopoly (The Ice- fauver Committee) .~ In December, 1059, the committee opened hearings on the ~ The dummy variables denoting "no growth" and "moderate growth" are introduced into equations (1) and (2) of table 1. When the dummies describe the rate of expansion in total research personnel, the estimated coefficients and standard errors are, for equation (1) : 0.371, 0.220 and 0.209, 0.184, and for equation (2) 0.399, 0.336 and 0.159, 0.297. The coefficient of the "no growth" dummy in equation (1) is statistically significant at the 95% level. Furthermore. the intercept, which encompasses the effect of the third clas- sification, is reduced from 0.422 to 0.199 by the introduction of the dummies into equation (1). *This article is drawn in part from my unpublished Ph.D. dissertation, "The Economics of Research and Development in the Pharmaceutical Industry" (Department of Economics, Harvard University, June 1963). I am grateful to J. W. Markham and R. B. Hefiebower, who acted as supervisors of the original study, and to Carl Kaysen for valuable counsel in the writing of this paper. I have also benefited from comments by Oswald H. Ganley. Mary Lee Ingbar, Lars Sandberg, and Harvey M. Sapolsky, and from the secretarial assistance of Anne Caneles. This study was undertaken with the support of the Science and Public Policy Program at Harvard. tlnstructor of economics, Harvard University. Currently on leave and associated with the Antitrust Division, U.S. Department of Justice. This article was written at Harvard and in no way does it reflect the views of the Department of Justice. 1 The pharmaceutical industry in the Congressional proceedings, and therefore in this paper, is limited to firms that produce ethical drugs, as opposed to proprietaries, and that distribute these products in dosage form. Pharmaceuticals, thus, are marketed and sold through the medical profession and require, for the most part, a written medical pre- scription. PAGENO="0537" COMPETITIVE PROBLEMS IN ~ THE DRUG INDUSTRY 2087 ii~dustry, ard 11,000 pages later, testimony was still being recorded. Now that the record has been closed and the dust from the contest `has been `allowed time to settle, it is useful to consider some of the questions raised in `the proceedings. This paper reviews the evidence on drug research. I A major point at issue between the industry and the committee concerned the medical value of industry research. Although forced to admit that expenditures on research were large,2 the coinniit'tee charged that little of social value came from industry laboratories. It maintained that nearly all the important new discoveries in recent years had been derived from research work performed out- side the industry and that commercial laboratories are concerned primarily with "molecule manipulations" or new drugs therapeutically quite similar to drugs already on the market.3 The committee implied that industry research was highly duplicate and that a large portion of this work could be eliminated without reducing very much the flow of important new drugs. In defense of its efforts, the industry argued that from industry laboratories come most of the new drugs that are extensively utilized in current medical practice.4 These opposing views on the nature and significance of the pharmaceutical research effort `collided in what may be termed the "battle of the lists." This contest was based on competing lists of new drugs prepared by the committee and the industry. One of the major points of `difference between the two is that the committee tended to concen'trate on drugs that embodied what it considered to be major therapeutic advancements at the time of introduction. As we would expect. the conunittee and i'ts lists emphasized the role played by foreign and non-industry laboratories.5 The `industry list, on the other hand. included new drugs that may not have embodied large steps forward but that are in frequent use and thereby seem to have the confidence of the country's physicians.6 A large majority of the drugs on `this list were dis~overed and developed within industry laboratories. While :tl~ere was some controversy over the origins of specific drugs. it seems likely that both sets of lists and their associated conclusions are essentially correct. To unravel the issues here, it is necessary to deal explicitly with the nature and functions of industry research. In the succeeding sections of this paper, we shall undertake this discussion in the light of the alternative pos'itions taken by the committee and the industry. Research and development in the pharmaceutical industry is carried on within a context of strong product rivalry, `and this factor has affected the type of work that is undertaken. Considerable attention is focused on the rapid introduction 2 In 1963. e.g., company-financed research and development totaled $282 million. This sum equaled nearly 9 per cent of total industry sales (Pharmaceutical Manufacturers As- sociation, "Pharmaceutical Industry Research and Development Activity, 1963-64" [Wash- ington, D.C., 1964 (mimeographed)], p. 1, and Pharmaceutical Manufacturers Association, "Ethical Pharmaceutical Sales, 1963" [Washington, D.C., 1964 (mimeographed)], p. 2). Although N.S.F. data on company-financed H ,& D as a percent of net sales in manufac- turing industries for 1963 is not yet available, 1961 data show that the highest percent- age for an industry other than drugs and medicines was 4.4 per cent (National Science Foundation, Research and Development in Industry, 1961 [February, 1964], p. 86). This position was based on statements such as the following, which appeared through- out the proceedings: "The question is what then is the goal of this admittedly large-scale laboratory effort of our industry? Partly to exploit and market those foreign and non- industrial advances and compounds that I have mentioned. Mostly, however, to modify the original drugs-just enough to get a patentable derivative, but not to change it enough to lose the original effect" (Testimony of Frederick H. Meyers, associate professor of pharmacology, University of California, in U.S. Senate, Administered Prices in the Drug Industry [hereinafter cited as "Administered Price Hearings"]. [Hearings before the Subcommittee on Antitrust and Monopoly (Washington: 86th Congress, 2d Session, 1960)] (Part XVIII, p. 10394). Pharmaceutical Manufacturers Association, Time Pharmaceutical Industry, Key Pacts at a alance (Washington, D.C., n.d.). The industry reported that of the ninety-four im- portant drugs developed since 1945 and in frequent use in 1960, forty-nine, or over half, were discovered by drug companies or by university scientists financed by drug companies. The original committee list appears in' Administered Price Hearings, Part XIX, pp. 10943-45, while the supporting comment is made in U.S. Senate Subcommittee on Antitrust and Monopoly, Report of the ytudy'of Administered Prices in the Drug Industry (Washington: 87th Congress, 1st Session, 1961), p. 115. 6 E.g., the original industry list included Achromycin as well as Terramycin, two anti- biotics that have similar therapeutic effects. This list appears in Administered Price Hearmngs, Part XIX, pp. 10840-54. PAGENO="0538" 2088 COMPETITIVE PROBLEMS IN THE DRUG INDUSTRY of new products, and, as a result, the number of introductions exceeded 3,800 in the decade ending in 1960. Of these, the largest number by far consisted of new combinations of existing drugs and of new dosage forms.8 The remainder, or 432 in total, were new chemical entities, and these absorbed the largest share of the research effort. It should be noted, however, that this category included salts anti other derivatives of known drugs as well as entirely new substances.9 The nature of industry research may be inferred to some extent from informa- tion on new products. As we might expect, there appears to be a large effort to invent around existing patents, and this has led to the introduction of a consider- able number of new drugs whose therapeutic effect is quite similar to that of products already on the market. We should note, however, that rarely do different chemical compounds have perfectly identical therapeutic properties and that small differences may be important for some patients. To a great extent, industry research activities build upon the scientific achieve- ments of the past. They are concerned, primarily, with the smaller, albeit less uncertain, steps forward. Although these activities may duplicate what has already been clone and also frequently may lead to new drugs that represent at best a very minor advance, it is still true that a number of important modifications and variations have been discovered within industry laboratories. The second major emphasis of industry research concerns the process by which new scientific know-ledge is translated into marketable products. Product rivalry is strong, and this has insured that new drugs will not lag far behind advances in scientific knowledge, no matter whether these advances originate in university, government, or foreign laboratories. This is an important area, because in most cases substantial research and development problems remain even after the original discovery has been made. Not only is there frequently a substantial gap between laboratory synthesis and large-scale manufacturing methods, but also considerable analysis and testing is required before a drug may be marketed. And these activities, which may well be more costly than basic research, are under- taken for the most part within industry laboratories. Whether the impetus for a new drug comes from competitive successes, from Jasic research carried on within the industry, or from new scientific knowledge arising from non-industry sources, the problems of synthesis and testing remain, and these absorb the major share of industry research expenditiures. Large quantities of the promising compound are normally required before efficient pro- duction techniques are available, and large sums are expended for this purpose. It has been estimated, in fact, that approximately one-half of total expenditures on research with the industry goes to support the work of synthesizing, purifying, modifying, and preparing suitable substances for subsequent physiological tests.'° Once this process is completed, the new substance is subjected to intensive bio- logical investigation. The first concern is to determine the therapeutic properties of the su,bstartce this encompasses the major purpose of the drug as well as side effects. In addition, it is necessary to test for toxicity levels and to determine the compound's potency in order to gain knowledge of appropriate dosages. For these purposes, the pharmaceutical industry used in its research facilities some nine million laboratory animals during l96l.~' There is, moreover, a great deal of uncertainty concerning the therapeutic properties of new substances. As would be expected, most of those tested in laboratories are found to be without sufficient promise to justify clinical testing on human beings. In 1958, for example, nearly 115,000 substances were subjected to biological tests by pharmaceutical firms, while only 1,900 were considered worth testing clinically.~ The Food and Drug Administration requires that manufac- turers undertake efficacy studies on new drugs and that the results of these tests Paul de Haen. "New Product Survey" (New York: privately printed by author. 1900). Althou~h new chemical substances are not involved in these cases, considerable atten- tion may still be required. E.g.. a good deal of work was done in an attempt to combine Glucosamine with certain antibiotics in the hope that this would increase the absorption of the drug into the blood stream (Administered Price Hearings, Part XVIII, p. 10257). During 1959. sixty-three new chemical entities were introduced. Of these, eleven were new salts of old products and twenty-three were derivatives of known drugs. The re- mainder of twenty-nine were entirely original products (ITS. Senate, Drug Industry Anti- trust Act [hereinafter cited as "Antitrust Act Hearings"} [Hearings before the Sub- committee on Antitrust and Monopoly (871i Congress, 1st Session, 1961) 1, Part II, p. 888). ~° Statement by Austin Smith. president of the Pharmaceutical Manufncturers Asso- ciation, in Administered Price Hearings, Part XIX, p. 10725. 1~ Ruported in New York Times, August 26, 1962. p. 60. 12 Administered Price Hearings, Part XIX, p. 10725. PAGENO="0539" COMPETITIVE PROBLEMS IN THE DRTJG INDUSTRY 2089 be submitted to the administration before a new chemical entity may be introduced. Clinical studies are normally carried out with the help of physicians and medi- cal institution:s throughout the country. New drugs are contracted out to clini- clans in medical schools or hospitals for controlled testing. While there may be a sizeable liaison group among the pharmaceutical research staff, the actual test- ing is done, in most cases, by non-industry personnel. In 195~, only one major company supported directly the maintenance of a clinical-research group having hospital facilities; this was the Lilly Laboratory for Clinical Research located within the Indianapolis City Hospital.13 Even when the testing was performed outside of industry facilities, however, the pharmaceutical firm concerned nor- mally supplied the compound that was to be tested. While it is difficult to determine directly the cost of clinical trials to pharma- ceutical firms, we can observe the division of industry research expenditures spent within and outside of the companies as some indication of the relevant magnitude. In this regard, less than 10 per cent of total R & D expenditures was in the latter category.'4 To this should be added, however, the cost of preparing the compounds used in the trials. Although the laboratory and clinical-research activities described above have a highly applied and developmental character, still they are necessary before medical benefits can be realized from new scientific knowledge. With regard to the major share of industry-research activities, these are best viewed as a comple- ment rather than as an alternative to those undertaken outside the industry. III To evaluate industry research activities, it is necessary to place them within the context of the total medical research effort. In this regard it should be stressed that the industry provides only about 30 per cent of total medical- research outlays within the United States,15 and the proportion would fall still further if foreign research expenditures were included in the total. Given the relative size and essential complementarity of industry and non-industry research, it follows that the two cannot adequately be evaluated by the same set of standards. Despite the fact that the same word is used to describe both sets of activities, they are essentially different. One of the major problems of the investigatory proceedings and the joust be- tween the committee and the industry concerned the failure of either side to recognize this difference. The committee criticized the industry on the apparent failure of industry scientists to provide many dramatic new breakthroughs on which major therapeutic advances are based and also on the lack of important new drugs in recent years. Whatever the validity of these charges, they are not founded on a realistic appraisal of the role and function of industry research. At the same time, however, the industry's apparent claim that its research is respon- sible in large measure for the great advances in medical science that have been achieved in recent years is equally inappropriate. Both of these positions imply that industry and non-industry research can be evaluated on a single set of standards, and this neglects the considerable differences in purpose and approach between the two areas of work. A valid judgment of industry research and development must necessarily be founded upon the nature of the complernentarities present within the larger boundaries of medical and health-related research. These complementarities are important, and it does appear that the differences that exist between the two areas of research may well constitute a necessary and desirable division of labor. The peculiar attributes of each probably create for it a comparative advantage for the specific activities which it, in fact, undertakes. While it is difficult to be definite in these matters, it does seem likely that basic or funda- mental research is better carried on in a university or government laboratory 13 American Foundation, Medical Research, A Mid-Century ~S'urvey (2 vols.; Boston: Little. Brown & Co., 1955), I, p. 592. " In 1958 the amount spent outside of Industry facilities was $13.4 million out of a total of $170.0 million, while in 1959 the comparable figures were $15.3 million and $190.0 million (Administered Price Hearings, Part XIX, p. 10724). 15 In fiscal year 1960, the pharmaceutical industry allocated $215 million to research and development activities out of a total national expenditure of $715 million. Govern- silent provided $400 million, while other private support equaled $100 million (Antitrust Act IIear~ngs, Part III, p. 1705). PAGENO="0540" 2090 COMPETITIVE PROBLEMS IN THE DRUG INDUSTRY than in a commercial facility. Within a profit-motivated firm, there is a latent conflict between an atmosphere of freedom, which is necessary for good scientific endeavor, and the necessity of direction from those who are responsible for the conduct of the firm.16 In addition, it appears that those firms which have had the greatest success in basic research have moved furthest toward creating a uni- versity environment within their facilities.17 While additional corroborating fac- tors could be mentioned, it is sufficient to remark that we should not be surprised if the major advancements in medical science are achieved in laboratories out- side the pharmaceutical industry. At the same time, moreover, there are a number of reasons w-hy highly applied research and development activities are best carried on within the confines of the industry. Not only do these tasks assume a highly routine character which tends, in a large number of cases, to make them uninteresting to university scientists, but also the relevant procedures are sufficiently well understood so that they may be reasonably well directed. An atmosphere of completely free inquiry is not crucial where both the goals and the procedures are relatively w-ell defined. These considerations, moreover, may explain the fact that although the pharma- ceutical industry provided approximately 30 per cent of the total funds for medical research in 1960, it utilized only 12 percent of the total number of persons doing medical research with the Ph.D. degree and only 5 per cent of the total number doing medical research with the M.D. degree.18 Even more important, however, are the economies that are likely to result from combining applied research with production in the same organization. Within a technical industry, the road from laboratory to factory is not level or direct, and large gains may result from institutional arrangements that provide the smallest possible barriers to the flow of information. To insure a free and unre- stricted flow- it is probably necessary that the latter stages of research and development, as well as production, are undertaken w-ithin the firm. It is interesting, in this regard. to note the conclusions of a recent study that contrasted the American and Soviet pharmaceutical industries. Unlike the United States, research and production are carried on by completely separate agencies in the Soviet Union. The authors report that as a result of this arrangement, `there are difficulties of communication and coordination between [research] institute and factory which constitute a major bottleneck in getting pharma- ceuticals into production." 19 While the divorcement of research from production represents merely one additional factor out of many that distinguish the Soviet industrial structure from that of the United States, still the apparent results from this divorcement do seem to point to a major set of problems. The division of labor that exists betw-een industry emphasis on product de- velopment and non-industry emphasis on the more fundamental areas of re- search provides a system that is rational in principle although we know little of whether the optimal share is more or less than the 30 per cent presently occupied by the industry. If, however, our concern is with the use and possible misuse of such scarce and limiting factors as highly trained personnel rather than merely with dollars, the quantitative pro~ilems appear in a different perspective. A large drug-industry effort does not divert, to a substantial degree, scarce resources from university and government laboratories to its own facilities but, rather, uses, for the most part, a different and non-substitutable class of personnel. Thus, the opportunity cost of industry research in terms of non-industry research is likely to be relatively low. When evaluated on its own terms, it seems clear that the industry accom- plishes an important research task in a generally effective manner. It has been responsible for a large number of pharmacological modifications and improve- ments that have been introduced as well as for an acceleration in the process of testing and developing new- drugs, and both of these gains provide a high degree 16 This "conflict" is discussed in John Jewkes, David Sawers, and Richard Stillerman, The Sources of Invention (New York: St. Martin's Press, 1958), chaps. vi, vii. ~ See, e.g.. the statement by James B. Fisk. "Basic Research in Industrial Laboratories." in Dael Wolfie (ed), Symposium on Basic Research (Washington, D.C. : American Asso- ciation for the Advancement of Science, 1959), pp. 159-67. 18 National Institutes of Health, Manpower for Medical Research-Reqnireineists and Resources, 1965-1970. (A Report for the Committee on Appropriations [Washington: U.S. House of Representatives, February. 19621), p. 24. ~9 Raymond A. Bauer and Mark G. Field, "Ironic Contrast: U.S. and U.S.S.R. Drug Industries," Harvard Business Review (September-October, 1962), p. 93. PAGENO="0541" COMPETITIVE PROBLEMS IN THE DRUG INDUSTRY 2091 of social value.20 It should be recognized, however, that these gains are on a very different scale from those related directly to significant advancements in medical science and to dramatic improvements in public health. Iv During the investigations, the Kefauver Committee proposed a major revision in economic policy toward the drug industry. This proposal, however, was re- jected by the parent Committee on the Judiciary and was not included in the 19G2 Drug Act. The change would have instituted compulsory licensing at reason- able royalty rates three years after a patent had been granted. It is useful here to note the implications of this change in the light of the analysis presented above. Although this proposal was intended to produce more competitive levels of prices and profits, the industry maintained, that it would also lead to much re- duced research expenditures. This position seems to be generally correct. The primary motive for large research efforts in this industry has been the drive to achieve effective product differentiation.21 With compulsory licensing, there would be a sharp decline in the extent of differentiation based currently on chemical differences among products. Nevertheless, it is not at all certain that compulsory licensing would signifi- cantly lower the rate of introduction of the most important new products. The largest proportion of these come originally from non-industry laboratories. It is true that pharmaceutical companies would have less incentive to undertake projects of a long-term nature, such as many of those in basic research, because of the diminished prospect for large gains over a prolonged period of time. These firms account, however, for only a minor share of the work which currently is done, and they also appear to lack a comparative advantage in pursuing basic research in an extensive manner. Even with compulsory licensing, research and development would still com- prise an important element of industry behavior. There would still be gains from achieving product differentiation. New drugs that embody a large element of therapeutic improvement are also likely to provide a high degree of product differentiation. Having developed a differentiated product, not only would firms benefit substantially from a head start in promotion and selling, but also this advantage would last for the initial, prelicense period of patent protection, dur- ing which time monopoly gains could still be attained. In addition, patent royal- ties might become an important element of the rewards resulting from successful research. Compulsory licensing appears on balance to be a useful and desirable policy to adopt. While the magnitude of research would decline, it does seem probable that projects of limited medical value and of lowest industry comparative advan- tage would be eliminated first. There are substantial social gains to be derived from industry research, but the marginal social productivity of research may well decline rapidly after a certain level has been reached. We should be wary of believing that much is to be gained from ever higher levels of research and development, and it is quite possible that the present effort may exceed that required to fulfill the major research functions and responsibilities of the industry. So long as the decline in research expenditures was not of over- whelming proportions, it may well be a small "price" to pay for a more competi- tive determination of pharmaceutical prices and profits.21 ~ This is in contrast to the Soviet pharmaceutical industry. Bauer and Field state that "the testimony of the well-informed Soviet sources is that the separation of research from production tends to produce substantial delays in the availability of drugs to physi- cian and patient" (Bauer and Field, op. cit., p. 94). At the same time, however, excessively rapid process leads to new drugs that have not been adequately tested, and this charge has been laid on the doorsteps of the American pharmaceutical industry. This point was com- mented on by a medical educator who noted sharply: "There is no short cut from chemical laboratory to clinic, except one that passes too close to the morgue" (Administered Price hearings, Part XVIII, p. 10418). 21 The role and function of product differentiation and its relation to research and de- velopment activities is discussed in William S. Comanor, "Research and Competitive Product Differentiation in the Pharmaceutical Industry in the United `States," Economica (November, 1964), pp. 372-84. Smaller research facilities may also lead to increased efficiency in pharmaceutical re- search and development. When research output is measured in terms of private rather than social gains, there appear to be substantial diseconomies of scale in research within the larger firms in this industry. See William S. Comanor, "Research and Technical Change in the Pharmaceutical Industry," Revieiv of Economics and Statistics (May, 1965), pp. 182- PAGENO="0542" 2092 COMPETITIVE PROBLEMS IN THE DRUG INDUSTRY [From The Review of Economics and Statistics, Volume XLIX, November 1967, Number 4, PP. 423-440] ADVERTISING MARKET STRTTCTURE AND PERFORMANCE (By William S. Comanor and Thomas A. Wilson*) This paper presents an empirical analysis of the role of advertising in con- sumer goods industries. The primary finding is that advertising has a statis- tically significant and quantitatively important impact upon profit rates which provide a measure of market performance as well as indicate the existence of market power. This result is robust, and the estimated multivariate equations account for half of the inter-industry Variance of profit rates. This finding has implications which are precisely the opposite of the conclu- sions reached by Telser in a recent important article.1 This contradiction is a reflection primarily of differences in the conceptual and statistical approaches adopted rather than differences in data or sample, for with minor exceptions, we used the same set of industries, and drew upon the same basic data for advertising outlays. We shall therefore proceed as follows. First, we shall describe the conceptual framework used. Then we shall examine the relationships which are likely to exist between product differentiation, advertising and entry barriers. Finally, w-e shall present the empirical results which are the core of this paper. FRAMEWORK OF ANALYSIS The analytical approach is to examine the joint effect of various dimensions of market structure upon profit rates. Not only do profit rates provide some indication of market performance in terms of the normal criteria of allocative efficiency, but also high returns signal the possible existence of market power.2 If exercise in the direction of profit maximization, market power should lead to rates of return which exceed those in competitive industries that are comparable in terms of risk and growth of demand. In this framework, concentration is simply one dimension of market structure and is not of itself a measure of monopoly or market power. Another major dimension is the height of entry barriers, which is determined in part by tech- nical factors such as the extent of production economies of scale relative to the size of the market, the absolute amount of capital required to operate a plant of minimum efficient scale, and other absolute production cost disadvantages of new entrants. Product differentiation, a third major dimension of market structure, plays a dual role. Not only does it directly influence the character of competition among established firms, but it also raises the height of entry barriers.3 In this study, however, w-e do not deal directly with product differentiation, hut focus instead upon advertising expenditures, which are both a sympton and a source of differentiation. Not only are advertising budgets influenced by product and market characteristics, but also they depend on the policies pursued by indi- vidual firms. In addition, past advertising outlays appear to be important deter- minant of the extent of product differentiation. Differences in advertising, therefore, reflect both structural and behavioral differences between industries. On these grounds, the empirical analysis which follows takes the form of multi-variate regression equations which explain the inter-industry variation in profit rates as a function of different combinations of the following variables: Seller concentration, The rate of growth of demand, *A preliminary version of this paper was presented at the December 1965 meetings of time Econometric Society. The authors thank Frank Edwards for his discussion of the paper at that time, and William Hughes and Lester Taylor for their comments on a sub- sequent draft. Special thanks are due Richard Caves and R. B. Heflebower, both of whom provided us with extensive and constructively critical comments. The authors are grateful for financial assistance provided by the small grant program of the Ford Foundation and a grant from the Alfred P. Sloan Foundation. 1 Lester Telser, "Advertising and Competition," Journal of Poltical Economy (Dec. 1964). Low or average profit rates (10 not necessarily indicate that market power is absent. Firms may become lax in minimizing costs when the discipline of competition is weak. For a discussion of such behavior, see Carl Kaysen. U.S. vs. United Shoe Machinery Cor- poration (Cambridge: Harvard University Press, 1956) 114-116. Joe Bain, Barriers to New Competition (Cambridge: Harvard University Press, 1956) 21, PAGENO="0543" COMPETITIVE PROBLEMS IN THE DRUG INDUSTRY 2093 Economies of scale in production in relation to the size of the market, Absolute capital requirements for a plant of minimum efficient scale,4 and Advertising. The specific variables used, the alternative functional forms, and other specifica- tions of the estimated equations are des'cribe~ below. The conceptual relationship between advertising, product differentiation, and the height of entry harriers is discussed in the next two sections. Before proceeding, however, it is useful to contrast the framework adopted here with that used by Telser. One of his major empirical findings is that the simple correlation between advertising outlays as a percentage of sales and the level of seller concentration is statistically insignificant. In each of the years studied, he finds that thi.~ coefficient is about 0.16, and from this, concludes that "There is little empirical support for an inverse association between advertising and competition." This approach raises the problem of whether concentration ratios are an adequate measure of the extent of competition. Telser justifies their use by stating that "Concentration of sales among the four leading firms is a widely accepted measure of monopoly." 6 While this statement is unfortunately correct, it ignores the fact that the concentration ratio measures only one dimension of market structure, and is therefore an inadequate indicator of market power, which depends on additional structural variables as well as on established be- havior patterns. The significance of advertising expenditures depends on whether they represent an additional factor affecting the achievement of market power. The weak correlation between concentration and advertising simply indicates that these are independent rather `than collinear variables. ADVERTISING AND PRODUCT DIFFERENTIATION The relationship between advertising outlays and product differentiation is important for an evaluation of the competitive effects of advertising because the former reflects the policies adopted by individual firms, while the latter is a dimension of market structure. The degree of product differentiation in a market is measured by the cross elasticities of demand and supply which exist among competing products. Low cross elasticities of demand between these products indicate that buyers prefer the products or brands of particular sellers and w-ill not switch in significant num!bers in response to small differences in price. Low cross elasticities of supply, on the other hand, signify that firms are unable to imitate the products of their rivals sufficiently well to eliminate these consumer preferences. While cross elasticities between the products of existing producers affect the character of the rivalry which exists between them, cross elasticities between the products of established firms and potential entrants influence the height of entry barriers posed by product differentiation.7 Product differentiation reflects two sets of factors: the basic characteristics of products within the market, and the present and past policies of established firms with respect to advertising, product design, servicing, and distribution. On the demand side, products are more likely to be differentiable when buyers are relatively uninformed about the relative merits of existing products. This is particularly important for differentiation achieved via advertising. On the supply side, differentiation is more likely where the products of rivals cannot easily be imitated and where new entrants have difficulties in producing products which are sirnular to those sold by successfully established firms. In producer goods industries, successful imitation requires investment in product design and adequate service facilities. In consumer goods industries, successful imita- tion may require investment in advertising as well. No attempt was made to measure any other absolute cost disadvantages of new entrants. Bain found that only in those industries in which established firms controlled scarce natural resources were these important. Bain, op. cit., 155-150. In addition, no attempt was made to measure risk. The sources of the data and various technical ad- justments are described in the appendix. Telser, op. cit., 544 and 55S. °Ibid., 542. It is important to distinguish product differentiation from product variety. The steel industry, for example, aroduces a great variety of products which are sold to knowledgeable buyers. but product differentiation is minimal. In contrast, the cigarette industry offem's a smaller variety of nroducts, but product differentiation-based largely on extensive advertising-is great. Bain, op. cit., 127-129. PAGENO="0544" 2094 COMPETITIVE PROBLEMS IN THE DRUG INDUSTRY It is noteworthy that Bath, in his authoritative examination of product differentiation in 20 manufacturing industries, found advertising to be the most important source of product differentiation in the consumer good industries in his sample. Distribution policies are also important where forward integration in prevalent, while customer services and product design play contributing but relatively minor roles. For typical consumer goods industries, then, a persistently high level of advertising expenditures can be viewed in two ways: a) If firms behave reasonably, high levels of advertising indicate that the product is differentiable. In this sense, advertising is a symptom of differentiation. b) The high level of advertising is itself an important determinant of the level of differentiation which is realized by established firms vis-~-vis potential entrants. In this sense, advertising is a source of product differentiation. Provided that firms act reasonably, observed advertising expenditures provide a useful measure of the extent of product differentiation. We write reasonably rather than rationally since, in an oligopolistic market, rational policies are not unambiguous. What is rational policy for the group acting in concert is not rational policy for the individual firm expecting to gain a march on its rivals. It is quite possible, moreover, that rivalry via advertising among established firms is carried to the point of diminishing returns in terms of group profit rates. However, even in this case, the result of extensive advertising rivalry may be to permit the achievement of higher future profits for the group by raising entry barriers. THE EFFECT OF ADVERTISING ON ENTRY BARRIERS Although advertising is only one source of product differentiation, it is espe- cially important in a number of consumer goods industries where it has a strong direct impact on entry barriers.8 In these industries, new entrants generally are forced to sell at a price below the established brands or else incur heavy selling costs. This explains the phenomenon of unbranded products selling at prices substantially below those of highly advertised products even where there is little "real" difference between them. On this account, established firms can set prices above existing cost levels, including advertising and other selling expenses, without inducing entry. Product differentiation via advertising affects entry barriers in three ways, each of which is analogous to the other determinants of overall entry barriers. First, high prevailing levels of advertising create additional costs for new entrants which exist at all levels of output. Because of buyer inertia and loyalty, more advertising messages per prospective customer must be supplied to induce brand switching as compared with repeat buying. Since the market which pros- pective entrants must penetrate is made up largely of consumers who have purchased existing products, advertising costs per customer for new entrants will be higher than those of existing firms who are maintaining existing market positions. Moreover, the costs of penetration are likely to increase as output expands and customers more inert or loyal need to be reached.° This effect of advertising creates an absolute cost advantage for established producers, since they need not incur penetration costs. In addition, the effect of advertising on firm revenues is subject to economies of scale which result from the increasing effectiveness of advertising messages per unit of output as well as from decreasing costs for each advertising message purchased. The first source of economies will exist whenever the effect of advertising on consumer decisions is sufficiently important that a threshold level of advertising is required for a firm to stay in the market and maintain its current market share. In such a situation, larger firms have the advantage of being able to spread this cost over more units of output and thereby spend less per unit sold. This advantage creates economies of scale at the firm level, since an established firm does not have to spend twice as much on advertising to 8lbid 114-143. These penetration costs depend on past as well as current advertising outlays by established firms. The importance of past outlays is examined by Kristian S. Palda who concludes that "distributed lag models both give a better fit to the Plnkham data and forecast better than he models which do not incorporate lagged effects." The Measurement of Cumulative Advertising Effects (Prentice-Hall, 1964), 94. PAGENO="0545" COMPETITIVE PROBLEMS IN THE DRUG INDUSTRY 2095 maintain a market share which is twice that of a rival. Higher output levels are associated with lower unit costs.'° As a result, smaller firms, including most entrants, are placed at a strong disadvantage." Economies of scale in advertising also result when the cost per advertising message declines as the number of messages supplied increases. An increased use of some forms of advertising leads to a lower most per message, and available evidence suggests that this is very important for advertising on national television and in national magazines.'2 If advertising in a particular industry is characterized by economies of scale for either of these reasons, an entrant will suffer an additional cost disadvan- tage if he enters at a relatively small scale. If he enters at a scale sufficient to realize available economies of scale in advertising, however, his actions are likely to influence the price or advertising policies of the established firms. The possible reactions of established firms increase the costs and risks of entry. Finally, if economies of scale exist either in production or in advertising, the need to obtain funds for advertising will give rise to capital requirements over and above those needed for physical plant~ and equipment. Furthermore, this investment in market penetration will involve a particularly risky use of funds since it does not generally create tangible assets which can be resold in the event of failure. The required rate of return on such capital will therefore be high. These various effects are illustrated diagramatically in figure 1. Curve APC represents average production costs for established and prospective firms, and l\1ESP is minimum efficient scale in production. Curve AAC describes average advertising costs for existing firms as well as for new entrants after they have become established. It denotes unit advertising outlays which are required in order to maintain a firm's market position and to preserve a given volume of sales once it has been established. This will depend on both the total level of advertising outlays and their distribution among established firms, and therefore, it describes prospective advertising costs for entrants only if existing firms do not react to any loss of market share. To the extent that they do respond, required advertising outlays will be higher. Curve ATC, the vertical sum of these two curves, represents average total costs for established firms." MES denotes the minimum efficient scale in both production and advertising for an established firm with a given market share. In addition, curve AMPC describes average market penetration costs for new entrants. Penetration costs represent an investment in establishing a market position and therefore depend on the opportunity cost of capital as well as on total penetration expenditures.'5 10 In the automobile industry, for example, the two smaller firms during the 1950's were forced to spend more than twice as much on advertising per car sold as did either Ford or General Motors. Between 1954 and 1957 Studebaker and American Motors spent annually on national advertising approximately $64.04 and $57.89, respectively, per automobile sold while G.M. spent $26.56 per unit and Ford spent $27.22 per unit. Chrysler was in an intermediate position, spending $47.76 per unit. Leonard W. Weiss, Economics and American Industry (New York: John Wiley and Sons, 1961) 342. 11 This result occurs within the relevant market. When a firm in a regionally segmented market expands its national market share by moving into new geographic areas, unit advertising costs do not decline. "The extent of discounts given to large advertisers is documented In Federal Trade Com- mission vs. The Procter 1 Gamble Company, Brief for the Federal Trade Commission in the Supreme Court of the United States (Dec. 1966), 12-13. 13 For simplicity, we assume here that advertising constitutes the only form of selling expense. 14 Penetration costs include extra advertising outlays which are required for entry. These outlays will represent total penetration costs if the price charged by the entrant is the same as that set by established producers. If the entrant is forced to set a price below that of existing firms, there are additional penetration costs which equal the price differential times the amount of output sold by the entrant at the lower price. 81-280-68-pt. 5-35 PAGENO="0546" 2096 COMPETITIVE PROBLEMS IN THE DRUG INDUSTRY FIGURE 1. - ADVERTISING AND ENTRY BARRIERS Unit Costs 0 This schedule therefore denotes the required rate of return on capital invested in market penetration times the total expenditure required to establish a given volume of sales, all divided by the number of units sold. The figure illustrates the case where average penetration costs rise throughout the relevant range of output. This assumes that the growing difficulty involved in winning over customers with stronger preferences for the products of established firms, reinforced by rising required rates of return as the absolute amount of capital required for penetration increases with the scale of entry, is not fully offset by economies of scale in advertising or by bandwagon effects for the new entrant's products. Curve ACN represents average costs, including peneration costs, for new en- trants, and MESN is the most efficient scale for entry if the reactions of estab- lished producers are neglected. From this, it follows that EP is the miniinvm price at which entry will occur. If MESN is a negligible fraction of the market, EP is the entry-inducing price. If, however, MESN is a significant fraction of the market, entry is unlikely to occur even at price EP because the entrant will expect established producers to contest the encroachment of their market position through an increase in advertising outlays or by a reduction in price. The gal) between EP and ATC' represents, therefore, the minimum price-cost margin which may induce entry. This figure demonstrates, moreover, that the interaction between rising pene- tration costs and economies of scale at the firm level is important even if no allowance is made for the reactions of existing producers. If economies of scale in both production and in advertising w-ere absent, the relevant price-cost margin would be simply M, w-hich is less than EP-ATC'. VARIABLES USED IN THE MULTIVARIATE ANALYSIS In this section, we brefly define and discuss the rationale for the selection of each of the specific measures. Profit Rates The profit rate variable used is profits after taxes as a percentage of stock- holders' equity,15 averaged within each industry for firms with assets exceeding 15 This profit rate variable was used originally in Joe S. Bain, "Relation of Profit Rates to Industry Concentration: American Manufacturing, 1936-4940," Quarterly Journal of Economics (Aug. 1951), 296-297, and Bain, Barriers to New Competition, 192. - .- ACN ATC MESP M~SN MES Output PAGENO="0547" COMPETITIVE PROBLEMS IN THE DRUG INDUSTRY 2097 $500,000. This procedure avoids the difficulty, noted by Stigler, of profit with- drawals in the form of executive salaries in small and closely held corporations.'0 Profit rates are also averaged for the period 1954-1907, which covers a complete business cycle. Although the profit rate on stockholders' equity is viewed as a more appropri- ate variable than the rate of return (including interest) on total assets,'7 we examined whether the empirical results would be sensitive to this decision. The simple correlation between the profit rate on stockholders' equity and the rate of return on total assets is 0.93. In addition, the correlation coefficients between each of these variables and the 1954 four-firm concentration ratio'8 are, respectively, 0.30 and 0.33. These results suggest that our empirical findings are unlikely to be. sensitive to the choice of a specific profit rate valuable. Advertising In light of the discussion in the preceding section, it is useful to examine the absolute volume of advertising expenditures by existing firms as well as~ the advertising-sales ratio. The latter variable probably provides a good indication of the absolute cost disadvantage of the new entrant at small scales of entry, but is likely to be a less accurate index of the economies of scale and absolute capital requirements effects of advertising. We have, therefore, calculated two measures of advertising intensity: adver- tising outlays per dollar of sales for firms with assets greater than ~50O,OOO, and average advertising expenditures per firm among firms which account for 50 per cent of industry output.'9 Both advertising variables are averages for the years 1954 through 1957. Seller Concentration Concentration is a sufficiently promin~nt variable in the literature to warrant. introducing it in three alternative formulations. First, a trichotomous classifica- Lion based on Kaysen and Turner's classification of market groups is used.2° Second, the average four-firm concentration ratios published by Stigler are intro- duced.2' Finally, a dichotomous classification is constructed on the basis of Bain's finding that a critical point is reached when the eight-firm concentration ratiG exceeds 70 per cent.22 Economies of Scale in ~Production Economies of scale in production presumably exist primarily at the level of the plant rather than the firm. In the absence of better estimates for most of the industries in the sample, a measure is derived from the size distribution of plants within the relevant industries. Since cost minimization is an element of profit maximization, large multi-plant firms should operate plants which are sufficiently large to realize available scale economies. Where demand is not a limiting factor, moreover, competition among firms should lead directly to plants. which equal or exceed minimum efficient scale.2' At the same time, however, small plants may exist. These may have been built in an earlier period, before demand had expanded or a technology which required large scale had been developed, or `~ George J. Stigler, Capital and Rates of Return in Manufacturing Industries (Princeton University Press) 1.25-127. 17 This is because firms presumably maximize profits, rather than the sum of profits plus interest payments. The rate of return on stockholders' equity will therefore be a. more sensitive indicator of the extent of freedom from competitive constraints. 18 In two cases out of 41, our industry classifications differed from those presented by Stigler. In both of these cases, "Screens and Venetian Blinds," and "Radio, T.V., and Phonograph," it appeared that Stigler had combined these industries with smaller, mis- cellaneous industries. In these calculations, therefore, we used data for the more aggregated industry to stand for its major component. `~ The procedure used was to select successive asset size classes of firms until 50 percent or more of industry sales was covered. The proportion of sales in the boundary size class required to reach this degree of coverage was used to determine the amount of advertising and the number of firms from that size class included in both the numerator and denomina- tor of the measure of advertising per firm. For some industries, the largest size class accounted for more than 50 per cent of sales. In such cases, the measure is simply advertising per firm in the largest size class. 20 Carl Kaysen and Donald F. Turner, Antitrust Policy, 27. 21 Stigler, op. cit., 214-215. 22 Bain, "Relation of Profit Rates to Industry ConcentratIon," 314. 23 Since the bulk of the evidence suggests that cost curves in manufacturing are "L shaped" rather than "U shaped," plants which exceed minimal efficient scale will typically be efficient plants. See J. Johnston, $tatistical Cost Analysis (New York: McGra,v-}Tfl1, 1960) 44-168. PAGENO="0548" 2098 COMPETITIVE PROBLEMS IN THE DRUG INDUSTRY they may result from the entry of small firms. They may also exist in pockets of the market which are geographically segmented or may specialize in narrow product lines which are not representative of the industry generally. It is there- fore important to select a measure which is insensitive to the entry of single- plant firms of sub-optimal scale. The measure used is based on average plant size among the largest plants accounting for 50 per cent of industry output. This average plant size is divided by total output in the relevant market to obtain the scale economies variable used in the regression analyses.21 A test of the reliability of this variable can be made by comparing minimum efficient scale as a percentage of industry output with Bain's estimates. Not only did Bain concentrate on a smaller number of industries, but also he used varied forms of information. Therefore, his estimates can be considered a benchmark against which to appraise various methods of estimating the extent of scale eco- nomies. We examine both the method described above and an alternative method, the Survival Technique, as used by Weiss.25 Of the 20 industries examined by Baiu, data on the size distribution of plants are available for 19. Across those industries, the correlation coefficient between Bain's estimates and those derived from the method proposed above is 0.89.26 Estimates b~sed on the Survival Technique are available for 13 of the industries~-~--~. studied :~y Bain, and the correlation between these estimates and those presented ~ by Bain is 0.66. Whan the comparison is limited to these same 13 industries, the correlation coefficient between estimates derived from the size distribution of establishments and those published by Bain is 0.86. The method propose'd above is more consistent with Bain's estimates than those computed from the Survival Technique,~ and it will therefore be used in the succeeding analysis.tm Absolute Capital Requirements This amount of capital required for entry at the scale of a single efficient plant is based upon the above estimates of economies of scale. The average output level of plants at estimated minimum efficient scale is multiplied by the ratio of total assets to gross sales for the industry.25 The Rate of Growth of Demand The rate of growth of demand is measured by the rate of growth of sales be- tween 1947 and 1957.~° A period of this length was chosen iu order to emphasize the long-run effects of the growth of demand, and the terminal years selected were both years of nearly full employment for the economy as a whole. Composite Variables Representing Technical Entry Barriers and Advertising Industries were classified into three groups on the basis of the two variables which measure technical entryl barriers (economies of scale and absolute capital requirements). Dummy variables identifying industries with high and moderate technical entry harriers were used in some regression models.3' 21 An alternative measure based on average plant size among the largest plants ac- counting for 70 per cent of output was also constructed. This variable was highly corre- lated with the variable used. 25 See Leonard Weiss, "The Survival Technique and the Extent of Suboptimal Capacity," Journal of Political Economy, (June 1964), 246-261. ~° In these calculations, Bain's estimates of minimum efficient scale for the Steel In- dustry refer to "Steel Works and Rolling Mills," while for the Copper Industry, they refer to `Primary Copper." 27 In addition, the survivor technique has the great disadvantage of frequently yielding indeterminate results. See Weiss, op. cit., 258-259. 28 Ideally, some account should be taken of the slope of the average cost function below minimal efficient scale. The variable used in this paper, in contrast to the estimates de- veloped by Bain, ignores this aspect of economies of scale. `~ This measure is likely on the average to understate capital requirements. The book value of total assets will normally be less than their replacement cost, as a result of in- flation in preceding years. In addition, a new firm is likely to have higher input costs while it is learning the production and distribution techniques required in the market. tm The ideal measure, of course, would be the rate at which the demand curve shifted over time. The rate of growth of sales is an exact measure only if the price elasticity of demand is unity or if prices did not change over the period. ~` High technical entry barriers are assumed to exist if either the scale of an efficient plant exceeded ten per cent of the market or the capital required for efficient entry equalled at least $50 million. In addition, if scale economies fell between six and 9.9 per cent and if capital requirements amounted to more than $25 million, the same classification is designated. Low technical barriers exist when either scale economies amounted to less than three per cent and capital requirements are less than $25 million or when the former was less than six per cent but the latter sum was under $10 million. Moderate technical barriers are assumed in all other cases. PAGENO="0549" COMPETITIVE PROBLEMS IN THE DRUG INDUSTRY 2099 Similarly, industries were classified into three groups on the basis of the two advertising variables. Dummy variables identifying industries with high and mod- erate advertising expenditures were also used as an alternative measure of ad- vertising intensity in some equations.32 Local Market Dummy T7ariable In some of the regression equations, a dummy variable was introduced to iden- tify the `three local market industries. THE SAMPLE OF thDIJSTRIES For the reasons presented above, the empirical analysis is confined to con- sumer goods industries.33 Of the 41 industries included in the analysis, 29 produced non-durable consumer goods, and the remaining 12 produced consumer durables. In size, the industries ranged from motor vehicles with average sales of over $20 billion per year to hats wi'th average sales of only $122 million per year. Spuri- ous size effects are `absent, however, since the dependent variable is expressed in ratio form. The core of the analysis is based upon unweighted interi'ndustry regressions. However, as su:bsequent tests indicate that lieteroscedasticity is pres- ent, weighted regressions are also estimated. Three o'f t'he industries-soft drinks, bakery products, and dairy products-sell in local markets, and this factor influences the appropriate measures of both con- centration and scale economies. Two `techniques `are used to handle this problem: a) Economies of scale are estimated in relation to the `typical local market, and the Kaysen and Turner concentration classifications which take into account the local market character of these industries, are used. b) In some equations, national concentration ratios are introduced along with a dummy variable which identifies the ideal market character of these three industries. One interesting characteristic of the underlying data is that the distribution of advertising-Sales ratios across the 41 industries is high'ly skewed. Twenty-five of the industries have ratios below three per cent while eight have ratios between three and six per cent and only six have ratios w-hich exceed six per cent. In the latter group, perfumes have an advertising-sales ratio of 15 per cent, cereals and drugs ten per cent each, soap nine per cent, malt liquor seven per cent, and soft drinks slightly over six per cent. Notable industries in the intermediate group include cigarettes and wines with ratios of about five per cent each. Although advertising per firm is positively correlated with advertising per dol- lar of sales, the positions o'f two important industries change radically depend- ing on which variable is used to measure advertising intensity. While tires and tubes and motor vehicles have quite low advertising-sales ratios, both are `among the small group of industries with high or very high average advertising outlays per firm among the leading firms. As a result, tires and tubes and motor vehicles are classified respectively as industries with moderate and high overall adver- tisi'ng barriers. MAJOR EMPIRICAL RESULT5 The simple correlation coefficients betweer~ profit rates and each explanatory variable are presented in table 1. All of the coefficients have the expected sign and all are statistically significant in at least one functional form. Moreover, the lograrithmic relationship appears appropriate' in the case of the growth of demand variable, while the opposite is the case with regard to the advertising-sales ratio. As the latter variable is already expressed as a percentage, it is measured in units comparable to the dependent variable. 32 High barriers are assumed either when the advertising-sales ratio exceeded eight per cent or when advertising expenditures among leading firms averaged more than $20 million. The same classification is given to industries where the average ratio fell be- tween four and eight per cent and average annual expenditures amounted to between $5 and $20 million. Low barriers exist when the advertising-sales ratio was less than two per cent and average expenditures less than $5 million or when the ratio lay between two and four per cent and expenditures did not exceed $1 million per firm. In other cases, moderate advertising barriers are assumed to exist. ~ Time Petroleum industry is excluded because of the difficulty of obtaining profit data comparable to other industries in view of the special tax treatment of that industry. PAGENO="0550" 2100 COMPETITIVE PROBLEMS IN THE DRUG INDUSTRY TABLE 1.-Simple correlation coefficients-Profit rates and various dimensions of market structure Market structure variables Correlation wi th profit rates Natural units 1 Logarithm 2 Growth of demand 0. 17 0.42 Capital requirements Economies of scale 3 0.43 0.25 0. 57 20.37 Advertising-sales ratio Advertisingperfirm Concentration ratio (4 firms) 0.42 3Q43 0.36 0.27 3Q5Q 4Q35 I The units of measurement are described in the text above. 2 In computing these coefficients, the structural dimension is measured in logarithms although the profit rate is measured in natural units. 3 Indicates coefficient is statistically significant at the 99 percent level. 4 Indicates coefficient is statistically significant at the 95 percent level. NoTE-Tests of significance are made on the basis of 1-tailed t tests. The correlation between profit rates and concentration in table 1 is based on a continuous four-firm concentration ratio. The relationship between these tw-o variables was also examined in terms of discrete groupings. Industries were divided according to both the three-way classification scheme proposed by Kaysen and Turner, and a two-way classification depending on whether the eight-firm concentration ratio exceeded or was less than 70 percent. The following results are obtained: Number of industries Average profit rates Kaysen and Turner trichotomy: Type I oligopolies Type II oligopolies TJnconcentrated 13 14 14 8.4 9.2 6.3 Dichotomy based on 8-firm concentration ratio at 70 percent: Concentrated 8 10.0 Unconcentrated 33 7.5 All industries 41 7.9 While the distinction between Type I and Type II Oligopolies seems, to the average, unimportant, there do appear to be substantial differences in profits be- tween concentrated and unconcentrated industries. These differences are impor- tant in both classification schemes. The core of the empirical work is the multiple regression equations which relate profit rates to various combinations of the explanatory variables. A. set of linear equations is presented in table 2. As may be ojserved, the advertising-sales ratio and the measure of capital requirements appear to be the most important explanatory factors. Their regression coefficients are generally significant even w-hen all other variables are included. The vari- able describing economies of scale seems from these results to be quite weak, although it has the expected sign in all cases. The advertising per firm coefficient is significant if the advertising-sales ratio is not included. Where both are in- cluded, it tends to be insignificant. Advertising outlays per firms are correlated w-ith absolute capital requirements (the simple correlation coefficient between these variables is 0.40) and this variable is not as statistically important in con- junction with the latter as is the advertising-sales ratio. PAGENO="0551" COMPETITIVE PROBLEMS IN THE DRUG INDUSTRY 2101 TABL E 2.-~-Multiple regression equations expla ining pr ofit rates-Linear results Adver- Econ- Absolute Growth Concentration Intercept tising- Advertising sales per firm ratio omies capital of scale require- ments of de- mand classsesa Type I Type R2 II Cor- rected R2 (1) (2) (3) - (4) - (5) - 0. 049 *0. 424 0. 00000059 (2.4) (1.1) 0. 052 0. 296 *0. 00000114 (1.6) (2.0) 0. 051 ~0. 437 0. 00000050 (2.1) (1.1) 0.048 **0 499 - (3.1) 0. 058 *0. 00000112 (2.1) 0. 113 ~0. 000281 (0.8) (3.0) 0. 113 (0.7) ~0. 000282 (3.0) 0.116 **0 009315 (0.8) (3.5) 0. 145 *0. 000227 (0.9) (2.3) 0. 0014 (0.6) 0. 0018 (0.7) 0. 0012 (0.5) 0.0016 (0.7) 0. 0012 (0.5) -0. 0158 0. 0084 **Ø 47 (1.0) (0.6) -0. 0058 0. 0115 *_0 32 (0.3) (0.8) -0. 0091 0. 0126 **0 46 (0.7) (1.0) -0.0117 0.0085 **045 (0.7) (0.6) -0. 0147 0. 0146 **0 38 (0.9) (1.0) **0 34 *0. 19 **0 35 **034 ~ 25 a Based on Kaysen and Turner groupings. NoTE-Figures in parentheses are t values. The statistical significance of the regression coefficients is tested by means of 1-tailed t test and of the multiple correlation coefficients by means of the F-ratio test. *Indicates coefficient is statistically significant at the 95-percent level. **Indicates coefficient is statistically significant at the 99-percent level. In none of the equations do the estimated coefficients of the concentration dummy variables exceed their standard errors. In addition, the coefficient for Type I Oligopolies has a negative sign throughout, which does not coincide with a priori expectations. While the impact of cOncentration is examined at greater length below, the linear results suggest that the partial effect of this variable may be relatively unimportant when it is intrbduced in conjunction with variables reflecting product differentiation, the height of technical entry barriers, and the rate of growth of demand. In an alternative formulation, four dummy varaibles were defined to represent high and moderate technical entry barriers and high and moderate advertising intensities, and these were introduced in place of the advertising, economies of scale, and absolute capital requirements variables. The results are presented in table 3. The dummy variables designed to measure the influence of technical barriers are not statistically significant, and in the second equation, the estimated coefficient for industries with high technical barriers is less than the coefficient for industries with moderate barriers. This result reflects in part the correlation between the high technical barrier dummy variable and concentration. The cor- relation coefficient between the dummy variable for high technical barriers and the Type I concentration dummy variable is 0.53. When the four-firm concentra- tion ratio is used, this coefficient rises to 0.68. This collinearity obscures the separate effects of concentration and technical entry barriers. TABLE 3.-Multiple regression equations containing composite variables Advertising Technical entry Concentration a Growth Cor- Inter- barriers barriers of de- R2 rected cept mand R2 High Moderate High Moderate Type I Type II (logs) (1) 0. 056 **0 0552 0. 0080 0. 0154 0. 0090 -0. 0132 -0. 00085 *0. 020 ~ 46 **0 33 (3.3) (0.6) (0.9) (0.5) (0.7) (0. 05) (1.9) Advertising Technical entry Inter- barriers barriers cept - High Moderate High Moderate Concen- Regional Growth Cor- tration industry of de- R2 rected ratio b dummy mand R2 variable (logs) (2) 0.044 *0.0403 -0.0049 -0.0150 -0.0013 0.000596 0.0311 *0.024 5*048 ~~036 (2.3) (0.3) (0.8) (0. 1) (1.4) (1.4) (2.4) Based on Kaysen and Turner groupings. Four-firm concentration ratio. NoTE-Figures in parentheses are t values. The statistical significance of the regression coefficients is tested by means of one-tailed t test and of tl1e multiple correlation coefficients by means of the F-ratio test. 5lndicates coefficient is statistically significant at the 95-percent level. *allsclicates coefficient is statistically significant at the 99-percent level. PAGENO="0552" 2102 COMPETITIVE PROBLEMS IN THE DRUG INDUSTRY To consider further the impact of concentration, the continuous concentration ratio was introduced into the analysis. As this variable is available only on a national basis, a dummy variable was also introduced to identify industries which sell in local rather than in national markets.34 It is interesting that the con- tinuous variable appears to have a stronger impact on profits than do the concen- tration dummy variables. Both the coefficient for concentration and for the local market dummy variable are significant at the 90 per level. In contrast to the weak effect of composite technical barriers, the composite advertising variable has a strong effect. In both equations, the high advertising variable is statistically significant and the coefficient appears relatively stable. In addition, the growth of demand variable becomes significant when introduced in logarithmic form. Additional sets of regression equations are presented in tables 4 and 5. In both sets, capital requirements are introduced in logarithmic form and the estimated coefficients are generally significant. However, some degree of collinearity exists between the measures of capital requirements and scale economies. When both variables are introduced into the equations, the variance of the estimates in- creases, and in some cases, the capital requirements coefficient becomes statisti- cally insignificant. In table 4, the advertising-sales ratio is used to measure the entry barriers created by high advertising expenditures, while in table 5, the dummy variable representing high advertising barriers is introduced. Both advertising variables are statistically significant in all of the equations presented, and the estimated coefficients of both variables are very stable. The dummy variable for high adver- tising barriers appears to be somew-hat stronger than the advertising-sales ratio. However, the coefficients of the other structural variables included in these equations are not sensitive to the particular advertising variable used. TABLE 4.-Multiple regression equations explaining profits rates- Major findings with advertising-sales ratio Inter- cept Adver- tising- sales ratio Capital Econo- require- mies of ments scale (logs) (logs) Growth Concen- Regional Cor- of tration industry 122 rected demand class dummy 122 (logs) variable (1) (2) (3) 0.042 0.042 0. 039 *0.362 (2.4) *0.362 (2.3) *0.343 (2.3) **0 0097 (3.2) **0 0096 0.000067 (2.5) (0.01) **0 0105 (2.8) 0.016 *4046 **040 (1.6) 0.016 *4046 **038 (1.6) 0. 015 0. 0043 0. 0278 4*0. 49 *40 40 (1.4) (0.3) (1.5) (4) 0.038 *0.341 (2.3) 0111 (3.6) 0.014 0.0280 **049 4*042 (1.4) (1.6) An industry is concentrated if the 8-firm concentration ratio equals or exceeds 70 percent; otherwise it is unconcentrated. 4lndicates coefficient is statistically significant at the 95-percent level. **Indicates coefficient is statistically significant at the 99-percent level. NoTE-Figures in parentheses are t values. The statistical significance of the regression coefficients is tested by means of one-tailed t test and of the multiple correlation coefficients by means of the F-ratio test ~ These market characteristics had already been accounted for in the Kaysen and Turner groupings. PAGENO="0553" COMPETITIVE PROBLEMS IN THE DRUG INDUSTRY 2103 TABLE 5.-Multiple regression equations explaining profit rates-Major findings with high advertising barrier Intercept High Capital Econo- Growth Concen- Regional adver- require- mies of tration industry R2 Corrected tising ments of scale demand class dummy R2 barrier (logs) (logs) (logs) variable (1) 0.053 **0 0379 *00066 *0.018 **048 **042 (2) 0. 069 (2.8) (2.0) (1.9) **0 0388 0. 0047 0. 0038 ~0. 019 - - **0 49 **0 42 (3) 0. 048 (2.8) (1.1) (0.7) (2.0) **0 0395 *0 0089 0. 015 -0. 0063 *0. 0318 **0 53 **0 45 (4) 0.048 (2.9) (2.4) (1.5) (0.5) (1.8) **0 0379 **0 008~ *0.016 *0.0316 **052 **046 (2.9) (2.5) (1.7) (1.8) An industry is concentrated if the 8-firm concentration~ratio equals or exceeds 70 percent; otherwise it is unconcentrated. NoTE-Figures in parentheses are t values. The statistical significance of the regression coefficients is tested by means of 1-tailed t test and of the multiple correlation coefficients by means of the F-ratio test. *Indicates coefficient is statistically significant at the 95 percent level. **Indicates coefficient is statistically significant at the 99 percent level. In both tables 4 and 5, a dummy variable identifying industries with eight- firm concentration ratios which exceed 70 per cent was introduced in order to examine the effect of this aspect of market structure in yet another specification. The coefficients, however, remain smaller than their standard errors. The regional industry dummy variable was also introduced into the equations in both tables. In all cases, its estimated parameters are significant at the 90 per cent level, and reach the 95 per cent level in table 5. While this variable was used originally to correct for the use of concentration ratios calculated on a national basis, it appears to have an independent effect which does not depend on the presence of the other variable. It is useful, therefore, to compare the structural features of the three local market industries included in our sample w-ith `the others. Relevant data are presen'ted in table 6. The scale economy vari- able is the most sensitive to this industry characteristic. As wo~ild be expected, these industries, on the average, have much higher estimates of the ratio of minimum efficient scale to market than do the national industries. We should expect, th'erefore, that the local market dummy variable represents the increased importance of economies of scale as well as the higher concentration levels in local markets. The multiple correlation coefficients for these equations are aways statistically significant. The included variables typically' account for about half of the total variation in industry profit rates. TABLE 6.-Local and regional industry characteristics Industry Concentra- tion class 1 Advertis- ing-sales ratio (percent) Advertis- ` ing per ` firm (mu- lions) Capital require- ments (millions) Economies of scale (percent) Growth of demand (ratio) Profit rate (percent) Soft drinks Type I~ - - Dairy Type IL~ - Bakery Type IL~ - Average of 3 indus- tries 6. 2 2. 2 2. 9 3. 8 80. 26 15. 12 1. 97 5. 78 $0. 75 2. 09 2. 57 1. 80 8. 2 14. 2 8.3 10. 2 1. 98 1. 16 1. 69 1. 61 10. 0 7. 9 9. 3 9. 1 All industries 3. 3 6. 03 ` 24. 32 4. 7 1. 83 7. 9 1 Kaysen and Turner groupings. PAGENO="0554" 2104 COMPETITIVE PROBLEMS IN THE DRUG INDUSTRY THE PROBLEM OF MULTICOLLINEARITY As was noted above, a number of the explanatory variables included in the analysis are collinear to some extent. While the simple correlations between the advertising-sales ratio and the other independent variables are typically low,35 it is useful to examine the sensitivity of the estimated coefficients for this variable to changes in the specifications of the regression equations. The results are pre- sented in table 7. The estimated coefficients are reasonably stable, ranging from 0.30 to 0.52. In addition, seven of the 15 coefficients presented are significant at the 99 per cent level, seven at the 95 per cent level, and the remaining coefficient at the 90 per cent level. TABLE 7.-Sensitivity of regression coefficients for advertising variables to changes in specification of regression equations A. ADVERTISING-SALES RATIO [N-variable in natural units; L-variable in logarithms; I and II denote Kaysen and Turner concentra- tion classes; H and M denote high and moderate dummy variables] Estimated coefficient Value of t Concentra- Regional Capital Scale Growth of Advertis- tion industry require- economies demand ing per dummy ments firm 0.42 0.30 0.42 0.43 2.27 1,11 N 1.61 1,11 N 2.40 1,11 N N N 2.50 1,11 N 0.48 2.75 N N N 0.52 3.24 N N N 0.50 2.92 N N N 0.34 2.21 I L N L 0.35 2.29 I L N L 0.49 3.28 N N L 0.41 2.41 N N L N 0.40 2.54 N N L 0.46 3.02 N N L 0.36 2.43 L 0.33 2.19 N N L B. DUMMY VARIABLES FOR HIGH ADVERTISING BARRIERS Estimated coefficient Other variables included - Value of I Concen- Regional Capital Economies Growth of Technical Moderate tration industry require- of scale demand barriers adver- dunimy ments tising barrier 0.010 0. 049 0.050 2.28 N N L H,M M 3. 67 I, II L 3.25 1,11 N L H,M M 0.045 3.39 N N L 0.038 2.80 L L 0.039 2.83 L L L 0.038 2.89 N N L L The stability of the estimated coefficient for the dummy variable denoting high advertising barriers is also tabulated. These coefficients appear to be insensitive to changes in the specifications of the equations. They lie between 0.038 and 0.050. Six of the seven coefficients presented are significant at the 99 per cent level and the remaining one at the 95 per cent level. The estimated effect of either the ad- vertising-sales ratio or the high advertising dummy therefore does not appear to be affected by which of the other variables are included in the equations. ~ The simple correlation coefficients between the advertising-sales ratio and other struc- tural variables are as follows: the log of economies of scale, 0.27; the log of capital re- quirements, 0.21; the log of growth of demand, 0.40; and the four-firm concentration ratio, 0.10. PAGENO="0555" COMPETITIVE PROBLEMS IN THE DRUG INDUSTRY 2105 While there may be other variables which affect the estimated relationships between advertising and profits, the importance of both advertising variables is relatively insensitive to changes in specification of the variables and the models examined in this paper.3° Significant correlations also exist between capital requirements, economies of scale, and concentration. This is not surprising, for one should expect the first two variabes to have some effect on the latter. At the same time, however, con- centration is influenced by other factors, such as the past record of merger ac- tivity in the industry. To examine the extent to which concentration is explained by scale economies and capital requirements,~ two multiple regression equations were fitted. The results, which are striking, appear in table 8. TABLE 8.-Multiple regression analysis-Concentration and technical entry barriers Concentration a Intercept Capital requirements (logs) Economies of scale (logs) ~ Regional industry dummy variable R2 Corrected R2 (1) Natural units (2) Logarithms 49*9 3. 85 **7 08 (3.9) **0. 244 (5.1) **6 91 (2.6) **O. 238 (3.4) -11.2 (1.2) -0. 294 (1.2) **0 71 **Ø 81 **0 68 **0. 79 a Four-firm concentration ratios. Figures in parentheses are t values. The statistical significance of the regression coefficients is tested by means of one-tailed I test and of the multiple correlation coefficients by means of the F-ratio test. a*Indicates coefficient is statistically significant at the 99-percent level. Absolute capital requirements, scale economies, and the local market dummy variable ~ together account for a substantial share in the variation in national concentration ratios. In logarithmic form, over 80 per cent of the variation is ex- plained by these variables. What is surprising is the small share of variation left to be accounted for by other factors. With this high a degree of inter-corre- lation, it is understandable that the estimated coefficients for concentration are not statistically significant. The role of cencenration appears closely linked to that of technical entry barriers and there is little remaining influence which is evident.E HETEROSCEDASTICITY AND WEIGHTED REGRESSIONS An examination of the residuals from a leading equation (number 4 in table 5) revealed that heteroscedasticity is present, as small industries typically have large residuals. There are two possible reasons for this phenomenon. The smaller industries may tend to have fewer firms, so that the variance of average profit rates is larger. The smaller Industries may also have smaller firms. Previous studies have indicated that the variance of profit rates among small firms is greater than among larger firms,30 and this would also account for a larger vari- ance for smaller industries. ~ This result, however, does not apply to average advertising expenditures per firm, which is more strongly correlated with the other explanatory factors. As a result, its statistical significance in regression analysis appears to depend on which of the other variables are included in the estimating equation. 37 The regional industry dummy variable was included because the concentration ratios are constructed on a national basis. The negative sign on the coefficient represents simply the downward bias of the national ratios in those industries. iS One should be wary of drawing any policy conclusion on the basis of this equation. Merger activity may be highly correlated with entry barriers. Furthermore, there is some element of spurious correlation between the scale economies measure and concentration. The scale economies measure used here is 0.5 times the reciprocal of the number of the largest plants required to account for one-half of industry output. It is, therefore, related to plant concentration. Since plant concentration and firm concentration may be expected to be correlated even in the absence of variations in relative scale economies, some spurious correlation exists between concentration and relative scale economies. (The authors are indebted to Joe S. Bain for the elaboration of this point.) ~9 Sydney S. Alexander, "The Effect of Size of Manufacturing Corporation on the Dim- trihution of the Rate of Return, this REVIEW, XXXI (Aug. 1949), 22.9-235. PAGENO="0556" 2106 COMPETITIVE PROBLEMS IN THE DRUG INDUSTRY To determine an appropriate weighting scheme, an empirical approach was adopted. The variance of the residuals was calculated for successive quartiles in the distribution of industry sales. From this tabulation, it was clear that the use of industry sales is inappropriate as a weighting variable as it would give too much emphasis to the largest industries. The square root of sales, however, is nearly proportional to the variance of the residuals, and was therefore chosen as the weighting variable. The weighted regressions were fitted both for all industries, and for all in- dustries except motor vehicles, as this industry is an outlying observation with respect to some of the variables, including the weighting variable. The results appear in tables 9 and 10. As is clear, the ft2 of each of the weighted regressions is considerably higher than the W of its unweighted counterpart. This is to be expected, since the weighting procedure deliberately emphasizes industries with smaller residuals and the ft2 measures the proportion of the weighted variance of the dependent variable explained by the regression equation. Another way of look- ing at this is that weighting essentially involves multiplying the equation by the root of the weights (in this case by the fourth root of sales) and proceeding by ordinary least squares. The ft2 indicates the success at predicting profit rates multiplied by the fourth root of sales.4° The results are impressive. About 75 percent of the weighted variance across all industries is accounted for by these equations and about 65 percent of the w-eighted variance is explained when the outlying auto industry is excluded. TABLE 9.-Weighted regressions with advertising sales ratio Inter- cept High ad- Capital vertising- require- sales meats ratio (logs) Econo- mies of scale (logs) Growth of de- Concentra- mand tion ratio (logs) Regional industry 132 dummy variable Cor- rected 132 (1) a. All indus- tries 0. 040 *0. 29 **0 013 (1.9) (4.8) 0. 0084 (1.0) 028 **9 76 (1.9) **0 72 b. Motor ye- hices cx- cluded~__ 0. 045 **0 44 *0 00~7 (3.1) (2.4) 0. 0096 (1.3) 0. 020 **0 67 (1.4) 62 (2) a. All indus- tries 0.056 *0.28 **0 010 (1.8) (3.1) 0. 0046 (0.9) 0. 0096 (1.1) **075 **071 b. Motor ve- hicles ex- cluded 0. 074 *w 42 0. 0010 (2.8) (1.2) 0. 0052 (1.2) 0. 011 (1.4) 67 62 (3) a. All indus- tries 0.040 *0.29 *0. 014 (1.9) (2.4) 0.0081 -0. 00003 (0.9) (0.08) *0.028 **076 (1.8) **072 b. Motor ve- hicles ex- cluded - 0. 017 **0 43 0. 0090 (2.9) (1.6) 0. 0088 -0. 00011 (1.1) (0.3) 0. 021 **Q 67 (1.4) **0 61 * Indicates coefficient is statistically significant at the 95-percent level. ** Indicates coefficient is statistically significant at the 99-percent level. NoTE-Figures in parentheses are t values. The high advertising barrier dummy variable and the advertising-sales ratio variable are introduced alternatively. The former is significant at the 09 per cent level in all equations, the latter at the 95 per cent level when the auto industry is included, at the 99 per cent level otherwise. The collinearity between capital requirements, economices of scale and concentration is again evident, but in contrast to the unweighted regressions, the economies of scale variable is some- times significant when introduced alongside capital requirements (and the latter variable is sometimes insignificant). °° It is important to note that the increase in 132 is no indication that the weighting used is the correct one. Indeed, a very lsigh 132 cnn be obtained by weighting with industry sales, which is clearly inappropriate. A subsequent test, moreover, was made on the extent of heteroscedasticity in the weighted regressions. The residuals from equation is in table 10 were calculated and the successive variances of these residuals were compared with the mean root sales in the relevant quartile. The fact that the two varinbies were nearly proportional provides some confirmation of the use of the square root of sales as the weighting variable. PAGENO="0557" COMPETITIVE PROBLEMS IN THE DRUG INDUSTRY 2107 These results, moreover, provide additional evidence of the stability of the co- efficients for these two advertising variables. They also point to the joint signifi- cance of technical barriers to entry and of concentration, but the collinearity among these variables prevents precise measurement of the separate effects of each of these variables. THE PROBLEM OF CAUSALITY We have found that the inter-industry variation in profit rates can be ex- plained quite well by a model incorporating the rate of growth of demand, some measure of advertising intensity, and variables reflecting the importance of con- centration and technical barriers to entry. The relationship between profits and either of the advertising variables introduced into the equations is quite robusL Throughout this paper, we have assumed that the direction of causality is from the independent variables to profit rates. Could the reverse be the case? A plausible case can be made that a significant feedback exists from profits to advertising expenditures, since advertising reflects the discretionary behavior of firms as well as the extent of product differentiation. Indeed, we should not be surprised if a time-series analysis, which emphasizes short-run effects, revealed that changes in profits preceded, rather than followed, changes in advertising expenditures. There are a number of factors, however, which suggest that the causality of the observed relationships runs largely from advertising expenditures to profits. A cross-sectional study tends to emphasize the long-run differences between in- dustries, and this in turn is more likely to reflect the structural rather than the behavioral aspects of advertising. Profit levels cannot influence those market and product characteristics which permit product differentiation via advertising. Firms with high profit rates will not have higher optimum advertising expendi- tures than firms with low profit rates in the same market situation. The pursuit of profits will hence limit the extent to which profits will be "spent" on advertis- ing, especially over a period of several years. TABLE 10.-Weighted regressions with high advertising barrier Inter- High adver- Capital Econ- Growth Concen- Regional require- omies of tration industry 112 Cor- rected cept tising barrier ments of scale demand ratio dummy (logs) (logs) (logs) variable 112 (1) a. All industries_ - b. Motor vehicles 0. 052 **0 035 (2.9) **0 0080 0. 012 *0. 027 ~0. 78 (2.5) (1.5) (1.9) **Ø 75 excluded 0.055 **0 032 *0 0064 0. 012 0. 025 **Ø 65 **Ø 59 (2) a. All industries - - 0. 090 (2.7) **0 037 (1.8) 1 (1.5) (1.7) 0. 0034 *0 0073 0. 013 **Ø 78 **0 75 b. Motor vehicles (3.1) (1.0) (1.7) (1.6) ~ excluded 0.097 **Ø 032 0. 0010 *0. 0080* *0. 014 **0 65 **0 60 (3) a. All industries - - b. Motor vehicles 0. 053 (2.7) *20 035 (2.9) (0.3) (1.9) (1.8) 0. 0093 - -- 0. 011 -0. 00010 *0. 028 **0 78 (1.6) (1.3) (0.3) (1.9) ~ **0 75 excluded 0. 057 **Ø 031 (2.4) 0. 0085 0. 011 -0. 00016 *0. 027 2*0 65 (1.5) (1.3) (0.4) (1.8) *20 58 *Indicates coefficient is statistically significant at the 95-percent level. **Indicates coefficient is statistically significant at the 99-percent level. Figures in parentheses are t values. In addition, if high profits lead to high advertising expenditures, we should expect that industries which have high profits for reasons other than product differentiation (e.g., concentration or technical entry barriers) would tend to have high advertising expenditures as well. Yet, as we noted above, advertising is only weakly correlated with the other dimensions of market structure. CONCLUDING COMMENTS On the basis of these empirical findings, it is evident that for industries where products are differentiable, investment in advertising is a highly profitable ac- tivity. Industries with high advertising outlays earn, on average at a profit rate which exceeds that of other industries by nearly four percentage points. This PAGENO="0558" 2108 COMPETITIVE PROBLEMS IN THE DRUG INDUSTRY differential represents a 50 percent increase in profit rates. It is likely, moreover~ that much of this proftt rate differential is accounted for by the entry barriers nreated by advertising expenditures and by the resulting achievement of market power. We note also the significant joint impact on profit rates of concentration and 4he entry barriers created by scale economies and high capital requirements. Al- 4hough the composite effect of these factors is clearly important, a more precise indication of the distinct effect `of any of these variables is hazardous because of the high degree of collinearity. As would be expected, the rate of growth of demand has an important positive impact on profits. Models which incorporate ~these variables fit the underlying data reasonably well, accounting for approxi- matey 50 per cent of the variation in industry profit rates. These empirical results suggest that factors which promote product differentia- flion may be as important as those which influence the size distribution of firms :in terms of their effect upon the achievement of market power. Current policies -which tend to emphasize the role played by concentration may need to be supple- :mented by those concerned directly with the nature and extent of product dif- ferentiation. Policies dealing with these matters would be an important compon- sent in a general policy designed to promote competition. APPENDIX DATA SOURCES AND TECHNICAL ADJUSTMENTS The industry data used are reported at or aggregated to the level of I.R.S. "minor industries," which are roughly comparable to S.I.C. three-digit industry groups. The source for each variable is listed in table Al. The sample was chosen originally to gain complete coverage of all consumer goods industries. All "miscellaneous" industries were eliminated, however, be- cause of the obvious conceptual problems. In addition, three other industries were dropped from the sample: newspapers, while technically a manufacturing in- dustry was considered to have sufficient "service" elements to make its inclusion inappropriate; petroleum refining, because of the unusual statistical problems which result from the tax treatment `of mineral depletion; and motor vehicle parts, because of the lack of comparable Census data. Average profit rates and advertising-sales ratios for the remaining 41 industries are presented in table A2. The variables are defined and explained in the text. The calculation of the technical entry barrier variables and the rate of growth of demand involved using both Census and I.R.S. data. The various specific adjustments made to reconcile data drawn from these two sources and reported at different levels of aggrega- tion are described in the next two sections. Technical Barriers to Entry These variables are based on data from the 1954 Census of Mann factures. To carry out `these computations, it is necessary to relate industries as defined by the Census Bureau to those of the Internal Revenue Service. This is done on the basis suggested by the Census Link Project.4' Within 5.1.0. four-digit industries, average plant size among the largest plants which account for 50 per cent of industry output is used as the estimate of mini- mum efficient plant scale (MES). Data on shipments are used in all cases where available. In the few remaining cases, the calculations are based on value added. When the ratios of MES to industry output are obtained, the average percentage among component four-digit industries within the relevant I.R.S. industry is calculated, using shipments as weights where available and value added as weights elsewhere. In determining the capital requirements variable, the scale of an efficient plant is measured in most instances by the value of the shipments but in a few by value added. In the latter cases, these figures are multiplied by the ratio of ship- ments to value added for the same four-digit industry but in a later year. 41Bureau of the Census, Enterprise Statistics (1958), Part 3. PAGENO="0559" COMPETITIVE PROBLEMS IN THE DRUG INDUSTRY 2109 TABLE Al-SOURCES OF DATA Variable Source (1) Profit rate Internal Revenue Service Source Book of Statistics of IncomeS Average values for 1954-57. (2) Advertising Do. (3) Concentration: (a) Trichotomous and dichotomous classi- Carl Kaysen and Donald F. Turner, Antitrust Policy, statistical fications. appendix. (b) Continuous 4 firm ratio George J. Stigler, Capital and Rates of Return in Manufacturing Industries, 206-215. (4) Economies of scale relative to market 1954 Census of Manufactures. (5) Absolute capital requirements 1954 Census of Manufactures and Internal Revenue Service Source Book of Statistics of Income. (6) Rate of growth of demand Internal RevenueServiceSource Book ofStatisticsof Income. (7) Local market dummy variables Carl Kaysen and Donald F. Turner, Antitrust Policy, statistical appendix. When estimates of MES measured in shipments for all four-digit industries are obtained, these are averaged, using value added as weights, to derive the value in the larger I.R.S. industry. These averages are then multiplied by the appropriate assets-sales ratio for the I.R.S. industry, and the resulting figures used to represent the level of capital required for efficient entry. In the case of the motor vehicle industry, Census data are unavailable. Bain's estimates for this industry are therefore used for both the extent of scale econ- omies and the level of capital requirements. For the three regional industries-soft drinks, dairy products, and bakery products-it is assumed that the appropriate market is the typical large metro- politan area. Output data are not available for four-digit industries by standard metropolitan area. Consequently, value added data for larger three-digit group- ings are used. Average value added in large metropolitan areas for the three-digit industry is multiplied by the ratio of total national shipments in the relevant four-digit industry to total value added in the associated three-digit industry to obtain the estimate of four-digit industry shipments within the typical local market. This figure is then used as the denominator in the estimate of M.E.S. to market for the four-digit industry.~ Where necessary, the resulting estimates are aggregated to the I.R.S. minor industry level as described above. Rate of Growth of Demand This variable is the ratio of I.R.S. gross sales in 1957 to that in 1947. In a few cases, however, LR.S. data are not available for both years and alternative pro- cedures are used. In 1947, the I.R.S. industries "Cigars" and "Cigarettes" are aggregated as "Tobacco Manufactures." To disaggregate the reported figure for gross sales, Census data on value of shipments (excise taxes excluded) for 1947 were ex- amined. The ratio of shipments in "Cigars" to total tobacco manufacturing is multiplied by I.R.S. gross sales for tobacco manufactures, and the resulting prod- uct used to denote gross sales in "Cigars." Gross sales in "Cigarettes" is obtained residually. ~ Output data were available for 56 standard metropolitan areas for the beverage in- dustry (S.I.C. 208) ; 59 such areas for bakery products (S.I.C. 205) ; and 61 such areas for dairy products (SIC. 202). ~ These corrections were made for the following subindustries: SIC. 2081-bottled soft drinks: SIC. 2021-creamery butter; SIC. 2027-fluld milk and other products~ SIC. 2051-bread and related products. PAGENO="0560" 2110 COMPETITIVE PROBLEMS IN THE DRUG INDUSTRY TABLE A2.-AVERAGE PROFIT RATES AND ADVERTISING-SALES RATIOS IN 41 CONSUMER GOODS INDUSTRIES, 1954-57 [In percent[ Profit rate Advertising sales ratio (1) Soft drink 10. 0 6. 2 (2) Malt liquors 7. 2 6. 8 (3) Wines 7.3 5.2 (4) Distilled liquors 5. 0 2. 1 (5)Meat 4.6 .6 (6) Dairy 7.9 2.2 (7) Canning 6. 4 2. 9 (8) Grain mill products 7.0 1.9 (9) Cereals 14.8 10.3 (10) Bakery prodscts 9. 3 2. 9 (11) Sugar 5.8 .2 (12) Confectionary 10.6 3. 5 (13) Cigars 5.3 2.6 (14) Cigarettes 11.5 4.8 (15) Knit goads 3.8 1.3 (16) Carpets 4.5 2.0 (17) Hats 1.6 2.2 (18) Men'sclothing 5.9 1.2 (19) Women's coithing 6. 1 1. 8 (20) Millinery -1.3 .8 (21) Furs 5.7 1.0 (22) Furniture 9. 7 1. 5 (23) Screens and venetian blinds 9.3 1.6 (24) Periodicals 11. 7 - 2 (25) Banks 10. 1 2. 4 (26) Drugc 14. 0 9. 9 (27) Soaps 11.7 9.2 (28) Paints 9.9 1.5 (29) Perfumes 13. 5 15. 3 (30) Tiresand tubes 10.2 1.4 (31) Footwear 7.6 1.5 (32) Handtools 11.4 4.2 (33) Household and service machinery (not electrical) 7. 3 1. 9 (34) Electrical appliances 10. 3 3. 5 (35) Radio, TV, and phonograph 8.8 2.2 (36) Motorcycles and bicycles 5. 2 1. 1 (37) Motor vehicles 15. 5 - 6 (38) Instruments 12. 0 2. 0 (39) Clocks and watches 1. 9 5. 6 (40) Jewelry (precious metal) 5. 3 3. 2 (41) Costume jewelry 1.4 4.0 I.R.S. data for "Screens and Venetian Blinds," in 1947, are included in "Miscellaneous Furniture." As a result, Census value added data for both 1947 and 1957 are used. Similar data problems exist in the case of "Drugs," "Per- fumes," "Instruments," and "Costume Jewelry." In all of these instances Census data in 1947 and 1957 are used. With regard to "Perfumes" and "Costume Jewelry," information on value of shipments is available in both years and is therefore used, while value added data are used in the other cases. Concentration The classification of industries into Kaysen and Turner concentration classes is based on data for 1954, and their definitions are used in most cases. Where I.R.S. minor industries include a number of Kaysen and Turner markets, weighted averages based on value of shipments are used to determine the appro- priate classifications. Further adjustments are made to account for the local market character of three industries: soft drinks, dairy products, and bakery products. In addition. the soap industry falls just below the boundary between Type I and Type II Oligopolies. Since the same major firms are dominant in the important product lines, the industry is reclassified as a Type I Oligopoly. The radio and television industry falls just above the same boundary. It is classified as a Type II Oligopoly since the largest single subindustry falls into that category. For appliances, the Kaysen and Turner definition differed substantially from that used by the In- ternal Revenue Service. After an appropriate adjustment the concentration data indicated that this industry is a Type II Oligopoly. PAGENO="0561" COMPETITIVE PROBLEMS IN THE DRUG INDUSTRY 2111 The classification of industries according to whether the eight-firm concentra- tion ratio exceeded or fell short of the 70 per cent level is based on the concen- tration ratios presented by Kaysen and Turner. (Whereupon, at 11 :55 a.m., the hearing was adjourned subject to call of the Chair.) S1-280-68-pt. 5-36 PAGENO="0562" PAGENO="0563" APPENDIXES APPENDIX I-CORRESPONDENCE BETWEEN MR. BENJAMIN GORDON AND DR. P. M. COSTELLO RE TECHNOLOGICAL PROGRESS IN THE PHARMACEUTICAL INDUSTRY MARCH 6, 1968. Dr. P. M. COSTELLO, Department of Economics, Smith College, Northampton, Ma$s. DEAR DR. COSTELLO: It is my understanding that you have done considerable research concerning technological progress in the pharmaceutical industry. I should be very grateful if you would submit to our Monopoly Subcommittee a brief summary of your findings in this field. Your contribution on this aspect of the subject would be very useful in our study of this industry. Your cooperation is appreciated. Sincerely, BENJAMIN GORDON, Staff Economist. SMITH COLLEGE, No-rthampton, Mass., March 13, 1968. Mr. BENJAMIN GORDON, Staff Economist, Select Committee on Small Business, U.S. Senate, Washington, D.C. DEAR MR. GORDON: In reply to your letter of the 6th, I am enclosing a paper which examines the pace of technological progress in the ethical drug industry. This study was developed from research for my doctoral dissertation which con- sidered the narrower question of patents in the antibiotic segment of the industry. Sincerely, P. M. CosTELLO. TECHNOLOGICAL PROGRESS IN THE ETHICAL DRUG INDU5TRY* (By P. M. Costello, Assistant Professor of Economics, Smith College, Northampton, Mass.) One critical test of an industry's market performance is its rate of tech- nological progreSs. For the ethical drug~ industry this dimension of performance is of added significance in view of the increasing role played by drugs in the management of human diseases. This paper examines a sample of 528 new drugs innovated between 1945 and 1965. The first section presents estimates, based upon medical evaluation and opinion, of the proportion of these innovations that represent significant advancement in the art of chemotherapy. The second section considers the problems of constructing an adequate criterion `to judge the pace of advancement from the antibiotic segment `of the industry. I. THE RECORD OF TECHNOLOGICAL PROGRESS The sample of new drugs introduced `to the market between 1945 and 1965 was compiled mainly from the American Medical Association's New and Nonofficial Drugs. In general the drugs selected for analysis were new chemical structures excluding combinations of old drugs and new dosage forn~s. The exclusion of the latter two categories was founded on the assumed lower order of scientific ability required to produce such products. But even within the category `of new chemical structures major differences in the effectiveness of these drugs were immediately apparent. *An earlier draft of this article received valuable comment from Professors Joel B. Dirlam and Robert Britt. 2113 PAGENO="0564" 2114 COMPETITIVE PROBLEMS IN THE DRUG INDUSTRY In an attempt to meet the qualitative problem two categories are defined. The first is designed to include drugs which were in one degree or another an im- provement over older products. These varied from drugs which were considered medical breakthroughs to drugs which reduced the incidence of undesirable side effects observed with established chemotherapeutic agents. Drugs which fell in this category were counted as examples of technological progress. The second category includes those drugs which from a technical point of view were no advancement over older products. The range in this category is from drugs which had the same degree of effectiveness and the incidence of side effects as older products to drugs which were removed from the market or limited in their application because of undesirable side effects. Also included in this category are drugs where medical opinion was divided or clinical experience too limited to form a statistically valid sample of the value of the drug. Drugs falling in this category were counted as examples of product differentiation which exhibit no significant technological progress. To classify each of the drugs in `the sample the Medical Letter1 and the AMA's New and Nonofficial Drugs were selected as authoritative publications that best summarized clinical experience with each drug. Thus the classification of a drug is directly dependent upon its technical efficiency relative to a particular medical problem. This scheme gives no weight to sales volume as an index of the relative value of a drug. This recognizes that physicians are responsive to the promotional claims of the manufacturers and will prescribe a new drug before sufficient clinical experience is amassed to establish the comparative value of the new drug relative to older products. As a result a drug significant in the management of a disease of low incidence in the population and a corre- spondingly low volume of sales receives the same weight as a drug which cures diseases of high incidence and high volume of sales.2 Biases in the sample are undoubtedly numerous and difficult to evaluate and could operate in either direction. First, the source p~iblications do not clearly indicate the basis of selecting a particular drug for discussion. Inclusions may well be in response to the promotional claims of the manufacturer with the recognition that the majority of practicing physicians have neither the facilities, the time, the training, nor the range of patients necessary to conduct statistically valid tests of such claims. Second, definitive judgment of the place of a particular drug in the physician's tool kit may require several years of experience before the population treated is sufficiently large to expose undesirable side effects. Thus, a drug currently considered significant may, with further testing, be found to have only a restricted application. Third, drugs now considered in- significant may, through modifications in techniques of administration, or in combinations with other drugs, be found valuable in the management of specific diseases. Finally, there is a bias in restricting the sample to new chemical struc- tures. It is possible that combinations of old drugs, for example, may increase the number of diseases controllable through chemotherapy. But there is no evidence `to date that this technique has produced anything of importance. In Table I of the sample of 528 new chemical drugs is divided according to the classification scheme outlined above. Column I contains estimates of the total number of new chemical structures innovated yearly between 1945 and 1965. 1 The Medical Letter is a publication designed to acquaint the practicing physician with the latest information on drug products. 2 For a different approach to the qualitative problem see W. S. Comanor, "Research and Competitive Product Differentiation in the Pharmaceutical Industry in the United States.' Economics, Vol. 31. November 1964, pp. 372-84; and "Research and Technical Change in the Pharmaceutical Industry." Review of Economics and Statistics, Vol. 47, May 1965, pp. 182-90. Comanor's investigation of the whole range of new drug Innovations weights each innovation by its sales volume for the first two years following Its introduction. My objection to following this approach In the present study is that the time required to develop experience with any particular new drug to support a definitive opinion on its technical worth seems to exceed the two-year period following introduction. PAGENO="0565" COMPETITIVE PROBLEMS IN THE DRUG INDUSTRY 2115 TABLE I-ESTIMATE OF TECHNOLOGICAL PROGRESS FOR NEW CHEMICAL DRUGS, 1945-65 Year Estimate of the oumber Estimate of the number Total new chemical representing product representing techno- drugs differentiation logical progress Col. III as a percent of col. I (I) (II) (III) (IV) 1945 3 2 1 0.33 1946 16 14 2 .31 1947 16 11 5 1948 14 10 4 .29 1949 28 21 7 .08 1950 25 23 2 1951 21 17 4 1952 22 21 1 .13 1953 38 33 5 1954 35 33 2 1955 27 24 3 .11 1958. 27 26 1 1957 38 35 3 1958 41 37 4 .10 1959 52 46 6 .12 1960 34 28 6 1961 18 16 2 .10 1962 20 19 1 1963 16 14 2 .13 1964 16 15 1 1965 21 20 1 .05 Total 528 465 63 .12 Industry figures for new chemical drugs are available for the period 1951-61.° Comparing these figures with those in Column I indicates that the industry data exceeds the sample data by a yearly, average of 10.8 drugs. In part this dis- crepancy is due to differences in the year of introduction reported for certain drugs and perhaps due in part to differences in the definition of new chemical drugs used by the industry and the AMA. It seems likely however, that if any of these innovations reported by the industry were important they would have been reported to the medical profession through the AMA's publications. The effect of including the industry data here would be to reduce the proportion of total innovation that represented technological advancement. Beginning in the late 1940's there is a persistent increase in the number of drugs marketed each year until 1959, followed by a marked decline up through 1964. The series representing technological progress, Column III, however fails to increase appreciably. For the period 1945-50 the average ratio of innovations to advances is 5 to 1, while for the period after 1950 this ratio increases to an average of 10 to 1. Both series exhibit ,a decided break appearing in 1960 and 1961 respectively. These broad movements can in part be explained by a combination of structural factors internal to the industry and external institutional factors. For the 1945-50 period the industry engaged in a process of forward vertical inte- gration. The objective of this integration was better control over prices through the elimination of the firms solely engaged in packaging. The major example was set by Parke Davis and Lederle, the innovators of the first broad spectrum drugs. Both of these firms were fully integrated into the retail pharmacy and hospital markets and held strong patents on their innovations. In refusing domestic li- censes on these products the traditional marketing pattern in the industry was altered. With the elimination of the packagers, direct control over price became possible. This example of the use of the patent and forward integration was un- doubtedly a factor in Merck's merger with Sharp and Dohne as well as Pfizer's entry by internal expansion. The structural change in the industry suppressed one form of competition- imitation from packagers or from integrated firms under license agreements. The alternative avenue open wa's to invent around the patent to produce a close substitute. Thus the increase in the proportion of innovations which represented product differentiation was partly in response to the closing of one method of competition. ~ Pharmaceutical Manufacturers Association, Prescripting Drug Industry Fact Book (Washington, PMA), 1963, pp. 2-5. PAGENO="0566" 2116 COMPETITIVE PROBLEMS IN THE DRUG INDUSTRY The decline in both innovation and technological progress observed in the early 1960's can be explained as the industry's reaction to criticism generated by the Kefauver investigation and the stricter controls imposed by the FDA during this period. Table II indicates that the rate of increase in expenditures for research and development fell measurably in 1960, leading to a decline in inno- vations as well as in the number that represent technological progress. TABLE Il-RESEARCH AND DEVELOPMENT EXPENDITURES AND EMPLOYMENT FOR ETHICAL DRUGS Year Expenditures Percentage R. & D. Percentage (millions) change employment change (thousands) 1956 $94 1957 104 0.11 4.7 1958 128 .23 5.1 0.08 1959 154 .20 5.9 1.6 1960 163 .06 6.0 0.2 1961 181 .11 6.2 0.3 1962 198 .09 6.7 0.8 1963 215 .09 6.8 .01 1964 238 . 06 7. 2 . 06 1965 268 .13 7.6 .06 Source: National Science Foundation, Review of Data on Science Resources, Washington, various issues. In general, if technological progress is measured as the percentage of ismo- vations w-hich find general medical acceptance, it is clear that over time the pro- portion of drugs that can be classified as representing meaningful progress de- clines. Or to state the case somewhat differently, the series on the number of new drugs of lasting value appears rather constant over time in spite of the marked increase in the number of innovations. These data then tend to support the observation that while the industry has increased its expenditures on re- search and development, much of the added effort appears to have resulted in product differentiation. IT. CRITERION FOR JUDGING PERFORMANCE Attempts to judge the "fairness" of prices or profits are generally made on the basis of comparisons of price to costs of production, or profits of one firm in comparison with those of another. But with technological progress as the per- formance variable no such objective criterion exists. In the present case we do not know whether the proportion of innovations representing technological prog- ress w-ould have been higher or low-er if the industry did not, for example, have access to patent protection. If the research resources which were directed to patent circumvention had been used to attack more significant problems, it is pos- sible that the research output would have been higher. But it is also possible that given the quality of these resources, product modification was the best they were capable of in their employment in the industry. The problem is of course a familiar one, in that w-e cannot rerun the industry's reactions under different structural characteristics. An alternative approach is to compare progress under different market struc- tures. But even here competing hypotheses of the effect of structure on perform- ance are in existence and debated. The Schumpeter version that only firms with some degree of market power have the resources to innovate is opposed by the hypothesis that innovation offers competitive firms escape from the rigors of competition. The antibiotic segment of the drug industry does shed some light on these competing hypothe~s and can aid in the judgment of whether prog- ress has been adequate. Table III presents data on the innovation record for the ant.ibiotic segment of the industry. Total innovations are divided between the so-called narrow- and broad spectrum drugs corresponding to a competitive segment and a monopo- lized segment of the market. In the narrow- spectrum market approximately 30 percent of these innovations represented some degree of technological progress, w-hile in the broad spectrum market one-half of the innovations were signifi- cant in this respect.4 In light of different market structures associated w'ith these The four broad spectrum drugs and their classification are as follows: Both chioramphenicol and chlortetracycline are counted as evidence of technological progress. Oxytetracycline and tetracycline are counted as evidence of product differentiation as both are very close substitutes for chiortetracycline. PAGENO="0567" COMPETITIVE PROBLEMS IN THE DRUG INDUSTRY 2117 two markets, the differences in the proportion of innovations representing an advance lends initial support to the hypothesis that monopoly is a necessary condition for progress. But examination of the development of the broad spec- trum market and its operation weakens this conclusion. TABLE Ill-PRODUCT INNOVATION IN THE NARROW AND BROAD SPECTRUM ANTIBIOTIC MARKETS Narrow spectrum Broad spectrum Year innovations innovations 1945 1 1946 1 1947 2 1948 1 1 1949 1 1 1950 3 1 1951 1 1952 1 1953 2 1 1954 2 1955 3 1956 1957 3 1958 3 1959 1960 3 1961 1962 2 1963 1 1964 2 1965 2 Total 34 4 The origins of the antibiotics are found in Alexander Fleming's discovery of penicillin in 1928 and in the development research undertaken by a group headed by H. W. Florey in England in the late 1930's. England's involvement in the war and problems of mass producing penicillin caused Florey to seek commercial sources of supply in the United States. The majority of firms contacted by Florey expressed little interest in pursuing the research necessary for large scale production. But he was successful in his contact with the Department of Agriculture's Northern Regional Research Laboratory (NRRL). With the entry of the United States into the war a cooperative program was instituted by the Committee on Medical Research (CMR). The program included a number of drug firms, university and government laboratories, as well as laboratories in England. The commercial laboratories of the drug firms initially concentrated their research on the synthetic production of penicillin,5 while the noncommercial laboratories concentrated on the problem of mass production. The development of improved mold stains by the university laboratories and the deep fermenta- tion' production process by the NRRL essentially solved the key problems of mass production.° By the mid-1943 the CMR shifted the emphasis in the penicillin program from research to production. To overcome the reluctance of the firms to abandon their synthetic research and adopt the available technology, the program was expanded to include firms not initially~ under contract with the government It has been argued that the drug firms concentrated on the synthete approach at the expense of the production technology suggested by the NRRL. The reason offered for this approach was that product patents were unavailable on penicillin, and the Department of Agriculture held the key process patents. Richard Harris, Annals of Legislation, the Real Voice, The New Yorker, for March 14th, 21st and 28th, 1964. Especially page 69 of the March 14th issue, ". . . the firms were too busy trying to corner patents on the various processes in the production of penicillin to produce much of it, and the government began to press them to work together." A similar conclusion was reached in Federal Trade Commission, Economic Report on Antibiotic Manufacture (Washington, Govt. Printing Office, 1958), pp. 37-38. In defense of the firms involved it should be noted that his- torically, the synthetic production of drugs had made the `natural technology' obsolete. For a description of Merck's experience and others see, Tom Mahoney, The Merchants of Life (New York, Harpers, 1959), Chaps. 11, 13, 16. Until the mid-1940's penicillin was produced by two methods. The less efficient `sur- face technique' involved the growth of penicillin on the surface of the culture medium. With the `deep fermentation technique' the medium was aerated allowing growth through- out the medium. PAGENO="0568" 2118 COMPETITIVE PROBLEMS IN THE DRUG INDUSTRY through the CMR and pressure on those under contract to begin production. In addition, government funds were provided for plant construction and modifi- cation,7 and as a result, the War Production Board permitted the sale of peni- cillin through normal trade channels in March of 1945. After the removal of government restriction on the safe of penicillin, market forces operated to reduce prices. Tariff Commission data indicate that between 1945 and 1950 the realized price of penicillin fell from $3,055 to $282 per pound, while production increasing from 11,746 to 329,746 pounds.8 The behavior of penicillin prices can be explained by two factors-conditions of entry and the existence of a bulk market. A product patent on penicillin was unavailable, and the process patents were held by the Department of Agriculture, w-hich follow-ed a policy of freely licensing applicants; entry could not be re- stricted in manufacturing or selling based upon patents. Available evidence indi- cates that there were 16 penicillin suppliers in 1945, and 13 in 1950.~ The second factor was that several of the major drug producers were not integrated forward into direct selling to hospitals and retail pharmacies. These firms generally sold a significant proportion of their output through the bulk market to packagers who resold under their own brands. The FTC study re- ported 27 packagers in 1950, increasing to 45 by 1956, concluding that entry at this level was relatively easy.1° Thus packagers purchased their supplies in the open market and sold in competition with the integrated producers. The emphasis on price competition in the penicillin market was not at the expense of product improvement. Various modificat~ions of penicillin were introduced. For example, Squibb introduced the first orally effective penicillin in 1945, Wyeth introduced benzathine penicillin which has a slower rate of absorption in 1951, and in 1955 Lilly marketed penicillin V claiming higher potency. Product patents were generally obtained on these specialties but usually only after interference proceedings had been declared by the Patent Office. Settlement took the form of withdrawal of a competitive patent application in return for a license to produce and sell under any patents that be issued. While data on realized prices of these products are not reported separately by the Tariff Commission, these prices generally drifted downward following the price of the basic drug although at a somewhat higher level. Thus in the penicillin market the underlying condition of freedom of entry and the bulk market produced a market competitive in price and with product improvement. The second major antibiotic following penicillin was streptomycin, discovered by Selman Waksman at Rutgers University in 1943. Under the terms of the research contract between Waksman and Merck, the latter received exclusive commercial rights to this drug. To exploit its position, however, Merck had to first convince the Patent Office that streptomycin was sufficiently important to warrant patent protection,11 and second, to integrate forward into the retail market to avoid dependence upon the bulk market. Both steps were necessary. For with the sale of a patent product to packagers, the patent monopoly was exhausted and price cutting could legally begin.12 The demonstration of the potential importance of streptomycin resulted in Waksman's request to Merck that right of commercial exploitation be given up. The sequent agreements provided for the assignment of the patent applica- tion to a nonprofit foundation at Rutgers, and royalty abatement of $500,000 to cover Merck's development costs. In 1948 the patent was issued to the founda- tion which followed an unrestrictive licensing policy.13 Seven firms in addition to Merck were licensed to manufacture streptomycin. Four of these firms, including Merck,14 were not integrated forward, selling primarily in the bulk market to FTC, op. cit.. pp. 47-56. U.S. Tariff Commisoion. PVntlzetzc Organic Chemicals, United Production. and ,S'ale, Reports No. 159 and 167 (Washington. Govt. Printing Office), various issues. ° The decline in the number of manufacturers was largely due to their adoption of inefficient production techniques. With the end of military purchases and the onset of price competition these firms wculd not cover costs and transferred their resources to other markets. 10 FTC. on. cit., p. 66. 11 Prior to streptomycin the Patent Office took the position that the isolation of products of naturo did not rise to the level of invenion and hence were unpatentabble. Merck successfully arrued that antibiotic products were transitory in neture and that their isolation constituted invention within tile purpose of the statutes. F. Cacciapaglia, Jr.. nail H. B. Rroakman. The Proposed Drug Industry Antitrust Act-Patents, Pricing and the PubPc, The George Washington Law- Review, Vol. 30. June 1962, p. 890. 12The doctrine that the first sale of a patented product exhausted the patent monopoly was set out in Boston Stores Co. v. American Gramophone, 246 US 8 (1917). 11 FTC. op. cit.. p. 229. 14 Merck did not enter the retail market until its 1952 merger with Sharp and Dohne. PAGENO="0569" COMPETITIVE PROBLEMS IN THE DRUG INDUSTRY 2119 packagers. The market results were similar to those in the case of penicillin. Tariff Commission data indicate that the, realized price of streptomycin fell from $2,866 per pound in 1946 to $160 in 1950. Competitive market conditions did not prevent Merck from further research on streptomycin. In 1948 dihydro- streptomysin was marketed as an improved version of streptomycin, and the product patent was issued to Merck in 1950. Six firms were licensed, three of which were bulk sellers. The price of dihydrostreptomycin has closely followed that of streptomycin. The marketing of the broad spectrum ailtibiotics in the late 1940's and early 1950's ended the competitive characteristic of the postwar antibiotic market. In late 1948 Lederle Laboratories, a division of American Cyanamid, introduced the first broad spectrum antibiotic, chiortetracycline, closely followed by Parke Davis with chioramphenicol. As both firms were fully integrated into direct selling the bulk market was bypassed. With the issuance of the product patents competition could be legally excluded. Evidence through 1960 indicated that neither company made bulk sales to packagers and all requests for domestic licenses were refused. The prices of these drugs were reduced from their intro- ductory level of $15 per bottle of 16-250-milligram tablets, reaching a floor in 1951 at $5.10 and remaining constant until the opening of the Kefauver investi- gation in 1960. In 1950 the Chas. Pfizer Company introduced the third broad spectrum drug oxytetracycline. While waiting for FDA clearance a sales force was organized, thereby avoiding sales in the bulk market. With the issuance of the product patent all requests for licenses were rejected. The marketing of oxytetracycline caused Lederle to reduce the price of its entry, with Parke Davis meeting the price cut. The last major broad spectrum drug was tetracycline, introduced in 1953 by Lederle, but also produced by Pfizer and Bristol Laboratories and sold by Squibb and Upjohn as a result of patent interference claims. The introduction of tetracycline did not disturb the broad spectrum price level nor were domestic licenses granted to other producers or sellers. Thus an examination of the early narrow spectrum market indicates a com- petitive market structure at both the manufacturing and packaging levels. Com- petitive behavior can be observed in price as well as in product innovation. The availability of patents and forward vertical integration provided a means by which the broad spectrum innovators could escape the rigors of competition. The examination of the antibiotic segment of the industry points up quite clearly the major flaw iii the measurement of technological progress employed earlier. If the market is subject to some degree or form of monopoly control where firms can reach agreement not to compete on the basis of price, and prices can be maintained at relatively high levels, there is a tendency not to disturb these agreements through product innovations. This has the effect of decreasing the base and overstating the proportion of innovations which constitute technological progress. Having achieved control over price and profits the major stimulant to further research is removed. There is evidence from antibiotics that once market control is established further research in the therapeutic area ceases. First, it seems doubtful that tetracycline, introduced in 1953, is the ultimate broad spec- trum drug. The strong patent positions' of the manufacturers and Pfizer's ag- gressive defense of its patent on tetracycline undoubtedly reduce the potential profitability of this segment of the market for firms contemplating entry through research. The broad claims allowed in antibiotic patents would be a major factor in discouraging entry directed research. The potential entrant to be free of patent infringement suits would be forced to discover antibiotic producing micro- organism unclaimed by the established manufacturers. Surmounting this prob- lem there remains the threat of infringement suits designed to harass. And unless the innovating firm is of equal size with its competitor the evidence indicates that a policy of harassment will be successful in forcing him from the market. Here one can cite the fact that Pfizer' drove 33 smaller competitors from the tetracycline market between 1960 and 1965 by this method. Second, there is evidence that Lederle, after the discovery and innovation of chiortetracycline abandoned research in this area until faced with serious com- petition from Pfizer in late 1952. In 1948 Lederle carried out experiments with chlortetracycline that produced a substance with antibiotic properties. In the patent dispute with Pfizer, Lederle claims that this substance was in fact tetra- cycline. But in 1948, Lederle dominated the broad spectrum market with chiorte- tracycline and evidently did not feel compelled to complete its experiment. But in 1952, with Pfizer's discovery of tetracycline Lederle ". . . resumed the 1948 PAGENO="0570" 2120 COMPETITIVE PROBLEMS IN THE DRUG INDUSTRY work, and in December 1952 or January 1953 . . . produced tetracycline . . ."" This lends support to the argument that once a dominant position is established and protected by patents, interest in further innovation declines until the basis of the dominance is threatened. Third, Parke Davis, one of the original innovators in the broad spectrum mar- ket, has never made any additional contributions. The observations from the narrow spectrum market suggests that where com- petition remains an active force innovation continues. And in comparing the approximate 30 percent of these innovations which were found to be significant with the 12 percent for the industry as a whole indicates that competition is a major determinate of technological progress. In the broad spectrum market where competitive forces were suppressed one finds a high proportion of the innovations significant. But to conclude from this observation that monopoly promotes prog- ress would be in error. First, the number of innovations were considerably fewer than in the narrow spectrum market. This in turn suggests that the agreement not to compete on price included an agreement not to compete in terms of product innovation as well. Second, an examination of the discovery of the broad spec- trum drugs indicates that associated with the establishment of monopoly is a decline in research within the therapeutic area. And only when the basis of the monopoly is threatened by competition does research resume. APPENDIX IT-PAPER, "RIsK AND CORPORATE RATES OF RETURN," BY I. N. FISHER AND G. R. HALL,' THE RAND CORPORATION, SANTA MONICA, CALIF. (Cited to text, p. 1820, supra.) I. INTRODUCTION Although economists have great interest in the correlation between risk and profits, few studies have attempted to quantify the relationship.2 Consequently, this paper considers the concept of risk differentials in corporate profit and pro- poses a model for measuring them. Using this model, the risk-rate of return re- lationship was estimated for a sample of firms in various industry groups. For each industry group, average risk-adjusted rates of return were also obtained.3 Risk is defined as the inability to predict the outcome of a forthcoming event with complete certainty. Entrepreneurs are viewed as making decisions in the face of uncertaintyly on the basis of probabilistic expectations about future out- comes.4 If certaintly is a situation where the entrepreneur's anticipation will as- suredly be fulfilled, then uncertainty can be measured by the likelihood that the actual outcome will differ from the anticipated outcome. The foregoing definition accords with economic models of risk (Refs. 6, 23, 24), and it suggests studying risk by examining distributions of corporate rates of return. Specifically, this approach intimates that the risk-rate of return rela- tionship can be analyzed statistically in terms of the relationship between the mean rate of return and higher moments of the distribution.5 15 American Cyanamid Company, Proposed Findings of Fact, Before the Federal Trade Commission, Docket No. 7211, 1960. Vol. I, p. 16. 1 Any views expressed in this paper are those of the authors. They should not be in- terpreted as reflecting the views of The RAND Corporation or the official opinion or policy of any of its governmental or private research sponsors. Papers are reproduced by The RAND Corporation as a courtesy to members of its staff. This paper was prepared for presentation at the 1967 Annual Meeting of the Econometric Society, December 27-30, Washington, D.C. Helpful comments and suggestions were provided by F. D. Arditti, B. A. Greenberg, R. E. Johnson, L. J. Kleiger, M. H. Kosters, A. Phillips and I. H. Plotkin. 2 Examples of the wide and diverse theoretical literature on risk and profits are Ref s. 6, 8, 11, 14, 17, 18, 20, 27, 32, and 34. Two important empirical studies of risk and rates of return are Refs. 4 and 30. In this paper "risk" will not be distinguished from "uncertainty." Compare Refs. 4, 6. 9. 11, 18, and 34. The 88 firms included in the sample were selected from Fortune magazine's list of the 500 largest industrial firms. Eleven industry groups were selected, comprising a variety of different types of industries. Firms in each group were chosen so that both middle- sized, as well as large, firms would be included. Economic theory contains two approaches to this problem: in one approach, the decision- maker balances the various moments of the probability distribution of potential outcomes on the basis of his utility function (Refs. 21 and 28), while in the other, the decision- maker chooses among a set of dated financial claims defined over all future states-of-the- world (Refs. 13, 14). We adopted the first, since the data do not justify the more elegant approach. It is assumed that the present and future sets of profit-generating opportunities for each firm are determined exogenously. That is, a firm may select opportunities but cannot influence the composition of any set of potential investments. Without this assumption, the concept of risk becomes more complex. If firms can influence the investment-choice set, however, presumably the observed variance of profits would decrease over time. Thus, concentration on fairly long periods of time and a large sample of firms should lessen the likelihood of this possible effect biasing the statistical results. PAGENO="0571" COMPETITIVE PROBLEMS IN THE DRUG INDUSTRY 2121 II. THE MODEL Assume that firms maximize not profits, but expected utility, and let U (P+W) be the firm's utility function.6 Utility is a function only of earnings, P (a random variable), and net worth, W. The risk premium, R(P,W), is that amount required to make the entrepreneur indifferent between the expected value of the uncertain earnings, E (P+W), and the certain amount E (P+W) -R(P.W), corresponding to the expected utility of the uncertain earnings (Ref s. 8, 26). Earnings distributions and utility functions are not important per se; it is their interaction that determines the risk component of profits. Suppose that both the probability distribution of potential earnings and the firm's utility function are known (illustrated in Fig 1(b) for a riskHaverse firm).7 Assume the probablity distribution is curve (1). Both the probability distribution of utilty, shown as (1) in (a), and its expected value, E(TJ1) are easily derived. Note that, although the probability distribution of earnings is symmetric about the expected value, E (P), the distribution of utilities is skewed to the left. This occurs because the utility function is concave, resulting in a non-linear transformation from earn- ings into utility. The expected value of the utility distribution, E (U1), is less than the utility of expected earnings, U (E (P+W)), and the difference, trans- lated into monetary terms, is the risk premium E (P) -P~. Now ~uppose that the probability distribution is not curve (1) but curve (2). This distribution is also symmetric about the mean, but the variance is larger. The distribution of utilities is curve (2) in (a) and, as before, it is not symmetric 6 Here P refers to profits in the sense of increments to net worth, rather than the profit rate. The question of "whose" utility function is moot. There are various candidates, e.g., managers, stockholders, the chief executive officer, as well as others (Ref. 10). It is assumed that each entrepreneur (management) is interested in maximizing the expected' utility of the net worth of the firm on the basis of his judgments about stockholders' preferences. Such an assumption permits us to explore the relationship of uncertainty to earnings without having to deal with the complexities of adding stockholders' utility functions, or the Modigliani-Miller view that stockholders can lever portfolios to offset corporate management decisions about risk (Refs. 22, 24. See also Ref. 6). If the firm is averse toward risk, the utility function is concave. This requires that U' >0 and U' `<0, or that utility increase with earnings and net worth, but at a decreasing rate. Fig. 1-Effect of dispersion and skewness on risk premium PAGENO="0572" 2122 COMPETITIVE PROBLEMS IN THE DRUG INDUSTRY about its mean, E (U2). The important point, however, is that the expected utility has decreased as a result of the increased dispersion of the earnings distribution. As a result, the risk premium, (E (P) -P) > (E(P) -P°). Con- sequently, greater variance in the distribution of earnings implies greater risk and, for risk-averse firms, leads to larger risk premiums. This implies that earn- ings should be larger, on average, for firms with greater variation in their earn- ings than for firms with little earnings variability. Skewness may also have an important effect on the risk premium (Refs. 3, 12, 31). The entrepreneur may prefer positively-skewed earnings distributions because the likelihood of extremely low earnings is smaller. This, also, is illus- trated in Fig. 1. Curve (3) in (b) has the same expected value as (1) and (2), but is skew-ed to the right. This function has been constructed so that the result- ing distribution of utilities is symmetric about its expected value. In this ex- ample, skewness offsets variance and the risk premium is zero, i.e., E (Us) U (E (P +W)). Thus, positive skewness results in smaller risk exposure, w-hile negative skewness leads .to greater risk exposure, implying that earnings should be smaller, on average, for firms with earnings distributions positively skewed but larger, on average, for firms with negatively-skewed distributions. The results of Fig. 1 suggest that once the form of the utility function is specified, risk exposure can be measured by characteristics of the probability dis- tribution of earnings. The required risk premium becomes larger as the spread of the earnings distribution increases, but the premium decreases as the dis- tribution becomes positively skewed.8 This illustrates that risk exposure, as de- fined here, can be measured by characteristics of the firm's earnings distribution. Before testing this hypothesis, one link in the discussion of the relationship between risk and earnings remains to be completed-that of the mechanism by which entrepreneurial preferences for risk and profits are translated into in- dustry profit differentials or risk premiums and discounts. Conventional economic theory indicates that with well-functioning capital markets the equilibrium rate of return will be identical among all activities. En- trepreneurs theoretically seek those investments yielding the largest rates of return. As capital is w-ithdrawn from less profitable activities, the rates of re- turn in such activities rise. Similarly, the inflow of capital into higher-yield investments forces the rates of return in these activities downward. Equilibrium occurs when the rates of return of investment are identical among all activities. When risk is considered, the adjustment process is more complex. Because differences in risk exposure exist among alternative investments, entrepreneurs balance risk against expected rates of return. Capital, therefore, is transferred from low-return, high risk activities to high-return, low risk investments until an equilibrium, characterized by a set of risk premiums reflecting differences in risk exposure, is achieved. In this equilibrium, risk-compensated rates of re- turn are equal among alternative investments, but observed or actual rates of return will differ by the amount of the risk premiums. 8 This can be demonstrated formally in the following manner: Expand U(P+JV) in a Taylor series A about the point (P± W) =-E(P± W), A A A A A fT" A A U(P+W)= tJ(P+JF)+U'(P+JV)(P-p)±-1 (P± W)(P_P)+-i-~.~ (P+ W) (P-P+ A Taking expected values and holding W, P constant, A U" A U" A E(U(P+JV))= U(P±JV)±~ -~- (P+JV)±~ -~- (P+JV)+ Rearranging terms, the difference between expected utility and utility of expected earnings Is A / U" A U'/A ~ (P±W)±o~ -~- (P+JV)+ Equation (3) is the risk premium, R(P, JV), and it becomes apparent that the second, third, and higher moments may affect the magnitude of the risk premium. Since U"0 and the risk premium becomes smaller as skewness increasesJligher moments add little information about the characteristics of the distribution and are ignored. (See references 3, 26). PAGENO="0573" COMPETITIVE PROBLEMS IN ~FHE DRUG INDUSTRY 2123 In short, we posit that capital markets respond to risk as they respond to expected rates of return. We should, therefore, expect to find a structure of risk- compensated rates of return that motivate or discourage investment. Part of the earnings differentials observed among alternative investments can be attributed to risk; these are the risk premiums that compensate for differences in risk exposure. III. EMPIRICAL RESULTS To test the hypothesis that profits are larger for firms with greater risk ex- posure, it is necessary to translate the theoretical definition of risk into statis- tical terms.° We can do this by assuming managers' anticipations, on average, are correct, thereby permitting the observed mean rate of return to be used as a proxy (Ref. 4). Risk exposure, as defined here, can then be measured by mo- ments of the distribution of earnings.10 The risk variables were calculated from r 2lh/2 I ~ ~r~-r~,) I o1L6T1 ] (1) and (r~i-~~,)~ ~ no-13 (2) where r~t= observed rate of return for firm i in year t; ~13=predicted rate of return from trend for firm i, year t; oi standard deviation of rates of return about trend, firm i; 81=skewness about trend for firm i; and n is the number of years included in the sample. The model can now be stated explicitly as ~1=ro+bio1+b2S1 (3) where ~= average rate of return on net worth for firm i; r0= intercept; and b1, b2 are the coefficients of the standard deviation and skewness, respectively-the risk coefficients. The signs of these coefficients are expected to be b1>O, b2