PAGENO="0001" 86th Con~ress} ~TOINT COMMITTEE PRINT STUDY PAPERS NOS. 4 AND 5 ANALYSIS OF THE RISING COSTS OF PUBLIC EDUCATION BY Werner Z. Hirsch TRENDS IN THE SUPPLY AND DEMAND OF MEDICAL CARE BY Markley Roberts MATERIALS PREPARED IN CONNECTION WITH THE STUDY OF EMPLOYMENT, GROWTH, AND PRICE LEVELS FOR CONSIDERATION BY THE JOINT ECONOMIC COMMITTEE CONGRESS OF THE UNITED STATES L L ~ ~. 1iistit~~ o~ W~;Tioemen~ NOVEMBER 10, 1959 L.thor RoI~H~s d for the use of the Joint Econom(h~~I2I&ti~6O UNITED STA?ESJTGER~) UN VERS IT'( GOVERNMENT PRINTING OFFICE WASHINGTON: 1959 F r sale by the Superintendent of Documents, U.S. Government Printing Office ~/ ~~1ngton25,D.O. .Price 30 cents 47761 PAGENO="0002" JOINT ECONOMIC COMMITTEE (Created pursuant to sec. 5(a) of Public Law 304, 79th Cong.) PAUL H. DOUGLAS, Illinois, Chairman WRIGHT PATMAN, Texas, Vice Ghairman HOUSE OF REPRESENTATIVES RICHARD BOLLING, Missouri HALE BOGGS, Louisiana HENRY S. REUSS, Wisconsin FRANK M. COFFIN, Maine THOMAS B. CURTIS, Missouri CLARENCE E. KILBURN, New York WILLIAM B. WIDNALL, New Jersey STUDY OF EMPLOYMENT, GROWTH, AND PRICE LEVELS (Pursuant to S. Con. Res. 13, 86th Cong., 1st sess.) O~o ECESTEIN, Technical Director JOHN W. LEHMAN, Administrative Officer JAMES W. KNOWLES, Special Economic counsel SENATE JOHN SPARKMAN, Alabama J. WILLIAM FULBRIGHT, Arkansas JOSEPH C. O'MAHONEY, Wyoming JOHN F. KENNEDY, Massachusetts PRESCOTT BUSH, Connecticut JOHN MARSHALL BUTLER, Maryland JACOB K. JAVITS, New York II PAGENO="0003" These are part of a series of papers being prepared for con- sideration by the Joint Economic Committee in connection with their "Study of Employment, Growth, and Price Levels." The committee and the committee staff neither approve nor disapprove of the findings of the individual authors. The findings are being presented in this form to obtain the widest possible comment before the committee prepares its report. m PAGENO="0004" PAGENO="0005" LETTERS OF TRANSMITTAL NOVEMBER 6, 1959. To Members of the Joint Economic Committee: Submitted herewith for the consideration of the members of the Joint Economic Committee and others are study papers 4 and 5, "Analysis of the Rising Costs of Public Education" and "Trends in the Supply and Demand of Medical Care." These are among the number of subj ects which the Joint Economic Committee has requested individual scholars to examine and report on to provide factual and analytic materials for consideration in the preparation of the staff and committee reports for the "Study of Employment, Growth, and Price Levels." The papers are being printed and distributed not only for the use of the committee members but also to obtain the review and comment of other experts during the committee's consideration of the materials. The findings are entirely those of the authors, and the committee and the committee staff indicate neither approval nor disapproval by this publication. PAUL H. DOUGLAS, Chairman, Joint Economic Committee. NOVEMBER 3, 1959. Hon. PAUL H. DOUGLAS, Chairman, Joint Economic Committee, U.S. Senate, Washington, D.C. DEAR SENATOR DOUGLAS: Transmitted herewith are the fourth and fifth in a series of papers being prepared for the "Study of Employment, Growth, and Price Levels" by outside consultants and members of the staff. The authors of these papers are Werner Z. Hirsch, Washington University, St. Louis, Mo., and Markley Roberts, American University, Washington, D.C. All papers are presented as prepared by the authors, for considera- tion and comment by the committee and staff. Orro ECKSTEIN, Technical Director, Study of Employment, Growth, and Price Levels. V PAGENO="0006" PAGENO="0007" CONTENTS STUDY PAPER NO. 4, "ANALYSIS OF THE RISING COSTS OF PUBLIC EDUCATION," BY WERNER Z. HIRSCH Page Introduction-Statement of findings 1 Chapter 1. General summary 2 Introduction 2 State and local government sector 2 Costs of public education 3 Teachers' salaries-the overwhelming cost factor 4 Ability to afford education-income elasticity 5 The near term prospect 5 Some implications 6 Chapter 2. The State and local government sector 6 Introduction 6 Nature of data 6 An overview 8 Functional breakdown Current versus capital expenditures 13 Cost comparisons in constant dollars 13 Chapter 3. The public primary and secondary education sector 15 Introduction 15 Nature of data 15 Factors affecting expenditure level 16 Number of pupils in average daily attendance 17 Sociological characteristics of population 18 Age structure 18 Geographical distribution 18 Economic characteristics 19 Price level of goods and services bought by schools 19 Income level 20 Physical characteristics 21 Productivity of school system 21 Governmental characteristics 24 Variety, scope, and quality of education 25 Addition or deletion of services 25 Scope of services 26 Quality of services 26 Test of hypothesis 30 Education cost in constant terms 32 Income elasticity of public education 35 The 1960 and 1965 outlook 39 Appendix-Economies of scale 41 CHARTS Chart 1. Total current expenditures (plus debt service) for public primary and secondary education, selected years, 1900-1958 16 Chart 2. Current expenditures (plus debt service) for public primary and secondary education minus expenditures, for auxiliary services, selected years, 1900-1958 TABLES Table 1. State and local government expenditures, selected years, 1902-58- 8 Table 2. General expenditures of State and local government by function, selected years, 1902-58 10 Table 3. Percent of general expenditures by service function 11 Table 4. Current general expenditures of State and local government by function 12 VII PAGENO="0008" VIII CONTrE~7TS Table 5. Capital outlay general expenditure of State and local govern- Page mentbyfunction 12 Table 6. State and local government purchases of goods and services in current and constant (1954=100) dollars, 1929-57. 14 Table 7. Pupils in average daily attendance (ADA), selected years, 1900- 1958 17 Table 8. Public high school relative to total public enrollment, selected years, 1900-1958 18 Table 9. Urbanization of school population, selected years, 1900-195&__ 19 Table 10. Average annual salaries of public school teachers, selected years, 1900-1958 20 Table 11. Per capita personal income, selected years, 1900-1958 21 Table 12. Auxiliary school service expenditures, selected years, 1900-1958_ 25 Table 13. Length of school term, selected years, 1900-1958 25 Table 14. Total number of principals, superintendents, and consultants per 1,000 pupils in average daily attendance in public primary and sec- ondary schools, selected years, 1900-1958 29 Table 15. Total current expenditures (plus debt service) for public pri- mary and secondary education, selected years, 1900-1958 33 Table 16. Total current expenditure (plus debt service) for public pri- mary and secondary education minus auxiliary services, selected years. 1900-1958 34 Table 17. Income elasticities of select local public services, St. Louis city-county area, 1951-52 and 1954-55 38 STUDY PAPER NO. 3, "TRENDS IN THE SUPPLY AND DEMAND OF MEDICAL CARE," BY MARKLEY ROBERTS Page Introduction and summary 47 Chapter 1. Demand for medical care 49 Total demand 49 Health consciousness 49 Total spending 50 Private demand 52 Total personal consumption spending 52 Faniily medical care spending 55 Growth of prepayment 56 Health needs of older people 59 Public spending 60 Public support for health and medical care 60 Federal support for research 61 Demand for hospital services 62 Trend to hospitals for health care 62 Hospitalization insurance 62 Utilization by older people 64 Mental illness and chronic diseases 64 Chapter 2. Supply of medical care 65 Organization of medical services 65 Quality changes in medical care 66 Supply of medical personnel 67 Physicians (medical doctors) 67 Dentists 68 Xurses 68 Other health workers 71 Supply of hospital facilities 72 Increases and shortages 72 Types of hospitals 73 Distribution of hospitals 75 Hospital services and costs 76 Hospital construction 77 Chapter 3. The price index of medical care 78 Trends 78 Limitations 80 Chapter 4. A look ahead 83 Expanding demand for medical care 83 Inadequacies of supply 83 Continuing price inflation 84 Providing an adequate supply of medical care 84 PAGENO="0009" CONTENTS IX CHARTS Page Chart 1. Spending for health and medical care 51 Chart 2. Personal consumption spending (for medical care) 53 Chart 3. Growth of health insurance coverage 57 Chart 4. Supply of physicians, dentists, and nurses, in relation to popula- tion, with projections on the basis of currently predicted output of medi- cal, dental, and nursing needs 69 Chart 5. Ownership of hospital beds 74 Chart 6. Medical care price index and selected components 79 TABLES Table 1. Private and I)ublic spending for health and medical care, selected fiscal years, 1929-58 52 Table 2. Personal consumption spending for medical care, 1946-58 54 Table 3. Charges incurred by families for health services, by income and health insurance status 56 Table 4. Family outlays for medical care as a percentage of family income, by income and health insurance status 56 Table 5. Growth of health insurance coverage, number of people protected against hospital expenses, surgical expenses and regular medical expenses 58 Table 6. Health problems of people aged 65 and over, 1957-58 60 Table 7. Public spending for health and medical care, selected fiscal years, 1929-58 61 Table 8. Physicians, dentists, and nurses: Trends and projections of supply 70 Table 9. Ownership of hospital beds 74 Table 10. Hospital personnel and payroll costs 77 Table 11. Ilospital construction, value put in place and value by source of funds 78 Table 12. Relative importance of medical care price index components as percent of Consumer Price Index, all-items total 80 Table 13. Medical care price index, selected items and groups, 1947-~- June 1959 82 PAGENO="0010" PAGENO="0011" STUDY PAPER NO. 4 ANALYSIS OF THE RISING COSTS OF PUBLIC EDUCATION (BY~WERNER Z~ HIRSCH) XI PAGENO="0012" PAGENO="0013" STUDY PAPER NO. 4 ANALYSIS OF THE RISING COSTS OF PUBLIC EDUCATION (By Werner Z. Hirsch) INTRODUCTION-STATEMENT OF FINDINGS 1 A study of the relation of the public school sector to the problems of employment, growth, and price levels leads to the following conclu- sions: 1. The huge increases in current expenditure (plus debt service) for public primary and secondary education since the turn of the century are mainly due to exogenous forces, such as factors raising salary levels and per capita personal income. 2. With the property tax as the main revenue source, general price level increases, and particularly rises in teacher salary level, are likely to lead to continued fiscal school crises. 3. Since the turn of the century, the percentage increases in educa- tional expenditures-more specifically, daily per pupil current expendi- ture, plus debt service---were on the average about the same as the percentage increases in per capita personal income. The former exceeded the latter by a very small aniount. Such low income elas- ticity of public education must be of deep concern to all those who are convinced that improvements in education are essential if the United States is to remain a leading world power. Since more funds need to be found, serious consideration must be given to policies aimed at changing people's attitude toward public education, and to shifting partial responsibility for financing education from the local school district to the State and possibly to the Federal Government. This will require the combined leadership of all branches of govern- ment and the institution of matching Federal funds combined with such eligibility criteria as minimum tax effort and statewide equaliza- tion of educational opportunity. 4. It appears that there have been no significant productivity changes in the education sector. 5. During the last 10 years general expenditures of State and local governments increased an average of about 15 percent per annum measured in current dollars and 8 percent if measured in 1954 con- stant dollars. During the same period the average annual increase of total current expenditures, plus debt service, for public primary and secondary education was about 17 percent in current dollars and 5 percent in 1954 constant dollars. 1 The author would like to express his appreciation to Elbert Segelhorst, who served as research assistant, and to Profs. 1-Jarold Barnett and Burton Weisbrod for their numerous useful comments. Much assistance was also rendered by officials in the Departments of commerce; Health, Education and Welfare; and Labor. They are too numerous to be enumerated. PAGENO="0014" 2 ANALYSIS OF RISING COSTS OF PUBLIC EDUCATION 6. A "medium" projection suggests that if annual teachers' salary increases of 3 percent are granted-and funds to finance them can be found-total current expenditure, plus debt service, for public primary and secondary education wifi have advanced in terms of 1958 dollars to about $17.4 bfflion by 1965; i.e., a 58-percent increase in 7 years. However, if people's attitude toward public education and the way of financing it remain unchanged, a 3-percent annual mcrease in per capita personal income may produce merely $15.3 billion m revenue in 1965. While this would be an increase of about 39 percent over the 1958 figure, it would be about $2.1 bfflion below the earlier projected expenditure figure. CHAPTER 1. GENERAL SUMMARY INTRODUCTION This study is designed to shed light on the cost and expenditure level changes of the public primary and secondary education sector of our economy-a sector employing more than 2 million men and women, who serve almost 35 million pupils. This sector in 1958 spent more than 30 percent of all the money spent by State and local governments, an amount representing about 3.3 percent of national income and 3 percent of gross national product. In order to trace and understand changes in the cost of education and relate them to the rest of the State and local government sector, the latter sector is examined first in some detail. This sector, for example, accounts for about 12.3 percent of gross national product, and perhaps 3 to 3J~ percent of the Consumer Price Index of the U.S. Department of Labor. There are indications that since 1952 the cost of State and local government services increased up to three times more rapidly than did the other items included in the index. To shed light on the cost behavior of this sector, a fimctional service breakdown is attempted, and the question is raised: "Which State and local government services provided the major monetary push?" STATE AND LOCAL GOVERNMENT SECTOR In terms of general expenditures, education, highways, public wel- fare, and health and hospitals have been the four most important State and local government services since the turn of the century. Education expenditures have accounted for an increasingly large share of all expenditures. They increased from about 25 percent in 1902 to about 35 percent in 1958. General expenditures for public primary and secondary education advanced at a somewhat slower pace than did those for higher education. The relative importance of highway expenditures showed great fluctuations. During the first quarter of this century a steady advance from 17 to 25 percent was registered. From then on until the end of World War II a steady decline set in, which has been reversed only in recent years. Public welfare expenditures, at least in relative terms, grew more rapidly than any of the other main services. Most of their advance took place during the first half of the 1930's. Finally, the relative share of health and hospital expenditures has shown a PAGENO="0015" ANALYSIS OF RISING COSTS OF PUBLIC EDUCATION 3 very steady, though not rapid, increase. Many other services have lost in importance. Next, expenditures are examined from a somewhat different view- point. The issue at hand is, "Did current or capital expenditures provide the major push"? The answer can be given in unequivocal terms. So far in the 20th century, capital outlays of State and local government advanced at a rate more than twice as fast as did current expenditures. The 1902-58 increase of capital outlays is 76-fold, while that of current expenditures is 37-fold. This not only holds if all services are taken together, but it is also true for single services except public welfare, sanitation, and local parks. What is the picture of State and local government expenditures if they are adjusted for price level changes; i.e., in constant dollars? The U.S. Department of Commerce has prepared a series of implicit defiators~applicable to State and local government purchases of goods and services which goes back to 1929. This series does not make allowance for quality changes. While State and local government expenditures -measured in current dollars -increased from 1929 to 1957 by almost 370 percent, expressed in 1954 constant dollars these expenditures have not even quite doubled. Between 1929 and 1945 the constant dollar expenditures moved relatively little and without definite direction. Since them a pronounced and steady rise has started. The average annual increase of the last 10 years-measured in constant dollars-was about 8 percent. COSTS OF PUBLIC EDUCATION What about cost changes in the single most important State and local government sector, i.e., public primary and secondary education? More specifically, what forces have decisively contributed to cost increases in this sector? A large number of factors can be identified as possibly affecting current expenditures, plus debt service, for public primary and secondary education. Some of them have assumed greatly different values since the turn of the century. Thus, for example, the number of pupils in average daily attendance-ADA-increased almost three times~since 1900,~high school~enrollment relative to total public school enrollment increased about six times, and the percent of pupils living in urban as against rural America has increased about 40 percent. Turning to some economic factors, it is noteworthy that the average salary of a teacher advanced about 14 times, while per capita personal income increased about 9 times. Finally, an examination of some characteristics of public education itself reveals that while in 1900 virtually no appreciable auxiliary services were rendered, in 1958 pupils were fed in school cafeterias, attended by school health services, and brought to school in school- buses, etc. In 1958, almost 14 percent of current school expenditures, plus debt service, were applied to auxiliary services. In addition, the school term was lengthened an average of 60 percent and the number of principals, superintendents, and consultants per pupil almost doubled. PAGENO="0016" 4 ANALYSIS OF RISING COSTS OF P1JBLIC EDUCATION TEACHERS' SALARIES-THE OVERWHELMING COST FACTOR Which of these factors have decisively contributed to cost increases? How important a factor is the salary of the average teacher and how important is people's ability to afford good education, i.e., their income? Has the efficiency-or productivity-of the public school system changed much? These are some of the questions for which answers are sought by relying on multiple correlation and regression techniques. First an education unit was developed, one that appears to have been fairly uniform over the years. It is so standardized that its variety and scope are held reasonably constant and it is expressed in per pupil in average daily attendance terms. Ever since the turn of the century the costs of this education unit-daily per pupil current expenditures plus debt service minus the cost of auxiliary services-appear to have been overwhelmingly determined by the prevailing salary level of the instruction staff, and perhaps the prices of all goods and services bought by schools. However, they, in turn, are determined by the general demand and supply situation and price level of the rest of the economy, as well as the particular demand for and supply of teachers. There is some evidence that the salary of the average teacher has advanced somewhat more rapidly than the earning of other workers. This suggests that the specific demand for and supply of teachers play an important role in addition to general price level changes. An over- whelming percentage of the 1900-1958 variation of the cost of the standardized education unit can be explained, on the average, in terms of teacher salary level increases, holding constant the effect of changes in the public high school-all public school enrollment ratio, urbanization, and number of specialists per pupil. If now current expenditures plus debt service are deflated by teachers' salaries, the 1930-57 advance is about 54 percent. It is less than the 1930-57 advance in all State and local government costs in constant dollars, which was about 88 percent. The average annual increase in constant dollars was 2.0 percent for education and 3.3 per- cent for all State and local government costs. The overwhelming importance of teacher salary level in determining the cost of a public education unit raises some intriguing questions of finance, since much of the revenue of schools takes the form of property taxes. The general price level together with specific demand and supply factors for teachers can advance at a more rapid pace than the assessed valuation of real property. Not only may real property increase at a slower pace than the general price level but, more impor- tantly, reassessments are usually at best a slow and belated process. As long as the property tax is the main revenue source of schools, general price level increases, and particularly teacher salary level increases, are likely to create major fiscal problems. The importance of salary level changes can also be i]iustrated by comparing 1900-1958 total public school expenditiu~e (minus auxiliary expenditures) in current dollars with those in constant 1954 dollars. In current dollars the 1900-1958 increase was 48-fold, while in con- stant dollars the increase was merely 23~-fold. If further adjustments are made for the number of pupils in average daily attendance and length of school term, expenditures in real terms are approximated. Expenditures in real terms exhibit amazing stability during 1900-1958. For the years for which data are available, 1922 was the low year with PAGENO="0017" ANALYSIS OF RISING COSTS OF PUBLIC EDUCATION 5 $1.37 daily expenditure per pupil, and 1913 was the high year with $1.60. Over these 58 years, an overall decline of about 3 percent was registered. To the extent that the standardization of the education unit has been successful and appropriate data in general were used, the analysis can reflect on the productivity variable as a residual. Apparently, productivity in the public schools has changed very little, if at all. ABiLITY TO AFFORD EDucATIoN-INcOME ELASTICITY Next the effect of per capita personal income on the cost of daily per pupil current expenditure plus debt service is considered. Once more multiple correlation and regression techniques are em- ployed. They point to the importance of income as another deter- minant. Most likely, income's main impact is twofold-it affects the amount of revenue that is raised for public education as well as its variety and quality. About 76 percent of the 1900-1958 varia- tion in the cost of daily per pupil current expenditure plus debt service can be explained on the average in terms of per capita per- sonal income changes, holding the effect of changes in the public high school-all public school enrollment ratio and urbanization constant. It must be remembered that these 76 percent, and the change associated with salary level, are not additive. Income elasticity of public education was estimated to have aver- aged plus 1.09 during 1900-1958. Thus, during 1900-1958 a 1 per- cent increase in per capita personal income was on the average asso- ciated with a 1.09 percent increase in daily total current expenditures for public primary and secondary education per pupil in average daily attendance. It is only slightly above unit elastic and apparently lower than that of some other public services, iiot to speak of such consumer amenities as air conditioning, automobiles, golf, speed- boats, etc. Such low elasticity must be of concern to all those who are convinced that improvements in public education are essential if the United States is to remain a leading world power, and that there~ fore an increasing portion of the American people's income must be channeled into public education. Certainly, people's attitude toward education and our general tax system are mainly responsible for the prevailing income elasticity of education. If public education is to he improved in the United States and more funds to finance education are to be found, serious consideration must be given to changing both, including further shifts in the responsibility for financing education from the local school district to the State and possibly to the Federal Government. THE NEAR TERM PROSPECTS Making projections is at best a hazardous undertaking, particu- larly in a case where the phenomenon awaiting proj ection so greatly depends upon exogenous forces, such as personal income and teacher salary levels. I-low hazardous predictions in the field of education are is well demonstrated by recalling James B. Conant's 1938 pre- diction: By 1960 or thereabouts we shall have a stationary population. The expansive pressure on our schools will soon be gone.2 2 James B. Conant, "The Future of Our Higher Education," Harper's magazine, May 1938. 47761-59------2 PAGENO="0018" 6 ANALYSIS OF RISING COSTS OF PUBLIC EDUCATION Three basic sets of projections for 1960 and 1965 have been pre- pared in 1958 dollars-low, medium, and high. The medium projec- tions assume a cumulative annual increase of 3 percent in per capita personal income, teachers' salary and length of school term. The number of pupils in average daily attendance is assumed to increase to 32 million in 1960 and 37.2 million in 1965. The other factors are assumed to increase very little. On the basis of these assumptions, equation 1.1, which emphasizes teachers' salary level, can be used to project expenditures. Equation 1.2, which emphasizes people's income and thus ability to afford good education, helps project revenue. Medium projections suggest that if annual teachers' salary increases of 3 percent are granted-and funds to finance them can be found-total current expenditure plus debt service for public primary and secondary education will have advanced in terms of 1958 dollars to about $12.7 billion by 1960 and $17.4 billion by 1965, compared to $11.0 billion expenditures in 1958. This amounts to about an 8 percent annual increase. However, if people's attitude toward public education and the way of financing it remains unchanged, a. 3 percent annual increase in per capita personal income is projected to produce merely $11.1 billion in revenue (in terms of 1958 dollars) in 1960 and $15.3 billion in 1965. While this would be a. 39 percent increase from 1958 to 1965, it would fall short by $2.1 billion in 1965 of what teachers' salary pressure would lead to in terms of expenditure. `Whether the public education sector will in 1965 account for about $17.4 or $15.3 billion will much depend upon whether the income elasticity of public education will change. SOME IMPLICATIONS Leadership will play a decisive role. The American people need to he persuaded to spend a. larger part of their income for public education. Congress can make an important contribution by voting into existence a system of matching Federal funds combined with carefully designed eligibility criteria. High on the list of these criteria should be the following two: States must make a minimum tax effort to finance education-i.e., a minimum ratio between per capita State and local school tax and per capita income will have to be set; and there should be a statewide equalization program which effectively assures sufficient funds to underwrite a floor below which the scope and quality of education cannot fall anywhere in the State. CHAPTER 2. THE STATE AND LOCAL GOVERNMENT SECTOR INTRODUCTION While it might not be generally recognized, the number of State and local governments in the United States is not only fairly large- 102,352 units in 1957-but they also perform a great variety of functions. State and local government direct expenditure in 1958 was $53.9 billion, compared to $81.2 billion spent by the Federal Government. These expenditures amount to 12.3 percent of gross national product and 15 percent of national income. Or, to use another yardstick, State and local governments employed in 1957 a full-time equivalent of 4,793,000 men and women. PAGENO="0019" ANALYSIS OF RISING COSTS OF PUBLIC EDUCATION 7 NATURE OF DATA Because of the volatile nature of capital expenditures by State and local governments, for analytical purposes it would be desirable to let debt service reflect the capital expenditure of these govern- ments. Since these capital expenditures are usually amortized over a 20- to 50-year period through bond issues and sinking funds, debt service would more nearly reflect the current benefits accruing from these capital outlays. However, the U.S. Bureau of the Census, in its census of governments, excludes payments for debt retirement, extension of loans, and purchases of securities from expenditure. Since the U.S. Bureau of the Census is the basic source for many State and local government expenditure data, the concept of debt service as a reflection of capital expenditure cannot be used in con- nection with census data. There are three general functional classifications of State and local government expenditure: general expenditure, utility and liquor stores expenditure, and insurance trust expenditure. Included in general expenditure are the services which are commonly associated with State and local government-education, highways, welfare, health and hospitals, police, fire, sanitation, etc. In 1958, these general expenditures accounted for $45.1 billion or 83.7 percent of total expenditures. Utility and liquor store expenditures in 1958 were $4.6 billion or 8.6 percent of total expenditures and include expenditure on such government-owned utilities as water supply, electric power, gas supply, and transit systems. Included in this category are expenditures for running State-owned liquor stores in 16 States and those owned by local governments in a few States. Other commercial-type operations of State and local government are included in general government. State and local government insur- ance trust expenditure consists chiefly of retirement payments for State and local government employees and unemployment compensa- tion payments to the qualified unemployed, who together with their employers, have contributed to the fund from which payments are made. In 1958, insurance trust expenditures were $4.2 billion and accounted for 7.7 percent of total State and local government expendi- tures. It might be helpful to ignore liquor store and insurance trust expenditures. Unemployment compensation, which in 1958 amounted to $2.8 billion or 66.2 percent of insurance trust expenditure, is not generally considered to be a State and local government expenditure since both revenue is received and payments are made by this separate insurance trust fund. The administrative cost of this insurance trust fund is, however, included in general expenditure. One must keep in mind, then, that the U.S. Bureau of the Census uses the following basic concepts and terminology in its statistics on expenditure of State and local government: 1. State and local government expenditure comprises all amounts of money paid out as between State and local governments and external individuals or agencies (net of correcting transactions such as recoveries or refunds), with the exception of amounts for debt issuance and retirement and for loan and investment, agency, and private trust transactions. PAGENO="0020" S ANALYSIS OF RIS~G COSTS OF PUBLIC EDUCATION 2. Although State and local governments act as agents of the Federal Government in withholding Federal income and social security taxes from their employees' pay, such amounts are excluded from State and local government revenue and expenditure, and are reported as Federal revenue and expenditure. 3. Although the Federal Treasury handles unemployment compen- sation accounts for the States, these. funds are omitted froni the Federal figures and are included with the State and local data. 4. The data. utilized for each individual government represent a consolidation of amounts for its various funds, and payments between funds have beeii eliminated for census reporting. 5. Intergovernmental expenditure has been netted out of aggregates comprismg the groups of governments concerned. Miost of the amounts so classified comprise fiscal aid in the form of Federal and State grants. The value of intergovernmental aid "in kind," such as commodities distributed by the Federal Government for school-lunch purposes, is not trea ted as intergovernmental revenue or exp encliture. Contributions by local governments to State-administered retirement systems that cover their employees are included as part of tile "current operation" expenditure of the. local governments involved, and are included with State insurance trust revenue. TABLE 1.-State a ad local government expen ditures, selected yeas's, 1902-58 [In millions of dollars] Year General expendi- tare (1) General expendi- ture plus employee retirement (2) Total expendi- tare (3) Year General expendi- ture (1) General expendi- ture plus employee retirement (2) Total expendi- ture (3) 1902 1913 1922 1927 1932 1934 1936 1938. 1940 1942 1944 1,013 2, 064 5, 218 7,210 7,765 7,181 7,644 8,757 9,229 9,190 8,863 1,013 2. 071 5, 24.8 7,760 7,840 7,277 7,757 8,886 9.369 9,359 9,05S 1095 2, 257 5. 652 7.810 8.403 7,S42 8,501 9, 98S 11,240 10.914 10, 499 1946 1948 1950 1952 1953 1954 1955 1956 1957 1958 11,023 17, 684 22, 787 26,098 27,910 30,701 33,724 36, 711 40,375 45,059 11,265 17, 981 23, 14S 26,628 28,495 31,380 34,446 37, 536 41,318 46,132 14,067 21, 260 27, 905 30,863 32,937 36,607 40,375 43, 152 47,553 53,857 Source: U.S. Bureau of the Census. "Historical Summary of Governmental Finances in the united States," 1959, 23 pp. and "Summary of Governmental Finances in 1958," 1959, p. 14. AX OVERVIEW As can be seen from table 1, during the last 56 years or so tile annual general expenditure of Stat.e a.nd local governments increased from $1 to $45.1 billion, about a 44-fold increase.3 Furthermore, total government expenditure. i.e., general expenditure plus utility and liquors expenditure plus insuranc.e trust expenditure, advanced from $1.1 to $53.9 billion during this period, that is, about 48-fold. But, as would be expected, the increase was far from steady. In rough terms, the rate of increase was very high from the turn of the century until the depression of the 1930's-about an average of 25 `Actually, although the census definition does not do so, iris appropriate to include in the general expendi- tare figure government contributions to employees' retirement fund. If this is done, the increase is some- what higher. PAGENO="0021" ANALYSIS OF RISING COSTS OF PUBLIC EDUCATION 9 iercent per year. Depression and World War II greatly changed the picture. In 1944, general State and local government expenditure was only $1.1 billion or about 15 percent higher than it had been 12 years earlier. But with the end of World War II, a very high average annual rate of advance was resumed, i.e., about 24 percent a year. The rate was slightly higlier in the forties than in the fifties of this 12-year period. No doubt the heavy backlog of demand from the war years was responsible for the exceptionally fast increase in the late forties. FUNCTIONAL BREAKDOWN What services are mainly responsible for this rapid advance in State and local government general expenditure? To supply answers to this queation, the general expenditure figures have been broken down by 16 main categories; 1902-58 data are given in table 2. PAGENO="0022" 10 ANALYSIS OF RISING COSTS OF PUBLIC EDUCATION c~. rjí3~ ~ cç~ C).Z~ ~ ~ ~`-~ ~-~t ~ ~o ~ t~ t C~ - N- N- N- C ~ C `-` N- ~- C'~ ~ - ~ C ~ C'1 - N- C `~- C N- N- -~ C C - C C C -~ C - N- -~ C C C C - C C C ~ ~- ~- r~- - ~ ,-~ ~;` - ~ C C - CN-N-N-CCC-~~~~ C~CC -CC~-~ C -~- ~s &tc ----C--- z ~ : :~:~ C ~ C c~ ~3 -CCC~'-~ ~ ~C ~ C - - - C - C C - C - - ~- -~- - -~-~-~-~ -~ ~ -~ ~- -~:~ - - CC ~ -- -C~ `- C -~ C C C C N~ C N- N- C C~ C ~ ~C-N:C- £ ~ ~ C N- ~ N- C C C N- N- ,-~ C C ~ C N- ~ - N- N~ - -~ N~ C C C CCCCCCCN-CCC~C~C ~C C C'~_~ t- C -~ C C NC Ct- NC Cl `N C C C NC~- C CCC~~N-CCNCC~~CN-C - - ~ Cl £ ~~CCC C0 ` PAGENO="0023" ANALYSIS 01? RISING COSTS 01? PUBLIC EDUCATION 11 Moneywise, the most important functions are education, highways, public welfare, and health and hospitals, in this order. As can be seen from table 3, education expenditures are responsible for an increasingly large part of all expenditures. Their share increased from about 25 percent in 1902 to 33 percent in 1922; then a gradual decline took place to about 26 percent in 1934. From that time on, education has accounted for an increasingly large share of all expendi- tures, reaching a peak of 36 percent in 1956. TABLE 3.-Percent of general expenditures by service function Year Education (1) Primary and second- ary cdii- cation (2) Highways (3) Public welfare (4) Health and hospitals (5) Other services (6) 1902 1913 1922 1927 1932 1934 1936 1938 1940 1942 1944 1946 1948 1950 1952 1953 1954 1955 1956 1957 1958 25. 2 28.0 32. 7 31.0 29. 8 25. 5 28. 5 28.4 28.6 28. 1 31.5 30.4 30.4 31. 5 31.9 33.6 34.4 35.3 36.0 35.0 35.0 23. 5 25.3 29. 5 28.0 26.3 22. 5 24. 8 24. 7 24.7 24. 1 26.3 26. 0 24. 6 25. 8 26.5 28. 2 29.3 30. 2 30.6 29.3 1 29.3 17. 3 20.3 24. 8 25. 1 22.4 21. 0 18. 6 18.8 17.0 16. 2 13.5 15. 2 17.2 16. 7 17.8 17. 9 18.0 19. 1 18. 9 19.3 19. 3 3. 7 2.5 2. 3 2. 1 5. 7 12.4 10.8 12. 2 12.5 13.3 12.8 12. 8 11. 9 12. 9 10.7 10.4 10. 0 9.4 8. 6 8.4 8.3 5. 9 5.2 4. 9 4. 9 5.9 5. 8 6. 1 6.3 6.6 6.4 7.4 7.4 6. 9 7. 7 8.4 8. 2 7. 8 7. 5 7. 6 7. 9 8. 0 47. 9 44.0 35. 3 36. 9 36.2 35.3 36.0 34.3 35.3 36.0 34.8 34. 2 33. 6 31. 2 31.2 29. 9 29. 8 28. 7 28. 9 29.4 29.4 1 Estimate. Source: U.S. Bureau of the Census. General expenditure for primary and secondary education behaved about the same way as did that for all education. Yet, its share has grown less rapidly during 1902-58 than has that of all education. For instance, between 1902 and 1956 the share of all education increased by about 40 percent while that of primary and secondary education rose by only 30 percent. Thus, higher education expenditures have increased more rapidly than the rest. The relative share of highway expenditures showed great fluctua- tions but no clear trend over the years. Until 1927 their share had steadily increased to reach an all time high of 25 percent. From then on, it rapidly declined and by 1944 it had lost almost half its earlier importance. The postwar period has shown a slow but steady advance to 19 percent in 1958. During 1902-58, public welfare expenditures grew relatively more rapidly than those of any other major State and local government service. But most of the advance took place during the early days of the Franklin D. Roosevelt administration, when their share jumped from about 2 percent in 1927 to 6 percent in 1932 and 12 percent in 1934. From then on, speaking in rough terms, a decline set in, leading to a low of 8.3 percent in 1958. PAGENO="0024" 12 ANALYSIS OF RISING COSTS OF PUBLIC EDUCATION Except for the 1920's the relative share of health and hospital expenditures has been increasing, although the pace was not fast. If all remaining services are lumped together, their relative impor- tance shows a steady decline from 48 percent in 1902 to 29.4 percent in 1958. Nevertheless, such services as housing and community redevelopment, natural resources development, etc., showed a. dis- tinctly different trend. TABLE 4-Current general expenditures of State and local government by function. [In millions of dollars] Year c ~ ~ ~ (1) (2) `1~.s ~° (3) - .~ ~ ~4) ` ~ (5) `0 ~ .~ ~ (6) ,~ ~ ~il ~ (7) (8) ~? I~ (9) 1902 1913 1922 1932 1942 1957 1958 1902-58 change, percent 214 48.8 1,374 2, 064 2,411 10,8S2 12,293 116 220 508 719 807 2,605 2,798 58 91 226 370 540 12,300 3,20S 27 51 105 135 191 2799 2840 9 14 61 137 182 760 814 12 22 52 110 110 429 480 25 222 235 414 723 1,508 2, 521 3.602 11,092 11,846 850 1, 609 3,834 6. 056 7,868 29 0S9 32,514 5, 600 2, 300 5, 400 3, 000 8, 900 3,900 2,800 3,700 Hospitals only. 2 Sewers and sewage disposal only. Source: Special computations in cooperation with the Data Classification and Research Branch of the U.S. Bureau of the Census. TABLE 5.-Capital outlay general expenditure of State and local government by function [In millions of dollars] . Year .~ ~ ,~ S - ~ g .~ .- ~ ~, ~ ~ ~ ,-~-- ~ S :~ u 0 Co `~ ~ ,, z - ~ ~ n5 `~ll o~' ~i ~i~i ~: ~ .552 ~0a ~ ~ ~. ,~ o - -`I L~ ~S d~ ~ ~n E~ (1) (2) (3) (4) (5) (6) (7) (8) (9) 1902 41 59 2 24 (1) 17 20 163 1913 89 199 17 46 (1) 35 69 455 1922 331 786 32 84 (1) 33 118 1,384 1932 247 1,022 86 8.8 28 37 201 1,709 1942 175 683 51 38 32 18 211 114 1,322 1957 3,252 5,211 2348 ~644 271 156 283 1,121 11,286 1958 3, 489 5, 896 2417 3 667 307 205 366 1, 198 12, 545 1902-58 change, percent 8,400 9, 9i0 20,700 2,700 1, 100 5,900 7,6o0 1 Not available. 2 Hospitals only. 3 Sewers and sewage disposal only. Source: Special computations in cooperation with the Data Classification and Research Branch of the U.S. Bureau of the Census. PAGENO="0025" ANALYSIS OF RISING COSTS OF PUBLIC EDUCATION 13 CURRENT VERSUS CAPITAL EXPENDITURES It is mainly current expenditure, or mainly capital outlay that is driving up the general expenditure of State and local governments? And if one type of expenditure had advanced more rapidly than the other, what services are responsible for it? Here is some of the evidence. While total general expenditure of State and local governments increased about 44-fold during 1902-58, the current elements advanced about 37-fold and the capital elements 76-fold. (See tables 4 and 5.) Capital outlays of the three important services-education, high- ways, and health and hospitals-which throughout this period ac- counted for between one-half and two-thirds of general expenditure, advanced at a much more rapid pace than did current expenditures. The 1902-58 increases for education were 84-fold versus 56-fold, for highways 99-fold versus 23-fold, and for health and hospitals 207-fold versus 54-fold. Sanitation and local parks and recreation behaved in an opposite manner. For natural resources only data since 1932 are available. Housing and community redevelopment is a relatively new function and not enough evidence is on hand to reach a conclusion. But capital outlays of all other services increased about two and a half times as fast as did their current expenditure. In summary, much of the rapid increase in the general expenditure of State and local government since the turn of the century can, no doubt, be traced to the capital outlay elements of education, highways, and health and hospitals. COST COMPARISONS IN CONSTANT DOLLARS An analysis of State and local government general expenditure has revealed major increases. But the entire analysis was in terms of current dollars, in many respects an imperfect yardstick, since tile value of the dollar changed substantially during this period. In recognition of this fact the question will now be asked, "How great, if any, were the cost changes of a given bundle of State and local government services during this period?" To answer this question, a deflation procedure must be used so that a cost comparison in real terms becomes possible. Basically, there are two approaches to the problem of deflating State and local government purchases. As George Cobren has so aptly put it- In the case of Government purchases, a basic dichotomy exists among natiocial income theorists as to the proper approach to the deflation problem. On the one side it is argued that the deflation procedure should focus on the products which the Government buys; on the other, that the procedure should measure the volume of services which the Government provides.4 He goes on to point out that the U.S. Department of Commerce's implicit deflators for the Government sector of gross national product follows the first of these alternatives. With the help of the implicit deflator applicable to the State and local government sector, the value of the goods and services purchased by these governments can be deflated. Apparently this alternative was selected by the Department 4 George MI. Cobren, "The Deflation of the Gross National Product by the Department of Commerce," American Statistical Association proceedings, business and statistics section, 1958, pp. 312-319. PAGENO="0026" 14 ANALYSIS OF RISING COSTS OF PUBLIC EDUCATION of Commerce, since while it presents formidable difficulties, the other is even more hazardous. One of the main shortcomings of the deflating procedure followed by the U.S. Department of Commerce stems from the fact that by far the largest expenditure item-employee compensation-is derived * * by extrapolating the base-year estimate by man-hours wher- ever possible, and by employment when man-hours were not avail- able. * * ~" This, admittedly, is equivalent to deflating the current value figures by indexes of average compensation. The result is a measure which makes no allowance for quality changes-in this case, for changes in the productivity of Government employees. On this point, Richard Ruggles and Nancy D. Ruggles made some revealing observations in their testimony before the Joint Economic Committee.6 They are greatly concerned about the fact that the implicit price deflators do not properly allow for quality and efficiency improvements. Consequently. they argue that almost every category of expenditure has an upward bias. In table 6 purchases by State and local governments, both in actual and constant dollars (1954= 100), are presei~ted for the period 1929-58. Money expenditures increased from $7.2 billion to $40.5 billion, i.e., slightly more than five times. In constant dollars the increase was much less, i.e., the 1958 figure is only a. little more than double that of 1929. Expenditures in constant dollars zigzagged from 1929 to 1945. But with the end of World War 11 they began a. pronounced rise which as yet shows no sign of reversal. The average annual cost increase in constant 1954 dollars during the last 10 years was about 8 percent. TABLE 6.-State and local government purchaser of goods and services in current and constant dollars (1954=100), 1929-57 un billions of dollars] Year Current dollars (1) Constant dollars (1954=100) (2) Year Current dollars (1) Constant dollars (1954=100) (2) 1929 1930 1931 1932 1933 1934 1935 1936 1937_ 1938 1939 1940 1941 1942 1943 7.1 7.8 7.7 6.6 6.0 6.8 7. 1 7. 0 7.2 7.5 S.2 7.9 7.8 7.7 7.4 15.6 17.1 17.9 16.6 14.6 15.8 16.3 16. 6 16.4 17.4 19.1 18.0 16.9 15.4 14.0 1944 1945 1946 1947 194S 1949 1950 1951 1952 1953 1954 1955 1956 1957 1958 7.5 8.1 9.9 12.7 15.2 17.9 19. 7 21. 7 23.2 24.9 27.7 30.3 33.2 36.8 40.5 13.8 14.0 15.8 17.8 19.2 21.9 23.5 24. 1 24.5 25.5 27.7 29.7 30.6 32.3 34.4 Source: U.S. Department of Commerce, "U.S. Income and Output," 1958, pp. 118-119. `IbId., p.316. 6 Richard Ruggles and Nancy D. Ruggles, "Prices, Costs, Demand, and Output in the United States, 1947-57," the relationship of prices to economic stability and growth. compendium of papers submitted by panelists appearing before the Joint Economic Committee (Mar. 31, 1958), p. 299. PAGENO="0027" ANALYSIS OF RISING COSTS OF PUBLIC EDUCATION 15 CHAPTER 3. THE PUBLIC PRIMARY AND SECONDARY EDUCATION SECTOR INTRODUCTION As was shown in the preceding section, public expenditures for primary and secondary education are by far greater than those for any other single service performed by State and local governments. Of the 1957 general expenditure of $40.4 billion, almost 30 percent were spent for public primary and secondary education; about 3.3 percent of national income. Education expenditures had increased much faster since the turn of the century than the other State and local expenditures. The pivotal position and importance of public education for the future of this Nation is all too clear and needs no elaboration. NATURE OF DATA A major portion of this study is concerned with pricing educational services in constant terms. In this connection it is helpful to clarify the concept of expenditure pertinent to the analysis and then embark upon an inquiry of the factors that in the past appear to have had an impact on them. The question of how the cost of a given bundle of educational services behaved over time is best answered by talking in terms of current expenditures, plus a reflection of the cost of providing physical facilities. The latter can be approximated by a debt service figure, which can reflect the long-run benefits of capital expenditures. Such data are compiled for education by the U.S. Department of Health, Education, and Welfare (see chart 1). PAGENO="0028" 16 ANALYSIS OF RISING COSTS OF PUBLIC EDUCATION While, no doubt, a large variety of factors can be expected to deter~ mine expenditure levels, the following are likely to be of particulai significance: 1. Number of pupils in average daily attendance (ADA). 2. Sociological characteristics of population: (a) Age structure, e.g., primary versus secondary school popu- lation. (b) Geographical distribution, e.g., urbanization of school population. 3. Economic characteristics: (a) Price level of goods and services bought by schools. (b) Income level, i.e., ability to afford services. FACTORS AFFECTING EXPENDITURE LEVEL PAGENO="0029" ANALYSIS OF RISING COSTS OF PUBLIC EDUCATION 17 4. Physical characteristics: (a) Productivity of school system. 5 Government characteristics. 6. Variety, scope, and quality of education services: (a) Addition or deletion of services-variety. (b) More or less of a given service-scope. (c) Superior or inferior services-quality. In the following pages each of these factors will be briefly discussed and an effort made to quantify each. TABLE 7.-Pupils in average daily attendance (ADA), selected years, 1900-1958 [In thousands] Pupils in average Pupils in average Pupils in average Year: daily attendance Year-Con. daily attendance Year-Con. daily attendance 1900 10, 633 1930 21, 265 1950 22, 284 1902 11, 064 1932 22, 245 1952 23, 257 1910 12, 827 1940 22, 042 1954 25, 644 1913 13, 614 1942 21, 031 1956 27, 740 1920 16, 150 1946 19, 849 1958 29, 859 1922 18, 432 1948 20, 910 Sources: U.S. Department of Health, Education, and Welfare, "Biennial Survey of Education in the United States 1954-56," cli. 2, table 20, p. 66; select years not appearing in table 20 of the 1954-56 Biennial Survey were taken from earlier issues of the survey beginning in 1920, on file at the U.S. Department of I-lealth, Education, and Welfare. All figures for years 1900-1956 are actual attendance data and are not estimates. The 1958 figure is an estimate made by the U.S. Department of I-Iealth, Education, and Welfare and is based on their annual fall survey. 1. Number of pupils in average daily attendance Since most public programs are related to the number of pupils in average daily attendance and not to enrollment, the first will be used, whenever possible, to reflect the public school population in this study. In table 7 the number of pupils in average daily attendance for selected years, 1900-1958, are given. During this period pupils in average daily attendance increased from slightly more than 10 to almost 30 million-that is, about 200 percent. The nature of the functional relationship between population size and expenditures deserves some consideration. It is possible that population increases lead to economies (or diseconomies) of scale. However, in a scale analysis, both school size and the size of the school district need to be considered. During 1900-1958, the U.S. population grew rapidly; it became in-S creasingly urban in character. Many school districts consolidated. Thus, during the 15 years from 1942 to 1957, more than half of the local districts disappeared.7 Likewise, the number of one-teacher schools rapidly declined from 190,655 in 1920 to about 26,000 in 1959-that is, by about 86 percent.8 Still, it must be remembered that during this period, one-teacher schools educated but a very small percentage of all pupils in average daily attendance in the United States. There exist no data about the size of the average school district. But there is reason to expect that once the effect of changes in urbani- zation and the accompanying tendency toward somewhat larger school districts is partialed out-for example, by means of multiple regression techniques-a by and large simple linear relation between per pupil expenditures and number of pupils will result. This contention is considered in some detail in the appendix, which examines the general U.S. Bureau of the Census, "Governments in the United States," No. 1, vol. 1, 1957, p. 1. National Education Association, "Status and Trends: Vital Statistics, Education, and Public Finance," Rept. 13, August 1959, p. 14. PAGENO="0030" 18 ANALYSIS OF RISING COSTS OF PUBLIC EDUCATION problem of economies of scale in public education in a deductive manner and on the basis of a case study, and finds that growth or consolidation of school districts are unlikely to produce significant economies of scale.9 2. Sociological characteristics of population (a) Age strn~ture.-Not only does the age structure of the population affect the size of the school-age population, it also affects the percent of pupils in high school as against that in primary schools. Smce on the average there are fewer high school pupils per teacher, more expensive equipment is needed, and salaries tend to be higher in sec- ondary education; changes in this ratio can affect overall education costs. During 1900-1958, the percent of the high school age population has changed and with it also high school enrollment as a percent of total enrollment in public schools)° By how much the latter ha~ increased can be seen by referring to table 8, which presents for 1900-1958 high school enrollment relative to total enrollment. The 1900-1958 increase was about 580 percent. TABLE 8.-Public high school relative to total public enrollment, selected years, 1900-19.58 Public high school Public high school Public high school as a percent of as a percent of as a percent of total public total public total public Year: enrollment Year-Con. enrollment Year-Con. enrollment 1900 3. 3 1930 17. 1 1950 22. 8 1902 3. 5 1932 19. 6 1952 22. 1 1910 5. 1 1940 26. 0 1954 21. 8 1913 6. 1 1942 26. 0 1956 22. 1 1920 10. 2 1946 24. 1 1958 22. 4 1922 12. 4 1948 23. 6 Sources: Worksheets of historical education series collected by Vance Grant, educational statistician, Division of Statistics and Research Services. Reference, Estimates, and Projections Section, U.S. Depart- ment of Health, Education, and Welfare, Office of Education. (b) Geographical distribution.-Rapid urbanization has affected our public schools and possibly their expenditures. For example, in urban communities children live shorter distances from school than they do in rural communities; many urban communities need no school buses, while most rural communities do. Furthermore, in many rural communities school enrollment is alarmingly small; some of them still have one-teacher schools. As a result, per pupil expendi- tures tend to be high, considering the low quality of education that is offered. Thus, it stands to reason that increasing urbanization has tended to lead to lower per pupil expenditures, particularly if the scope and quality of education is held constant. `Werner Z. Hirsch, "Expenditure Implications of Metropolitan Growth and Consolidation," Review of Economics and Statistics (41), August 1959. `C This information could only be obtained on the basis of enrollment and not pupils in average daily attendance. However, both series are likely to be highly correlated. PAGENO="0031" ANALYSIS OF RISING COSTS OF PUBLIC EDUCATION 19 TABLE 9.-Urbanization of school population, selected years, 1900-1958 Percent of pupils Percent of pupils Percent of pupils in average daily in average daily in average daily attendance (age attendance (age attendance (age 5-19) from 5-19) from 5-19) from Year: urban areas Year-Con. urban areas Year-Con. urban areas 1900 39. 6 1930 53. 3 1950 52. 1 1902 39.8 1932 53.6 1952 51.5 1910 40.4 1940 54.6 1954 51.8 1913 42. 0 1942 55. 1 1956 51. 7 1920 45.9 1946 55.5 1958 51.5 1922 47. 4 1948 54. 3 NOTE.-The 1940 definition of an urban area was utilized so as to make the series comparable. Since an urban-rural breakdown by age groups was available only for census years 1910-50, the oIl-census years were estimated by assuming a constant rate of growth between consecutive censuses. Compound interest formulas were utilized in making these estimates. In extending the 1940 definition of "urban area" forward from 1950-58, the rate of increase of the 1950 definition of "urban area" was applied to the 1940 definition. Since much of the area in metropolitan areas as defined by the census in 1950 as "rural" will be redefined as "urban" in 1960, the method used to extend the 1940 definition probably understates the real urban popula- tion rate of increase. This fact was pointed out by Comad Taeuber, Assistant Director, U.S. Bureau of the Census. Sources: U.S. Department of Commerce, "Statistical Abstract of the United States, 1958," ch. 1: Area and Population. U.S. Department of Commerce, "Population: Characteristics of the Population," U.S. Summary, 16th Census, 1940, tables 7 and 9. U.S. Department of Conimerce, "Population: U.S. Summary," 1950 Census, Rept. P-B1, vol. 11, pt. 1, cli. B, table 38. rfhe extent to which the percent of pupils (5 to 19 years old) in average daily attendance living in urban America has undergone changes since the turn of the century can be learned from table 9. The percentage increased steadily from about 40 percent in 1900 to about 56 percent in 1946. Since then it has slowly declined to about 52 percent in 1958.11 In addition to urbanization, other geo~raphic shifts have taken place. The extremely rapid growth of California and industrialization of parts of the Southern States are two of many possible examples. However, because of its complexity, this type of change in the geo- graphic distribution of the United States will not be included in the analysis. 3. Economic characteristics (a) Price level of goods and services bought by schools. -To the extent that the prices paid by schools for teachers, equipment, etc., increase, also the overall cost of operating schools will tend to rise, unless hese increases are compensated for by such factors as productivity increases or quality deterioration. In order to fully gage changes in the prices paid by State and local governments for public education inputs, it is best to separate the various inputs into major categories, establish their importance during the period under consideration, and develop cost data for each major category. Main categories that come to mind are salaries of instruc- tion staff, salaries of maintenance personnel, equipment, etc. Great difficulties are encountered in trying to estimate salaries of school maintenance personnel and equipment. Since by far the overwhelm- 11 Although in all other instances the 5:to 17 years age group was considered of school age, early census data were only available on a 5- to 19-year basis. For this reason, the latter group was used here. PAGENO="0032" 20 ANALYSIS OF RISING COSTS OF PUBLIC EDUCATION ing part of all current expenditures take the form of salaries to the instruction staff, and since reasonably good data on annual salaries of instruction personnel are available, these data are used as the main source of information on price level changes. It is possible to find data on average annual earnings of full-time equivalent State and local government employees in public education. However, this information goes back only to 1930 and these data not only include salaries to the instruction staff but they cover all salaries. At the same time, these data cover salaries paid not only to the personnel of public primary and secondary schools, but also to the personnel of State and city colleges and universities. In table 10, information on average annual salaries per member of total instruction staff in primary and secondary schools for selected years, 1900-1958, is given. These data are supplemented for selected years from 1930 on by data on the average annual earnings of full- time equivalent State and local employees in public education. To facilitate comparisons, for both series index numbers have been developed with 1954 = 100. Judging by the index numbers, the two series resemble one another rather closely, especially since 1950. Prior to that period, increases in the first series were more rapid than those in the second. TABLE 10.-Average annual salaries of public school teachers, selected years, 1900-1958 Average Average annual annual salary per earnings member of Index full-time Index instruction 1954=100 equivalent 1954=100 Year staff in pri- State and mary and local govern- secondary meut public shools education (1) (2) (3) (4) i~oo 5325 8.5 1902 9. 1 1910 485 12. 7 1913 512 13.4 1920 871 22.8 1922 1,166 30.5 1930 1,420 37.1 $1,455 41.5 1932 1,417 37.0 1,399 39.9 1, 441 37. 7 1, 435 40. 9 1942 1, 507 39. 4 1, 512 43. 1 1, 995 52.2 2,025 57. 7 1948 2, 639 69. 0 2, 538 72. 3 ioso 3, 010 78. 7 2, 794 79. 6 1952 3,450 90.2 3, 169 90. 3 1954 3,825 100. 0 3, 510 100. 0 1956 4, 139 108. 2 3, 799 108. 2 1958 ~ ~00 122. 8 4, 343 123. 7 Sources: Average annual salary data per member of total instruction staff were computed by Eugene P. McLoone, Division of State and Local Scbool Systems, School Finance Section, U.S. Office of Education. Average annual earning data of full-time equivalent State-local government public education was taken from: U.S. Department of Commerce, ~ational Income, 1954, a supplement to the Survey of Current Business, table 15. U.S. Department of Commerce, L.S. Income and Output, a supplement to the Survey of Current Busi- ness, 1958, table \1-15. In 1900, the average teachers' salary in public primary and second- ary schools was $325. By 1958, it was about 13 times larger, i.e., about $4,700. (b) Income level.-There can be no doubt that throughout the United States the richer communities and States spend on the aver- PAGENO="0033" ANALYSIS OF RISING COSTS OF PUBLIC EDUCATION 21 age more for education than do the poorer ones. It is true that this is not a perfect linear relationship. Sonic States and communities make greater tax efforts for education than do others. Nevertheless, by and large, this relationship holds. The relation between income and expenditures is not a simple one. Thus, for example, rich com- munities will tend to insist on good education, which in turn will add to the cost of education. In such a case, statistical methods are unlikely to determine whether changes in income level or quality level are originally responsible for expenditure level changes. A good understanding of the impact of income level changes on education is of paramount importance. In spending his income, a person faces a virtually unlimited number of different goods and services. As his income increases, his spending habits are likely to change, too. Under those circumstances, the ques- tion is whether the average person, whose annual income increased by a given percent, will increase his education budget by the same per. cent, by less, or by more. If for simplicity's sake education expendi- tures are equated, in a rough way, with the amount of education pur- chased, the concept of income elasticity of public education becomes applicable. Education demand can be unit elastic, inelastic, or elastic. The question, therefore, is, "How elastic or inelastic is the demand for education?" TABLE 11.-Per capita personal income, selected years, 1900-1958 Per capita Per capita Per capita personal personal personal Year: income Year-Con. income Year-Con. income 1900 $199. 2 1930 $624. 1 1950 $1, 506. 2 1902 218. 5 1932 401. 4 1952 1, 739. 0 1910 300. 6 1940 595. 5 1954 1, 784. 5 1913 325. 1 1942 915. 7 1956 1, 965. 3 1920 712. 3 1946 1, 268. 1 1958 2, 030. 3 1922 552. 4 1948 1, 435. 1 Sources: Years 1900-1928. National Industrial Conference Board. Economic Almanac 1958. New York: Thomas Y. Crowell Co., 1958, p. 401. 1929-57. U.S. Department of Commerce, Office of Business Economics. Survey of Current Business 38: pp. 3-17; July 1958. Tables 3 and 4, pp. 6 and 7. Year 1958. From Economic Indicators, May 1959, pp. 2, 3, and 5. Personal income data and their index numbers are presented in table 11 for a select number of years from 1900 to 1958. Except for the depression years of the early thirties, per capita personal income steadily increased from about $199 in 1900 to about $2,030 in 1958, i.e., about nine times. 4. Physical characteristics (a) Productivity of school system.-Productivity in general relates output to any or all of the inputs employed in producing goods or service. Since the aim is to learn about the efficiency with which re- sources are utilized, input as well as output must be measured in physical terms. Perhaps the most common productivity measure is the partial productivity "output per man-hour." Output is related to but one input, i.e., labor, which in the case of education over- shadows by far all other inputs. To estimate the net saving of all cost elements, or inputs, and thus the change in overall productive efficiency, output should be related to all inputs, i.e., labor, equipment, physical plant, etc. 47761-59-3 PAGENO="0034" 22 ANALYSIS OF RISING COSTS OF PUBLIC E1)UCATION Productivity of schools is most difficult to measure. The main measurement problems center around output. An automobile assembly plant turns out on a given day a given number of cars of virtually uniform characteristics. They are the plant's output. The output of education is much less tangible and much less stand- ardized. Everyone will agree that education has changed since the turn of the century, and yet few will be able to say, in quantitative terms, how great the change has been. Whether it is because or in spite of the great difficulties of measuring educational outpuL, rela- tively little work has been done in this area. One of the most serious studies of productivity in government was undertaken by Solomon Fa.bricant in 1952. He concluded that a- Review of some factors affecting the trend of government productivity-the use of improved technology and equipment, the spread of the merit system, the introduction of centralized purchasing, and various other advances in public administration-leaves the impression that labor savings have been made. Indeed, it is hard to think of any factor tending in the opposite direction except possibly the very increase in the scale of governsnent operations. Nor does it appear that these savings of labor reflect merely increase in the volume of other resources-capital goods and other purchases-used per worker. Total produc- tivity, output per combined unit of all resources, appears to leave risen in government.12 An empirical effort to measure the productivity trend of Federal agencies was undertaken by Henry D. Lytton.'3 Lytton estimated changes of the last 10 years for the Commodity Stabilization Service, Social Security Administration, Post Office Department, Veterans' Administration, and Internal Revenue Service. Most of the outputs of these agencies are readily quantified. Thus, for instance, a post office handles a given number of pieces of mail having a given volume and given other characteristics. The productivity advances of these agencies during the last 12 years averaged about l~ percent a year. The range extended from 1~ to 8 percent, compounded annually. The Commodity Stabifization Service showed the greatest productivity increase; i.e., of about 8 percent a year. Much of this change appar- ently was due to the introduction of electronic processing equipment. The same general view was expressed by Richard Ruggles and Nancy D. Ruggles, in their paper to the Joint Economic Committee when they maintained that- There is good reason to believe, however, that the productivity of Government workers has increased substantially in this period. For one thing, the introduc- tion of data handling machines and computers speeds up the operation of many stages of Government work. Statistics in the Government now, in large part, are handled mechanically rather than by clerks. The mechanization which is so characteristic of current development in business is also occurring in Govern- ment.14 In 1955 Fortune magazine carried an article, "Productivity: The Great Age of 3 Percent." In this article it was stated that- Government efficiency, because electronic tabulating and computing can take over the bulk of its clerical work, can and should rise enormously. * * * The productivity of schools and teachers, by contrast, cannot and properly should not increase very much.15 12 Solomon Fabricant, "The Trend of Government Activity iritheUnited States Since 1900" (New York: National Bureau of Economic Research, Inc., 1952), pp. 98-99. 13 Henry D. Lytton, "Estimating Recent Federal Agency Productivity Trends" (Washington, D.C., 1959), 46 pages. 14 Richard Ruggles and Nancy D. Ruggles, op. cit., p. 299. "Gilbert Burck and Sam Ford Parker, "Productivity: The Great Age of 3 Percent," Fortune, November 1955, pp. 249-250. PAGENO="0035" ANALYSIS OF RISING COSTS OF PUBLIC EDUCATION 23 Instead of measuring the productivity of our public primary and secondary schools in any direct manner, an indirect method might be tried. It would move along the following lines: time series data of current education expenditures plus debt service would be so adjusted and standardized that price level and variety, scope, and quality of education-to the extent that they affect expenditures- are held constant and are expressed in per pupil in average daily attendance terms. This series could be called education costs in constant terms. It could trace over time cost changes of a given bundle of public primary and secondary education. But it can also testify to productivity changes. For example, should education costs in constant terms have consistently fallen, it stands to .reason that productivity increases would have occurred and vice versa. While it is not too clear whether and, if so, how much productivity in Government in general increased, it is even more difficult to assess the productivity increases in public education. Many of the forces that apparently work in general State and local governments do not work in education. Thus, for instance, electronic computers have not been introduced nor do they warrant introduction; likewise, ad- ministrative improvements appear to have been minor. While we cannot identify many steps taken in the past with a view to improving productivity, a few new methods have recently been discussed and tried. Perhaps the most important one is television. The Ford Foundation and the Fund for the Advancement of Education have been instrumental in experimenting with what has been described as the most important new educational tool since the invention of movable type. In their May 1959 report, "Teaching by Television," Ford Foundation and the Fund conclude that- Today the question is no longer whether television can play an important role in education. That question has been answered in the affirmative not only by the experiments supported by the Fund and the Foundation but also by the many other programs in which the medium is being used successfully for direct classroom instruction. The question that now needs fuller exploration is what kind of a role television can play most effectively. * * * Television can be used-and in some places is being used-to do the traditional job of education, and to do it well. However, those who have had experience with the medium know that, if wisely and imaginatively used, it also can bring to students educational experiences far beyond what is possible in the conventional classroom. * * * Also, television makes possible exciting new developments in the team approach to teaching, in which the particular skills and competencies of many teachers are used cooperatively in planning and presenting courses. The status and rewards of teaching can be vastly enhanced by this new medium." Another method that can possibly increase productivity in schools is the use of teachers' aids, who are employed to relieve teachers of such burdensome and nonteaching chores as compiling attendance records, collecting various kinds of contributions and fees, cleaning blackboards, watering plants, etc. To make fuller use of the physical plant, the length of the school term has been expanded. Some schools are beginning to have summer sessions. Some are considering a four-quarter system. The entire school year would be divided into four school quarters; schools would be open all year and students would attend any three quarters they choose. If this is successful, the present physical plant could thus acconnnodate as much as one-third more students. 16 Ford Foundation and Fund for the Advancement of Education, "Teaching by Television" (New York 1959), pp. 60-61. PAGENO="0036" 24 ANALYSIS OF RISING COSTS OF PUBLIC EDUCATION 5. Government characteristics Governmental structure and relationships can greatly affect the amount of money available for public education. At the moment not too much is known about the specific governmental characteristics which affect services and expenditures. But there ~an be no doubt that State-local fiscal interrelationships can play an important role. Clearly, the more the expenditures are financed by the State, and the more the State subsidies are based on the equalization principle, the more uniform the availability of funds throughout the State will be. Julius Margolis advanced the thesis in a paper before the 1959 National Bureau of Economic Research's "Conference on Public Finances: Needs, Sources, and LTtiization," 16a that governments with many functions may find it easier to raise funds b taxation, whereas such single-purpose special districts as school districts may not have sufficient fiscal strength to fulfil their goals. As is well known, public education is offered by 1)0th independent governments, i. e., school dis- tricts, and as special functions of other governments, i.e., counties, municipalities, and townships. Only about 4 percent of the public school systems in the United States are nonindependent school de- partments. But they are significant in size-they have 22 percent of the school enrollment.. According to Margolis, the comparison between fiscally independent school districts and fiscally dependent school departments is of inter- est since it indicates whether the public will spend more for a~ public service when it is presented as a single unit or as part of a package for which the specific benefit of the marginal tax dollar is uncertain. Agreeing that the evidence is far from conclusive, he reproduced some data that were used first by Henry B. Woodward.'7 They indicate that fiscally dependent departments spend more per pupil than fiscally independent school districts. Woodward had made an analysis of expendittues in 85 cities between 100,000 and 1 million population over the period 1929-30 to 1943-44. During this period the mean per pupil expenditure was highest in the fiscally dependent cities. They spent about 4 percent more than the independent districts in 1929-30 and nearly 12 percent more in 1943-44. In addition, Margolis refers to the following census data. The April 1957 payroll of fiscally dependent school departments was $22.90 per pupil while for fiscally independent school districts it was $20.90. Ten of the forty-one largest cities have municipal school departments, but they spent in 1952-53 $346.11 per pupil in average daily attend- ance against $293.40 spent in the other 31 largest cities with independ- ent school districts.18 Another example given by Margolis relates to New Jersey where there are dependent and independent systems of all size classes. The cities of over 100,000 had municipal school departments. They spent $383 per pupil in average daily attendance. In the remaining cities, the municipal school departments spent $343 per pupil in average daily attendance, while the independent school districts spent $293 per pupil in average daily attendance. 16a Public Finances: Needs. Sources, and Utilization-A Conference of the Universities-National Bureau Committee for Economic Research; A Report of the National Bureau of Economic Research to be published by Princeton University Press (1960). 17 Henry B. Woodward, "The Efiects of Fiscal Control on Current School Expenditures," Ph. D. thesis. New York Teachers College, 194S. Cited in Powell B. Mort and Walter C. Reusser, "Public School Finance," second edition (New York: McGraw-Hffi, 1941), p. 60. 15 u.s. Bureau of the Census, 1957 Census of Government, vol. 2. No. 1, table 4, and vol. 1. No.1, tabl&10. And U.S. Office of Education, Annual Survey of Education, 1953-54. "Statistics of City School Systems," pp. 36-45. PAGENO="0037" ANALYSIS OF RISING COSTS OF PUBLIC EDUCATION 25 The great shortcoming of these data is that they are not subjected to a multiple correlation and regression analysis. Much more work remains to be done on this score before Margolis' thesis can be ac- cepted or rejected. 6. Variety, scope, and quality of education services (a) Addition or deletion of services.-There are a number of different ways in which an attempt can be made to measure the scope and qual- ity of public education. But before they are explored, it appears desirable to distinguish between two types of services offered by schools, i.e., services directly designed to improve education versus those that do not directly affect educational quality. Among the hatter are the~ transportation of children in schoolbuses, the feeding of children in school cafeterias, health seivices, etc. No one will claim that these services are not important. However, it is helpful to recognize that they do not directly contribute to the quality of educa- tion. From now on, they will be called auxiliary school services. At the turn of the century they were hardly known. Today they amount to almost $1.5 billion, the most rapid advance having taken place during the postwar period. Expenditures for auxiliary school services are given in table 12. In many respects it is desirable to deduct expenditures for auxiliary school services from total current expenditures (plus debt services). If this is done, current expenditures (plus debt service) data are re- duced to a more common denominator. TA B LII 12.-Auxiliary school service expenditures, selected years, 1900-1958 [In thousands] Auxiliary school Auxiliary school Auxiliary school service service service Year: expenditures Year-Con. expenditures Year-Con. expenditures 0 1930...~_ $152, 263 1950____ $713, 132 1902..~ 1 0 1932....... 144, 273 1952_... 917, 440 1910... 1 $5, 000 1940~.~ 179, 257 1954_~. 1, 020, 037 1913...... 110, 000 1942....~ 219, 670 1956..._ 1, 304, 244 1920....... 45, 673 1946...~ 349, 097 1958_... 1 1, 490, 000 1922___ 69, 266 1948__... 526, 999 1 Estimate. Source: U.S. Department of Health, Education, and Welfare, "Biennial Survey of Education in the U.S.-1954-56," chapter 2, pp. 19-21, 80-82, and worksheets made available by the U.S. Department of Health, Education, and Welfare. TABLE 13.-Length of school term, selected years, 1900-1958 Year Average length of school term (days) Average number of days attended by pupils enrolled Year Average length of school term (days) Average number of days attended by pupils enrolled 1900 1902 1910 1913 1920 1922 1930 1932 1940 144.3 144.7 157.5 158.1 161.9 164.0 172.7 171.2 175.0 99.0 100.6 113.0 115.6 121.2 130.6 143.0 144.9 151.7 1942 1946 1948 1950 1952 1954 1956 1958 174.7 176.8 177.6 177.9 178.2 178.6 178.0 178.7 149.6 110.6 155.1 157.9 156.0 158.9 158.5 Source: U.S. Department of Health, Education, and Welfare, "Biennial Survey of Education in the United States, 1954-SI," ch. 2, table 1; for select prior years; these data were taken from earlier biennial survey reports. PAGENO="0038" 26 ~i~sis OF RISING COSTS OF PUBLIC EDUCATION (b) Scope of seruices.-As a first approximation it might be claimed that the longer the school term, the more can be learned. Table 13 indicates that, since the turn of the century, major changes have taken place in the length of the school term. In 1900, the school term lasted an average of 144 days; by 1958, it had been lengthened to 179 days, i.e., by about 24 percent. At the moment further lengthening is con- templated by many school boards. The change in the average number of days attended by pupils en- rolled in public schools during the school term has been even more striking during this period. The average increased from 99 days in 1900 to 158 days in 1956, i.e., about 60 percent. (c) Quality of ser~ices.-Perhaps the most feasible approach would involve identifying and then measuring ingredients that make for good education. On that basis an index of quality of education could then be developed. Here are some of the factors that might he included in such an index of the quality of public education: (1) Class size: While the average pupil-teacher ratio in average daily attendance is a reasonably good indication of class size, like any other average it conceals much detail. For instance, since high school classes are smaller than primary school classes, the age distribution is important. Another telling item would be whether the school offers a course in advanced mathematics or a foreign language if only 10 or 15 students are enrolled. Also the number of teachers' aids is to be considered. (2) Grouping: Many educators maintain tha.t good education re- quires that, within limits, students of common ability and interest are grouped together.'9 Even after this philosophy is generally adopted, schools will differ as to whether they are able to deal with very small groups of very able students. (3) Quality of the teaching staff: A good teaching staff has a number of important characteristics, some more tangible than others. The percent of experienced teachers is revealing. Do they tend to stay on for a career in the school system or do they move often? Also, the background of the teaching staff plays a dominating role. What is the percent of teachers who are graduates of strong liberal arts colleges with majors in the field or fields in which they are teaching? The methods used for selecting new teachers and of appraising the quality of the existing staff are also of interest. A good school inter- views many persons for a position; it looks outside the borders of its district and even outside its State for personnel. Thus, an important indicator would be the average number of outside-the-area candidates interviewed for each teacher hired, or the percent of outsiders on the staff. In addition, the better school systems have an organized pro- gram of teacher appraisal on which merit increases are based. Teaching load is another indicator. In many schools 20 hours of teaching a week and dealing with about 175 students is considered normal. It might be helpful to relate the average teaching load of a given school to the national average. Finally, the quality of the teaching staff depends to no small extent on the number and the variety of specialists among it. Is there a full-time librarian, or more than one, and what is the number of "See James B. Conant, "The American High School Today" (New York: McGraw.Hill, 1959), 140 pages. PAGENO="0039" ANALYSIS OF RISING COSTS OF PUBLIC EDUCATION 27 students per librarian? Likewise, how many counselors are there, and how many students per counselor? Similar figures should be collected for other specialists. (4) Quality of school administration: The leadership offered and ability of the school superintendent and his principals cannot be neglected; yet it is most difficult to appraise their contributions. Usually, principals who are not relieved from all teaching cannot do a superior job. (5) Teaching program: Schools with a good college preparatory program insist that college-bound students carry four courses a year in English, mathematics, science, history, or foreign languages. What percent of the students have such a program? How far does the mathematics program go? What percent of the students take mathe- matics courses beyond a second year of algebra and 1 year of plane geometry? Are they taught calculus? In the foreign language field, what percent of the students take 3 or 4 years of foreign languages? How much writing is done in the English classes? While it is possible to develop an ideal hypotheis about the quality of education, the task of implementation is enormous. It was at~ tempted in a study of the public schools in the St. Louis city-county area.2° This quality index was composed of six subindexes: A==number of teachers per 100 pupils in average daily attend- ance.21 B=number of college hours of average teacher. 0= average teacher salary. D=percent of teachers with more than 10 years of experience. E=number of high school credit units. F=percent of high school seniors entering college. In this index, A is supposed to reflect class size, B, 0, and B quality of teaching staff, and E quality of teaching program. It is true that Fis not an ingredient of good education. Instead it is a way of crudely approximating an achievement-native ability comparison. It may, however, be a better measure than appears at first glance. It is widely assumed that whether a high school senior enters college de- pends mainly upon the financial situation of his parents. Yet the rank correlation coefficient measuring the relation between percent of high school seniors entering college in 1955 and average assessed valu- ation of the 27 school districts, which were considered in this case study, is +0.34, and is statistically barely significant at an a of 0.05. This low correlation may have a number of causes. College education has become more and more universal and is no longer available pri- marily to the sons of the well to do. Also, school districts are not entirely homogeneous in regard to property values. And finally, assessed valuation is not only the product of residential property values, it is also affected by the relative importance of commercial and indus- trial property located within the district. Atiother subindex tried was one reflecting average capital expendi- ture per pupil in average daily attendance for the past 4 years. As 25 For detailed discussion, see Werner Z. Hirsch "Measuring Factors Affecting Expenditisre Levels of Local Government Services" (St. Louis: Metropolitan St. Louis survey, 1957), 94 pages. 21 Teachers per 100 pupils in average daily attendance and average teacher salary are usually expected to be highly correlated with the general level of expenditures. An analysis revealed that the correlation, while high,~isUar~from perfect. It is;+0.71 in the first~case~and +0.68 in the second ease. Both are not very high. Thus, the inclusion of factors A and C in the scopeand quality index appears warranted. PAGENO="0040" 28 ANALYSIS OF RISING COSTS OF PUBLIC EDUCATION was to be expected, a simple scatter diagram revealed a very high cor- relation between this factor and school population increases. This factor appears to measure primarily growth and not quality; and it was therefore discarded. The weighting problem is truly difficult. Who can confidently assess the relative importance of the number of teachers per 100 pupils, as compared to the number of high school credits offered by the school? Since no useful basis could be found to give each of the six factors specific differential weights, equal weights were used. Fortunately, with six components to the index, the weighting system is no longer so very crucial. Doubling the weights of any one or two of the components will not greatly affect the magnitude of the index numbers. This simplified index was subjected to a test which sheds some light on its reliability. A number of educators who are well acquainted with public schools of the St. Louis city-county area were interviewed separately and asked to rate the various school distficts in terms of excellent, good, medium, poor, and very poor. These ratings, when compared with the scope and quality index data, showed very close consistency. In no case was the school district with an index number above 1, i.e., average, given a subjective rating of poor or very poor, and vice versa. In the few cases with difference in the rating, the index number was equivalent to excellent and the subjective rating was good. Although this is a very simple validation procedure, it might well be that if there are inaccuracies in the index numbers, they are not in excess of from 5 to 10 percent either way. To the extent that discrepancies e~st, they are likely to be randomly distributed and no bias appears to have crept into the index. On the national level, the paucity of appropriate and consistent data, useful for such an analysis, is very great. Virtually all the data that are available, and that at first glance appear pertinent, are ex- pressed in dollars. They pose serious problems of circularity. It is a simple tautology to state that high salaries, designed to reflect variety, scope, and quality of education, affect expenditures. However, one measure, expressed in physical units and most likely an ingredient of good education, is the number of principals, super- intendents, and consultants per 1,000 pupils in average daily attend- ance. These are the educational experts who evaluate the curriculum and introduce changes. provide guidance, supervise teachers, etc., and their number is considered to l)e as reliable an indication of the quality of education as can be obtained at present; superior to any other simple measure. as, for iiistance, teacher-pupil ratio. About the latter, Ralph Cordiner, of the General Electric Co., is reported to have made the controversial statement- There are some educators who are actually proud of the declining ratio of students to faculty. * * * To my knowledge, there is no other field in human endeavor which actually prides itself on declining productivity.2' Returning to the number of principals, superintendents, and consultants per 1,000 pupils in average daily attendance as a sole measure of the quality of education, in the narrow sense, the following assumptions should be kept in mind. Quality of output is highly 22 Quoted by J. A. Livingston in "Get With It, Professor, Your Shows Slipping," the Washington Post and Times Herald. Nov. 27, 1957. PAGENO="0041" ANALYSIS OF RISING COSTS OF PUBLIC EDUCATION 29 correlated with the quality of certain crucial inputs, one of which is the number of principals, superintendents, and consultants per 1,000 pupils in average daily attendance. Also, superior school districts not only hire more specialists than do inferior ones, but they also hire better ones and, in general, they are rational and approximate a marginal calculus by using high-quality ingredients all around, i.e., the quality of none of the ingredients making for good education is far out of hue.23 As can be seen from table 14, the number of principals, superin- tendents, and consultants per 1,000 pupils in average daily attendance has about doubled since the turn of the century. TABLE 14.-Total number of principals, superintendents, and consultants per 1,000 pupils in average daily attendance in public primary and secondary schools, selected years, 1900-1958 Number of principals, Number of principals, Number of principals, superintendents, and superintendents, and superintendents, and consultants per 1,000 consultants per 1,000 consultants per 1,000 pupils in average pupils in average pupils in average daily attendance daily attendance daily attendance Year: Year-Con. Year-Con. 1900 1 1. 31 1930 1. 78 1950 2. 18 1902 1 1. 31 1932 1 1. 33 1952 ~2 13 1910 1 1. 33 1940 1. 65 1954 2. 18 1913 1 1. 32 1942 11 86 1956 2. 32 1920 1.25 1946 1 1. 82 1958 2. 70 1922 1 1.44 1948 12.22 1 Estimate. No~~.-The estimates were made by extending the biennial survey data backward based on Stigler's study. Estimates for certain years was based on simple linear extrapolation. The number of consultants, superintendents, and principals was derived as a "residual" where data on total instruction staff and the number of teachers, librarians, and other nonsupervisory instruction staff were available. Sources: U.S. Department of Health, Education, and Welfare, "Biennial Survey of Education in the U.S., 1954-56," ch. 2, table 1; George 1. Stigler, "Employment and Compensation in Education," Occasional Paper 33, NBER, Inc. (1950). At least in theory, there is a much superior approach to measuring scope and quality changes of public education. It involves comparing the students' achievement with their native ability. Specifically, a pupil's achievement test results obtained in high school could be compared with results of his IQ test taken early in his primary grade school days. One complicating element in this approach is that school districts do not have identical educational objectives. Further- more, some schopls consider it their task to prepare youngsters for college, while other schools, particularly those in rural America and to some extent in the core cities, have as their main objective the provision of vocational training. There can be no doubt that students will score differently on an achievement test depending upon the school's objective. Implementation of this approach would require that all schools in the United States use the very same achievement and IQ tests and that they have done so for a long period of time. Since these require- ments are not met, this method does not lend itself to practical application at the moment. 23 The author is obliged to Prof. Wffliam Fellner, who pointed out these explicit assumptions. PAGENO="0042" 30 ANALYSIS OF RISING COSTS OP PUBLIC EDUCATION TEST OF HYPOTHESIS The working hypothesis which was enunciated in the preceding pages will now be tested in the light of available national data. In many respects, availability more than appropriateness of data deter- mines their application. Time series data for selected years during 1900-1958 will be used and subjected to multiple correlation and regression analyses. By insisting on intervals of at least 2, and in many instances 7 or 8 years, it is hoped that much otherwise existing serial intercorrelation is eliminated.24 But with only data for 17 years at hand, one is somewhat handicapped in including all the independent variables that appear appropriate. Care must be exercised that no unduly large number of degrees of freedom is lost. Because of this con- sideration, it was decided to use as dependent variable, daily total current expenditures plus debt service per pupil in average daily attendance. In a sense, the dependent variable is already adjusted for possibly two independent variables-number of pupils in average daily attendance and length of school year. While there are some risks in taking such a step, they were considered outweighed by the advantages of retaining more degrees of freedom. In order to shed light on such questions as, "Why has the public education sector had such large expenditure increases?" and "What are the underlying relationships for the near term?" two working hypotheses have been enunciated and empirically tested. The first hypothesis is as follows: Xia-daily total current expenditure plus debt service for public primary and secondary education per pupil in average daily attendance is a function of- X2-percent of public high school enrollment relative to total public school enrollment; X3-percent of pupils (5 to 19 years old) in average daily at- tendance living in urban areas; X4-average annual salary for member of instruction staff; and X6-number of principals, superintendents, and consultants per 1,000 pupils in average daily attendance. In brief, Xia=f (X2, X3, X4, X6). The following multiple regression equation was obtained for 17 selected years during 1900-1958: 7~ 0.002067 X2 0.005288 X3 A la_O.l64999+-~j~0257) (0.1069) +0.000441 X4+0.022391 X825 1 1 (0.9895) (0.0271) The coefficient of multiple determination, adjusted for degrees of freedom lost_R*22.2346_is 0.998 and is statistically highly significant. Thus, about 99.8 percent of the variation of daily total current ex- 24 Although no detailed checks were made, it appears that also the other two important assumptions underlying the method are met. Thus, the values of the dependent variable of the population appear to be normally distributed around the least square line, and also the standard deviations of these normal popula- tion distributions appear about equal. 25 The figures in brackets are partial correlation coefficients. Since there are 17-5 or 12 degrees of freedom, coefficients are statistically significant at an u of 0.05, when they arelarger than 0.532. PAGENO="0043" ANALYSIS OF RISING COSTS OF PUBLIC EDTJCATION 31 penditure plus debt service per pupil in average daily attendance is explainable in terms of these four variables of which salary level, X4, is statistically significant. All relationships but one are positive The significant net regression coefficient 0.000441 can be interpreted in the following manner-during 1900-1958, a $1,000 increase in average annual teacher salaries was on the average associated with a 44.1-cent increase in daily total current expenditures per pupil in average daily attendance, holding the effect of changes in the other three factors constant. Also, about 98 percent of the variation of the cost of the standardized education unit can be explained in terms of teacher salary level increases, holding the effect of changes in the other three factors constant. It is somewhat surprising that neither the percentage of high school enrollment, nor urbanization, nor the number of specialists per 1,000 pupils in average daily attendance, which was designed to measure quality of education, had a statistically significant effect on expendi- tures. The beta coefficients, i.e., net regression coefficients divided by their respective standard deviations, are, in order of magnitude: 1314.236=0.9944, /313.246- -0.0504, 1312.346-0.0273, and /316.234-0.0161. The simple correlation coefficient matrix, which follows, is of some interest: x~ x2 x~ x4 0.6892 x.. 0.5707 0.9659 x4 0.9992 0.7027 0.5881 xo 0.9484 0.7307 0.6150 0.9487 This matrix reveals exceptionally high intercorrrelations; for ex- ample, r23=0.9659 and r46=0.9487. All simple correlation coefficients are high and statistically significant at an cc of 0.05. The second hypothesis, a slight modification of the first one, takes explicit recognition of the fact that expenditures for auxiliary services and the number of principals, superintendents, and consultants per 1,000 pupils in average daily attendance are unlikely to be related. Thus, instead of relating X2, X3, X4, and X6 to daily total current ex- penditures, it relies upon a new dependent variable, i.e., X35-daily total current expenditures plus debt service minus expenditures for auxiliary services for public primary and secondary education per pupil in average daily attendance. On this basis, the following new multiple regression equation was obtained for 1900-1958: X' -0 006929 0.000603X2 0.000336X3 0.000372X4 1b . - (0.0030) (0.00058) + (0.9889) (1.2) ~ 0.0l3847X6 m (0.01110) PAGENO="0044" 32 ANALYSIS OF RISING COSTS OF PUBLIC EDUCATION The statistically significant net regression coefficient is b14.236. The beta coefficients are, in order of magnitude: 1314.236= 0.9970, 1316.234=0.0118, 1312.346=0.0094, and 1313.246=0.0036. The coefficient of multiple determination adjusted for degrees of freedom lost is highly significant at an a of 0.05. It is 0.998. The following is the simple correlation coefficient matrix: Xlb ~ 0.6963 0.5808 0.9659 0.9994 0.7027 0.5881 0. 9484 0. 7307 0. 6150 0. 9487 The foregoing multiple correlation and regression analysis clearly points to the overwhelming importance of the salary level variable and perhaps the prices of all goods and services bought by schools. However, they, in turn, are determined by the general demand and supply situation and price level of the rest of the economy, as well as the particular demand for and suppiy of teachers.26 EDUCATION COST IN CONSTANT TERMS A truly serious shortcoming of the deflating efforts of the U.S. Department of Commerce is its admitted inability to adjust for anything but price level changes. It expresses State and local gov- ernment purchases in constant dollar terms, but neglects changes in the variety, scope, and quality of services. Thus, an effort will be made to develop first a. series which in concept resembles that. of the U.S. Department of Commerce, i.e., cost of public primary and secondary education in constant dollars. Thereafter, more terms than merely price level will be held constant, leading to an education cost index in real terms. By deflating the actual expenditure series by an average annual salary index-with, for instance, 1954 = 100-a new series results which testifies to the expenditures that would have been incurred if 1954 salary conditions had prevailed throughout the period. ~ According to some calculations made by Roger A. Freeman; for instance, during 1929-57 teachers' salaries advanced more rapidly than the earnings of all workers-94 versus 82 percent. This indicates that the demand for and supply of teachers cannot be neglected. See Roger A. Freeman, "School Needs in the Decade Ahead" (Washington, D.C.: Institute for Social Science Research, 1958), p. 131. PAGENO="0045" ANALYSIS OF RISING COSTS OF PUBLIC EDUCATION 33 TABLE 15--Total current expenditures plus debt service for public, primary, and secondary education, selected years, 1900-1958 [In thousands of dollars] Year Unadjusted (1) Adjusted for pupils in ADA (1954= 100) (2) Adjusted for salary ]evel (1954=100) (3) Year Unadjusted (1) Adjusted for pupils in ADA (1954=100) (2) Adjusted for salary level (1954=100) (3) 1900 1902 1910 1913 1920 1922 1930 1932 1940 193. 9 214. 2 384.8 473. 0 938. 5 1,332.0 2,095. 0 2, 147. 9 2, 192.0 467. 2 497. 0 769.6 890. 8 1, 489. 7 1,852.6 2, 527. 1 2,477. 4 2,548.8 2, 281. 2 2, 353. 8 3,029.9 3, 529. 9 4, 116. 2 4,367.2 5, 646. 9 5, 805. 1 5,814. 3 1942 1946 1948 1950 1952 1954 1956 1958 2, 378. 6 2,995. 1 4,092.4 5, 090. 7 6, 219. 8 7,458.3 9,054. 6 1 11, 020. 0 2,900. 7 3, 869. 6 5,021.3 5, 858. 1 6, 857. 6 7,458.3 8, 368. 4 1 9, 467. 4 6, 037. 1 5, 737. 7 5,931.0 6,408. 5 6, 895. 6 7,458.3 8, 368. 4 8, 973. 9 1 Estimate. Sources: Worksheets made available by the U.S. Department of Health, Education, and `Welfare, and tables 7 and 10. Such a deflated series is presented in column 3 of table 15 and can be compared with the actual expenditure series in column 1. While between 1900 and 1958 actual expenditures had increased 56-fold, the increase of the salary adjusted series extends merely from about $2.3 billion to $9 billion-less than threefold-and only 5 percent of that actually incurred. The U.S. Department of Commerce's constant dollar series of State and local government purchases extends back to 1929. During 1930-57 it advanced by 88 percent, while during the same period the education cost series in constant dollars advanced distinctly less, i.e., 56 percent. By the way, in the same manner adjustments have been made for the number of pupils in ~average daily attendance. (See col. 2 of table 15.) In terms of 1954 average daily attendance, expenditure increases from 1900 to 1958 would have been twentyfold, i.e., slightly more than one-third of the actual increase. What can be said about expenditures for public, primary, and second- ary education in constant terms? To answer this question an attempt will be made to measure the cost of an education unit, so standardized that its variety and scope are held reasonably constant, and expressed in per pupil in average daily attendance terms. At an earlier stage it was pointed out that since the turn of the century an increasingly large number of auxiliary services-schoolbus transportation, cafeteria, health services, etc.-have been introduced. Their cost has progressed from near zero in 1900 to almost $1.5 billion in 1958 (see table 12). But these auxiliary services have not neces- sarily improved the quality of education itself. PAGENO="0046" 34 ANALYSIS OF RISfl~G COSTS OF PUBLIC EDUCATION If expenditures for auxiliary services are subtracted from the other current expenditures plus debt service, the 1900-1958 increase is only 48-fold (col. 1 of table 16). By adjusting this series for changes in salary level, the increase is further reduced to less than 2~-fold (col. 2 of table 16). TAIILE 16.-Total current expenditure plus debt service for public, primary, and secondary education minus auxiliary services, selected years, 1900-1958 Year Unadjusted annual, thousands of doilars (1) Adjusted for salary level (1954=100), thousands of dollars (2) Per pupil expenditures, dollars (3) Daily per pupil ex- penditures, dollars (4) Daily per pupil ex- penditures In 1954 dollars (5) 1900 1902 1910 1913 1920 1922 1930 1932 1940 1942 1946 1948 1950 1952 1954 1956 1958 193.9 214.2 379. 8 463.0 892.8 1,262.7 1,942.8 2,003.6 2. 012. 7 2,158.9 2, 646. 0 3,565.4 4, 377. 5 5,302.4 6,438.2 7, 750. 3 `9,530.0 2,281.2 2,353.8 2, 990. 6 3,455.2 3,915. 4,140. 5,236. 5,41.5.1 5, 338. 5,479.4 5, 069. 0 5,167.2 5, 562. 3 5,878.5 6,438.2 7, 162. 9 7,760.6 18.233 19. 357 29. 607 34.007 55.281 68.506 91. 359 90. 069 91. 311 102. 655 133.308 170. 511 196. 443 227. 991 251.062 279. 392 319.167 0.126 .134 . 188 .215 .341 .418 .529 .526 . 522 .588 . 754 .960 1. 104 1.279 1.406 1. 570 1.786 1.482 1.473 1.480 1.604 1.496 1.370 1.426 1.422 1. 385 1.492 1. 444 1.391 1. 403 1.418 1.406 1.451 1.454 `Estimate. Sources: Tables 7, 10, 12, 13, and 15. In addition to excluding auxiliary services, it appears proper to use an education unit which is neither affected by the number of pupils in average daily attendance nor the length of the school term. Thus, the education unit which wifi be used is daily public primary and secondary education per pupil in average daily attendance, leaving aside auxiliary services. In column 3 of table 16, per-pupil annual expenditures are given, and in column 4 they are reduced to a daily basis. Finally, in column 5, daily per-pupil expenditures (minus auxiiaiy services) in 1954 dollars are given. This series might be quite appropriate to indicate the cost of a given bundle of public primary and secondary education m constant terms. This series together with the current and constant dollar series is presented in chart 2. Costs in real terms exhibit amazing stability during 1900-1958. For the years for which data. are available, 1922 was the low year with $1.37 daily expenditure per pupil, and 1913 was the high year with $1.60. Over the 58 years, an overall decline of about 3 percent was registered. To the extent that the standardization of the educa- tional unit has been successful and appropriate data in general were used, the analysis can reflect on the productivity variable as a residual. Apparently, productivity m the public schools has changed very little, if at all. This finding appears to bear out the claim made by Burck and Parker some 4 years ago.27 27 Gilbert Burck and Sam Ford Parker, op. cIt., pp. 249, 2.50. PAGENO="0047" ANALYSIS OF RISING COSTS OF PUBLIC EDUCATION 35 For policy as well as projective purposes, there is much interest in measuring the income elasticity of public primary and secondary education, a concept which was briefly discussed above. Measuring income elasticity requires that a net regression coefficient relating expenditures to personal income be computed and then multiplied by the mean personal income of the period divided by the corre- sponding mean expenditures figure. It assumes that expenditures are representatives of the "amount" of education. To the extent that Xe-number of principals, superintendents, and consultants per 1,000 pupils in average daily attendance-is a measure of the quality of education, it should not be included in an equation designed to esti- INCOME ELASTICITY OF PUBLIC EDUCATION PAGENO="0048" 36 ANALYSIS OF RISING COSTS OF PUBLIC EDUCATION mate income elasticity. Income changes must be permitted to find their expression in quality changes, which, in turn, can then lead to expenditure changes. With these considerations in mind, the following hypothesis was enunciated and tested: X1-daily total currently expenditures plus debt service for public, primary, and secondary education per pupil in average daily attendance is a function of, X2-percent of public high school enrollment relative to total public school enrollment, X3-percent of pupils (5 to 19 years old) in average daily attend- ance living in urban areas, and X5-per capita personal income. In brief, Xia=f(X2s~X3, X5). The multiple regression equation was found to have the following values for selected years during 1900-1958: / - 0.004274 0.005200 0.000924 X ia=.112~68+ (.0024) A2 (.0022) (.8715) A5. (1.3) The statistically significant coefficient is /315.23.27 The coefficient of multiple determination adjusted for degrees of freedom lost_R*21 235-is 0.969. It is statistically highly significant at an ~ of 0.05. About 76 percent of the 1900-1958 variation in the cost of daily per pupil current expenditure plus debt service can be explained on the average in terms of per capita personal income change, holding the effect of changes in the public high school-all public school enrollment ratio and urbanization constant. The following beta. coefficients were found: i315.23=.9625, /313.25=-.0565, and 1312.35=-.0474. The following is the simple correlation coefficient matrix: x'~ x2 x~ X, 0.6592 . 5707 0. 9659 . 9746 . 7049 0. 5854 Income elasticity of education is defined as, ~ ~x5 ~ and for 1900-1958 it is + 1.09, just. a little above unit elastic. Thus, it can be concluded that during 1900-1958, a. 1 percent increase in per capita. personal income was on the average associated with a 27 With 17-4 or 13 degrees of freedom lost, coefficients are statistically significant at an ~ of 0.05 if they are larger than 0.514. PAGENO="0049" ANALYSIS OF RISING COSTS OF PUBLIC EDUCATION 37 1.09 percent increase in daily total current expenditures for public primary and secondary education per pupil in average daily attend- ance. It assumes that the effect of the other variables of the equation is held constant. This figure can be compared with income elasticities obtained from cross section data. For example, Solomon Fabricant, using 1942 data of the 48 States, found the income elasticity coefficient of education to be +0.78.29 Fabricant's analysis included two addi- ti onal independent variables--urbanizati on and density. Income elasticity of all State and local government services was +0.90; but if education was excluded, it was somewhat higher, i.e., +0.96. Such services as come under the heading of general control, health and hospital, highways and police, were income elastic, while in addition to education, fire, public welfare, and sanitation were income inelastic. Harvey E. Brazer estimated the income elasticity of education for 40 large cities using 1949 median family income and 1953 per capita operating expenditure data. He computed the income elasticity to be +0.73.~° Other variables included in the analysis were: Ratio of city population to metropolitan areas' population in 1950, Students in average daily attendance per 1,000 of 1950 popula- tion, and intergovernmental revenue per capita, for education, 1953. The present author used 1951-52 and 1954-55 cross-section data to estimate income elasticity of education for the St. Louis City- County area. It was found to be +0.58. In the analysis the follow- ing additional variables had been included- Pupils in average daily attendaiice, I-ugh school pupils in average daily attendance as percent of all pupils in average daily attendance, Pupils in average daily attendance per square mile, and Percent increase in pupils in average daily attendance, 195 1-56. The coefficient of + 0.58 is distinctly below the + 10.1 for fire protection and + 0.98 for police protection found in the St. Louis study. (See table 17.) This study also estimated the elasticity with respect to per capita income of major education expenditure categories, which are also reproduced in table 17. It is interesting to note that, with but one exception, the income elasticity for instruction is lowest. Fixed charges are the exception, where a low income elasticity is not surprising. By definition these charges vary little with income. The fact that the income elasticity of instruction is so low gives cause for concern, as long as the quality of instruction requires great improvement. 29 Solomon Fabricant, op. cit., p. 125. `° Harvey E. Brazer, "City Expenditures in the United States," Occasional Paper 66, (New York: National Bureau of Economic Research, Inc., 1959), p. 58. 47761-59----4 PAGENO="0050" 38 ~i~sis OF RISING COSTS OF PUBLIC EDUCATION TA1~LE 17.-Income 1 elasticities of select local public services, St. Louis City-County area, 1951-52 and 1954-55 Services Income Education: elasticitji Total current plus debt service +0. 56 Total current without debt service +. 52 General control +. 50 Instruction +. 42 Auxiliary services + 1. 15 Plant operation and maintenance +. 55 Fixed charges +. 29 Fire protection + 10. 1 Police protection +. 98 ~1 Since'assessed valuation of real property is highly correlated with income and information about it is much more readily available, the former is used for measurement. The rank correlation coefficient, relating 1949 median income with 1954-55 per pupil assessed valuation of 10 of the districts about coterminous with municipal or census tract boundaries, is +0.905, which is highly significant at an x of 0.05. As was to be expected, income elasticities computed with the aid of cross section data are lower than those based on time series data.3' Actually, depending on the type of data used, the coefficient answered slightly different questions. Income elasticities of education based on cross section data reflect the average percentage change in per pupil education expenditures that is brought about as average per capita personal income changes from one school district to the next by 1 percent within a given year or sets of years. If based on time series data, the coefficient testifies to the average percentage change in per pupil education expenditures that results as average per capita per- sonal income changes over time by 1 percent. The first provides a static picture, while the second reflects dynamic changes in our society. But no matter which of the two concepts is used, there can be little doubt that the income elasticity of public education is quite low. (Especially low appears to be the income elasticity of instruction.) It is low in comparison to income elasticities of other public services and in particular such consumer amenities as air conditioning, auto- mobiles, golf, speedboats, etc. It is also low compared to what it must be if public education in the United States is to be improved. No doubt, people's attitudes toward education and the existing tax system, which relies so heavily on proportional property and sales taxes, are mainly responsible for the low income elasticity of education. The United States can readily afford the expenditures needed to raise the level of education as required by our status as a leading world power; but the necessary funds will only become available if an increas- ing portion of people's income is channeled into public education. This wifi require the combined leadership of all branches of govern- ment and educators. Further shifts will be needed in the responsi- bility of financing education from local to State and possibly Federal Government. ~vIatching Federal funds combined with specific per- formance criteria for eligibility for aid appear to offer a useful approach. Among the States' performance criteria two will be mentioned- tax effort and partial statewide equalization of educational oppor- tunity. Only States that raise a given percent of personal income for education should be eligible. In addition, a certain minimum educa- tion budget needs to be underwritten by the State for each and every child in the State. 31 Marguerite c. Burk, "Some Aualvsis of Income-Food Relationships," Journal of the American Sta- tistical Association, 53 (December 1953), pp. 905-927. PAGENO="0051" ANALYSIS OF RISING COSTS OF PUBLIC EDUCATION 39 THE 1960 AND 1965 OUTLOOK In spite of serious misgivings, 1960 and 1965 projections have been made with the aid of two of the equations given above. The fact that predictive equations have been derived offers flexibility and makes possible later projections by others who might want to change the assumptions. By and large the following projections are based on the same general assumptions that were made by Otto Eckstein in his "Trends in Public Expenditures in the N ext Decade." 1. The degree of international tension is not expected to change much. 2. The political attitudes toward expenditures will not undergo a revolution. 3. Prices are assumed constant at 1958 levels. 4. The economy is to grow at a 3 percent annual rate. 5. The division of functions between Federal and State, and local governments is assumed to persist unchanged.32 Three basic sets of projections will be developed. The first or "low" projection assumes that by and large 1958 conditions will per- sist, except that the number of pupils in average daily attendance will in all three cases increase as proj ected by the National Education Association .~ The "medium" projections assume a cumulative annual increase of 3 percent in the magnitude of per capita personal income, teachers' salaries and length of school term. The other independent variables are assumed to increase very little.34 Finally, the third projections are "high," in that they make rather optimistic assumptions about the future, i.e., cumulative annual increases of 6 percent in per capita personal income, teachers' salaries and length of school term. On the basis of these assumptions daily per pupil expenditures can be projected with the aid of equations 1.1 and 1.3. They will be con- verted into annual total current expenditures plus debt service esti- mates by assuming that the number of pupils in average daily attend- ance will advance from 29,859,000 in 1958 to 32 million in 1960, and 37,200,000 in 1965. Under these assumptions the following sets of proj ections were derived, which can be compared with 1958 expenditures: [In billions] Low projection Medium projection High projection I 32 Otto Eckstein, "Trends in Public Expenditures in the Next Decade" (New York: Committee for Economic Development, 1959), p. 4. 33 Enrollment projections were taken from "Status and Trends: Vital Statistics, Education, and Fi- nance," (Washington, D.C., National Education Association, 1959), p. 9, and with the aid of an enroll- ment-pupils in average daily attendance regression, pupils in average daily attendance figures were estimated. 34 X,-percent of public high school enrollment relative to total public enrollment-is assumed to be 23 In 1960 and 1965, compared to 22.4 in 1958, and X~-percent of pupils in average daily attendance living in urban areas-Is assumed to be 52 in 1960 and 1965, compared to 51.5 In 1958. (1958) 1960 1965 $11.8 $13.8 $11 12.7 17.4 13.8 22.8 PAGENO="0052" 40 ANALYSIS OF RISING COSTS OF PUBLIC EDUCATION Equation 1.1 emphasizes teachers' salary level and, in certain respects, it mainly reflects expenditure conditions. Using equation 1.3, which emphasizes the income side, projections are produced which more nearly reflect the revenue situation. As is to be expected, projec- tions by these two equations show differences. Thus, with the aid of equation 1.3 the following three sets of projections are derived and can be compared with 1958 expenditures: {In billions] (195S) 1960 1965 Low projection $10. 4 $12. 1 Medium projection $11 11. 1 15. 3 High projection 12. 1 20. 3 It is interesting to compare these sets of projections with 1958 expenditures, which amounted to 811 billion; 1960 and 1965 medium projections (in terms of 1958 dollars) emphasizing the expenditure side are 16 and 58 percent, respectively, above the 1958 figure. The corresponding projections emphasizing the income revenue side are 1 and 39 percent, respectively, above 1958. PAGENO="0053" APPENDIX EcoNoxilEs OF SCALE To shed light on the question whether economies of scale are likely in the public schools, a working hypothesis will be established in the light of detailed deductive reasoning. It then will be subjected to empirical tests by using the St. Louis City-County area as a case study. An examination of the operational characteristics of school districts suggests that, within limits, growth need not lead to economies of scale. Here are some of the reasons. A school district can be looked upon as a horizontally integrated government which controls a number of units all furnishing a single service, i.e., education; a unified policy is pursued with regard to all its units. For the sake of efficiency, a school is usually built in a location which minimizes the average distance of its pupils. Growth and evemi consolidation, except in the case of one- teacher schools, usually take the form of more horizontally integrated units. With the advent of consolidation, some local schools will tend to be closed and their functions performed by consolidated schools, particularly high schools. Ilowever, with minor exceptions, the consolidated government will use the existing plant. Time consolidated school district, like any government that adds facilities to already existing ones, will seldom operate under genuine longrun conditions. Thus, growth and consolidation, in the strict sense at least, takes place under quasi-longrun conditions which, depending upon the relative im- portance of fixed factors, and their degree of fixity, can approach longrun condi- tions. A school district in a small community faces a shortrun expenditure function until it reaches a size where an additional school is needed. While very little is known about the shape of the shortrun per pupil expenditure function of a of a school, deductive reasoning suggests that it should have a flat-bottomed shape. The declining phase on the lefthand side is due to the spreading of over- head. The flat bottom can be traced back to a certain amount of flexibility in most schools. The end of the flat bottom will curve up; from that point on there are more students than can be readily accommodated in the existing facilities. But schools will seldom operate in the rising expenditures phase. Location con- siderations, i.e., the distance a child has to travel to school, lead to diseconomies of scale, and in turn will tend to result in the opening of a new school. Schools have indivisible but highly adaptable fixed plants. The law of diminishing returns applies and leads to a U-shaped shortrun expenditure function. Since schools aie basically flexible, the average expenditure function will tend to be reasonably flat. What about the per pupil expenditure function of horizontally integrated schools? If we assume that services of equal quality are rendered regardless of the scale of operations and that the various plants are all of about equal size, have about equal service functions, and tend to be operated at about optimum capacity, schools can be readily added and closed and that factor prices are fixed, the longrun per pupil expenditure function will tend to be horizontal.1 To what extent are these longrun assumptions met? It has been argued above that location considerations result in schools of about equal size; upon inspection it appears that each school has quite similar service functions. There should be a tendency to operate near optimum capacity. But whether the schools are of exactly equal size and operate at optimum capacity is not too important a factor affecting the shape of the longrun per capita expenditure function, as long as the shortrun expenditure function is flatbottomed. Except for financing problems faced by local school districts, adding schools poses few problems. These new schools are additions to existing plants, making for quasi longrun conditions. This is not of major importance, since overhead of hori- zontally integrated school districts is relatively small compared to operating `Don Patinkin, "Multiple-Plant Firms, Cartels, and Imperfect Competition," Quarterly Journal of Economics, 61 (February 1947), pp. 173-205. 41 PAGENO="0054" 42 ~u~sIs OF RIS1~G COSTS OF PUBLIC EDUCATION expenditures. Closing of schools is not much of a problem in most areas, because of rapid population growth. Next, the relative importance of economies and diseconomies which can be associated with different degrees of integration and scales of operation deserves consideration. On the whole, it appears that conditions which help horizontally integrated manufacturers and marketers benefit from net economies-lower factor costs, larger and more efficient plants, and induced circular and vertical integra- tion-do not. exist when local government grows and consolidates. Since schools purchase a highly diversified array of factors, hut virtually none in quantity, few large-scale factor purchases and even fewer significant. price concessions are likely to result. While pecuniary economies are likely to he minor, the only factor purchased in large quantities, i.e., manpower, tends to unionize and produce pecuniary diseconomies.2 Also, the nature of public education, particularly the importance of location, tends to keep schools relatively small. This aspect together with legal restrictions on salary levels and permissible debt will tend to allow only small technological economies. At the same time, serious technological diseconomies can accompany large school districts, which tend to lose efficiency because of political patronage and general administrative top heaviness. On a priori grounds, growing or consolidating school districts can approximate the conditions under which the longrun expenditure function will tend to be horizontal. Since plant and caliber of the school superintendent are virtually fixed, the quasi-longrun function will resemble a U with a fiat bottom over a very wide range. To the extent that relatively little overhead exists, the shortrun and longrun functions tend to approximate one another. They coincide in their fiat-bottomed portion. Net economies are responsible for a negative slope to the left of this area and net diseconomies for a positive slope to the right of it. The more units are horizontally integrated, the flatter the shortrun function. The conclusions of this deductive reasoning have been subjected to some em- pirical verification. With the aid of multiple regression and correlation tech- niques a working hypothesis was tested in relation to the school districts of the St. Louis area. First, a few words might be said about the St. Louis City-County area and its 27 school districts which offer both primary and secondary education. The area has a total population of about 1~i million. St. Louis City with a population of about 850,000 has a single school district, while St. Louis County has 29 districts, 26 of which operate both primary and secondary schools, while the other 3 have merely primary schools. The 1954-55 enrollment of the St. Louis City district was 84,000 while the smallest of the 26 county districts with both primary and secondary schools had an enrollment of 600, and the largest had one of 7,000. Total expenditures per pupil in average daily attendance in St. Louis City was $261 and in the county districts it ranged from $121 to $728. Some school dis- tricts, including that of the St. Louis City, had hardly grown between 1951 and 1956, while in one school district a 225 percent growth had taken place. There existed great differences in assessed valuation of real property per pupil in average daily attendance; the low was $1,500 and the high $24,000. Finally, much varia- tion in the tax rate levied upon property within these school districts existed. The school district of the city of St.. Louis had the lowest rate. In the county school districts it ranged from S1.68 to $3.58 on $100 of assessed valuation. The State of Missouri has a school aid program, under which relatively small subsidies are paid to a district per pupil in average daily attendance. In addition, the equalization phase of the program is designed to assure each district $110 per pupil in average daily attendance. During the period under discussion, the State of Missouri contributed virtually nothing under its equalization program, i.e., less than $60,000 in 1954-55. The virtual absence of State equalization payments made the St. Louis City-County area useful for the purpose of this case study. It excluded many knotty problems connected with identifying determinants of the progress of State subsidy. The following working hypothesis was established in order to learn about the presence or absence of economies of scale in the St. Louis schools: Total current expenditures plus debt service for public primary and secondary education per pupil in average daily attendance is a function of- X2-numher of pupils in average daily attendance in public primary and secondary schools, X3-high school pupils in average daily attendance as a percent of all pupils in average daily attendance, is recognized that teachers throughout~the country are organi~~~, but in the larger school districts also maintenance workers, etc. tend to unionize. PAGENO="0055" ANALYSIS OF RISING COSTS OF PUBLIC EDUCATION 43 X4-number of public school pupils in average daily attendance per square mile, X5-percent increase in public school pupils in average daily attendance, 1951-56, X6-average assessed valuation of real property per pupil in average daily attendance, and X7-quality index of public education in primary and secondary schools. As was mentioned above, the quality index is composed of six subindexes- A is equal to number of teachers per 100 pupils in average daily attendance, B is equal to number of college hours of average teacher, C is equal to average teacher's salary, D is equal to percent of teachers with more than 10 years of experience, B is equal to number of high school credit units, and F is equal to percent of high school seniors entering college. Since testing this hypothesis involves the loss of 8 degrees of freedom, a large loss when merely 27 observations are available, it was thought desirable to work with as large a sample as possible. For this reason, data for two different periods-that is, 1951-52 and 1954-55, with a 2-year interval between, were used. Since expenditures and their determinants increased overtime at somewhat different rates in the different school districts, the cross-section data pertaining to the 2 staggered years are devoid of major serial intercorrelation. In order that a test as to the presence of economies of scale can be made, the following functional relationship was assumed: X1=f(X2, X~, X3, X4, X5, X6, X7). The following results were obtained for total current expenditures plus debt service (X15) and total current expenditures without debt service (Xlb), respec- tively, per pupil in average daily attendance. - 240 966 0.00347 K2 0.000000317 X~ 4.090 X3 1a - (0.0916) + (0.0696) + (0.363) 0.0293 X4 0.335K5 0.0135K6 213.888 X7~ (0.0253) (0.0253)+ (0.679) + (O414)~ The coefficient of multiple determination-R2-is 0.85 and this coefficient ad- justed for degrees of freedom lost_R*i_is 0.82. It is highly significant at an o of 0.05. Thus, about 82 percent of the variation in total current expenditures plus debt service per pupil in average daily attendance is explainable in terms of these six stated independent variables, of which financial ability, service level, and percentage of high school pupils are statistically significant. 0.00181 K2 0.0000000154 X~ 2.999 X3 lb=26.956 (0.0584) (0.0412) + (0.329) 0.0198 X4 0.143X5 0.0116 X6 156.40 X7~ (0.148) (0.131) + (0.697) + (0.375) R2=0.84 and R*2==0.82. They are highly significant at an o of 0.05. Thus, the empirical correlation and regression analysis did not reveal significant economies of scale in the school districts of the St. Louis city-county area. 3 The figures In brackets are partial correlation coefficients. Since there are 54-8 or 46 degrees of freedom, the coefficient is statistically significant at an o of 0.05 in case It is larger than 0.28. Statistically significant regression coefficients are associated with X~, X6,and X7. 4 footnote 3 above. PAGENO="0056" PAGENO="0057" STUDY PAPER NO. 5 TRENDS IN THE SUPPLY AND DEMAND OF MEDICAL CARE (BY MARKLEY ROBERTS) 45 PAGENO="0058" PAGENO="0059" STUDY PAPER NO. 5 TRENDS IN THE SUPPLY AND DEMAND OF MEDICAL CARE (By Markley Roberts) INTRODUCTION AND SUMMARY Medical care is becoming wonderfully effective and appallingly ex- pensive. Rising demand for medical care and shortages of medical personnel and medical facilities require sound public policy decisions to bring about a balance which is fair to those who need care and successful in providing a rising standard of medical services. This study, suggestive rather than comprehensive in scope, is an effort to outline some of the problems which require attention and action. These are public policy problems because the adequacy of medical care vitally affects the general welfare. Improving health has been an important factor in the advance of American productivity, and continuing improvement of health standards will contribute to further economic growth. Also, public policy must recognize the primary public interest in an adequate supply of medical care and must assure that those needs that private agencies and individuals do not meet, and perhaps cannot be expected to meet, will be fulfilled. The American people are increasingly conscious of health problems and eager to share in the advances of medical science. Total spend- ing on health and medical care is approaching $25 billion a year, or about 5 percent of gross national product. The price of this por- tion of the GNP has been rising rapidly, and thus has served to raise consumer prices, and no study of inflation would be complete without examination of this sector. Private spending on medical care is rising, with the biggest in- creases in spending for hospitalization and for prepaid health insur- ance programs. The average family spends 5 percent of its income on medical care, but medical costs are not spread evenly. Spending on medical care rises with rising family income, but declines as a pro- portion of rising family income. Voluntary health insurance covers more than 70 percent of the population, and pays about 25 percent of the Nation's private medical care bill. Federal, State, and local community spending make up about a quarter of total spending for health and medical care. A substantial part of this public spending comes from Federal funds to care for veterans and military personnel and their families. Federal support for medical research is growing and pays an increasing proportion of medical school operating expenses. Hospitals are increasingly utilized in medical care because medical progress and medical services are hospital oriented, because of the prevalence of hospitalization insurance and because of long-term hospital care needs of older people with chronic and degenerative diseases. New drugs have made treatment possible for mental pa- tients in general hospitals and outside hospitals. 47 PAGENO="0060" 48 TRENDS Th~ SUPPLY AND DEMAND OF MEDICAL CARE Organization of medical services is separated from the market price system by a combination of public and voluntary nonprofit supply factors. To achieve greater efficiency, certain aspects of the organiza- tion of supply may need revision, and a trend toward greater efficiency in supplying medical services appears in the growth of group medical practice. Although medical science is raising the quality of medical care, there has been a decline in the ratio of active physicians and dentists to the rising population. Expansion in the number of profes- sional, technical, and auxiliary health workers matches advances in medical science, hut lengthy training, low wages, and low salaries con- tribute to shortages of personnel and high turnover in these occupations. The number of hospitals and hospital beds increased between 1948 and 1959. However, the number of beds per 1,000 population dropped from 9.7 in 1948 to 9.1 in 1958 and hospital utilization rose from 115 to 137 hospital admissions per 1,000 population. Almost half of existing hospital beds are occupied by mental and tubercu- losis patients in State and Federal hospitals, but most hospital care for the general public is provided by short-term nonprofit hos- pitals. These general hospitals are community-owned or under re- ligious or other voluntary associations which depend on hospital fees and on philanthropy for their operation. Expenses increased from $10.04 per patient-day in 1940 to $29.24 in 1958 in the voluntary short- term hospitals, primarily because of rising payroll costs. Demand for higher levels of quantity and quality of medical care can be expected to continue. This demand will generate increasing public support for medical research, medical education, and health facilities. Rising costs of hospital care will continue to increase demand for insurance protection against hospitalization. Demands for protection against other medical care expenses will increase the comprehensivenesss and cost of health insurance. In spite of progress in medical science, the supp1y of available med- ical care, in terms of medical personnel and medical facilities, is declining in relation to population growth and rising health conscious- ness. Shortages of supply exist already and will grow more serious in the future, unless there is a broad national effort of public and private support. Such a national effort will require increased Federal assistance for medical research, medical education, and construction of medical facilities. It will require Federal aid to encourage recruitment, train- ing, and more effective use of auxiliary health workers such as physical therapists, rehabilitation specialists, practical nurses and home aides. \Tolmrtary health insurance, already widespread, will become in- creasingly comprehensive in financing the Nation's private medical care needs. Nevertheless, the greater health care. needs and lower income of the aged population will preclude private insurance plans from solving~ t.he financial side of the medical problems of this group. There will continue to be a need for some form of public participation, probably in the form of social insurance. Other health needs include development of community services to provide home care for the chronically ill and aged, and public assistance for the special problems of disabled children. PAGENO="0061" TRENDS IN SUPPLY AND DEMAND OF MEDICAL CARE 49 CHAPTER 1. DEMAND FOR MEDICAL CARE TOTAL DEMAND Health consciousness Personal and social ethics combine to put reverence for life and health high on the current American scale of values. Higher stand- ards of living, rising levels of education, and widespread public infor- mation about progress in medical science contribute to increasing general public awareness of health and medical care needs, thus stimulating a large and growing desire for health and medical care for individuals and for communities. There is a new awareness of interaction between health of the community at large and the health of its individual members. There is growing realization of benefits achieved for individuals and for society by maintaining a high, rising level of health through effective medical care and preventive health measures. Disease, pain, and disability are increasingly regarded as avoidable and death as postponable. These attitudes are reflected in the concept that a society can buy the level of health it wants for its members if it is willing to pay the price. Such expectations raise the problem of thstinguishmg between needs for health and medical care and the demand for such care. Needs may be large, or even infinite, but demand is a measure of financial ability and willingness to meet the needs. A revolution of rising expectations is in progress in the field of health and medical care. New expectations and new attitudes are raising demand for medical care, increasing utilization of medical services, and generating support for public health activities and medical research. Widespread concern about cancer, heart disease, mental illness, alcoholism, accidents, and other health hazards reflect public awareness of social costs as well as personal and family costs. Widespread concern about the chronic and degenerative health prob- lems of the aging reflects an awareness of the human and social costs imposed by increases in the older age groups of the American popu- lation. Rising expectations stimulate support for medical progress through research. These expectations also support private and public action to raise availability of health and medical care toward higher levels of adequacy. Good health and adequate medical care are regarded as part of a high standard of living. Naturally, individuals want to gain the advan- tages of progress in medical science for themselves and their families, but the American people increasingly appear to believe that their fellow citizens have a right to adequate medical care, although the concept of adequacy may be subject to wide variations. This belief rests primarily on humanitarian welfare considerations,' but there is growing awareness of the economic costs of illness as well.2 1 Herman M. Somers and Anne R. Somers, `Private Health Insurance, Part I, Changing Patterns of Medical Care Demand and Supply in Relation~to Health Insurance." Reprinted from California Law Review, August 1958, vol. 46, No. 3, p. 382. 2 Selma Mushkin, "Toward a Definition of Health Economies." Public Health reports; vol. 73 No. 9 September 1958. 3 But see the views of Eli Ginzberg, "What Every Economist Should Know About Health and Mcdi- cine." American Economic Review, March 1954. pp. 1O4-119;~also see his testimony in hearings before the Joint Economic Committee, Oct. 1,1959. "But we are'coming to the nonproductive areas of medicine. * * * We are moving up against the old age, and what~we are really trading is a kind of easing of life in the terminal years, and really trading one kind of illness for another." PAGENO="0062" 50 TRENDS IN SUPPLY AND DEMAND OF MEDICAL CARE Total spending Total public and private spending on health and medical care indi- cates aggregate effective demand for medical care. Such spending is currently about 5 percent of the gross national product, or close to $25 billion annually in recent years. (See table 1.) Combined private and public health care spending rose from 3.6 percent of gross national product in fiscal 1929 to 4.7 percent in fiscal 1950 and to 5.2 percent in fiscal 1958. Total spending rose from $3.6 billion in 1929 to Whillion in 1950 and to $22.7 billion in 1958.~ It is clear that health and medical care is taking vastly increased amounts of the Nation's output in absolute and relative terms. A rising proportion of total spending on health care comes from public funds. In 1929, public spending, excluding funds for medical facilities, amounted to $414 million. In 1958, public spending was more than $4.9 billion, or about 13 times the 1929 level. On the other hand, private spending on health care in 1958 was more than five times the 1929 level, increasing from $3.1 billion to $17.2 billion. In fiscal 1929, public expenditures by Federal, State, and local gov- ernment units were about 15 percent of the total. Since 1946, public expenditures have been about 25 percent of the total, amounting to $5.4 billion in fiscal 1958. (See chart 1.) The rising amount of spending for health and medical care indicates that the American people want to raise their health standards. Even after allowance for changes in the purchasing power of the dollar, a substantial increase in total demand is indicated by the amount and proportion of medical care spending. Spending on medical care increases as national income rises, but the increase in the proportion of income devoted to medical care also reflects changes in distribution of income as well as new sophistication about the need for health care and about its social and economic value. 4 Ida 0. Merriam, "Expenditures for Health and Medicaircare, 1928-29--1957-58." Department of Health, Education, and Welfare, Social Security Administration, Division of Program Research, Research and Statistics Note No. 26, Sept. 17, 1959. PAGENO="0063" PRENDS TN SUPPLY AND DEMAND OP. MEDICAL CARE 51 CHART 1. Spending for health and medical care. Billions of Dollars 25 20 IS 10 5 0 1928-29 1939-40 Seirce: Social Security Administration, Research and Statistics Nate Na. 26-1959. 957-58 PAGENO="0064" 52 TRENDS IN SUPPLY AND DEMAND OF MEDICAL CARE TABLE 1.-Private and pnblic spending for health and medical care, selected fiscal years 1929-58 [Dollars in millions] 1029 1040 1950 1955 1956 1957 1958 Total spending Private expenditures Healthandmedicalservices Direct payments Insurance benefits Expenses for prepay- ment Industrial in-plant services Philanthropy 7vledlcal facilities construc- tion Public expenditures Health and medical serv- ices Medical facilities construc- tion Total expenditures as percent of gross national product Public expenditures as percent of the total Percent of personal health care spending from- Private spending Insurance benefits Public spending 53, 625.0 83. 914. 7 512, 407. 1 517,764.7 519, 220.6 521, 057.3 522, 737. 7 3, 112. 0 3. 010. 0 3, 023. 0 2,992. 0 9, 042. 0 8, S27. 0 13, 455. 0 13, 130. 0 14. 711. 0 14, 399. 0 16, 082. 0 15, 693. 0 17, 294. 0 16, 785. 0 2, 900. 0 30. 0 80. 0 2, 900. 0 40. 0 52. 0 7, 125. 0 878.0 274. 0 150. 0 400. 0 9,388. 0 2,357.0 595. 0 210. 0 580. 0 10, 176. 0 2,776.0 611. 0 221. 0 615. 0 10, 937. 0 3,245.0 639. 0 232. 0 640. 0 11, 555. 0 3,675.0 645. 0 245. 0 665. 0 102. 0 513. 0 31. 0 891. 7 215. 0 3, 365. 1 325. 0 4, 309. 7 312. 0 4, 509. 6 389. 0 4, 975. 3 509. 0 5, 443. 7 414.0 99. 0 836.6 55. 1 2,780. 1 585. 0 3,923.4 386. 3 4, 148.3 361. 3 4,497.3 478. 0 4,918.5 525. 2 3. 6 14. 2 90.5 9. 5 4. 1 22. S 82. 0 18. 0 4. 7 27. 1 78. 2 8. 1 21. 8 4. 7 24. 3 78. 0 14. 8 22.0 4. 7 23. 5 78. 9 16. 1 21. 1 4. 9 23. 6 79.4 17. 3 20. 6 5. 2 23.9 78.8 18. 1 21. 2 Source: Social Security Administration, Research and Statistics Note No. 26-4959. PRIVATE DEMAND Total personal con-sun4ption spending Total personal consumption spending for medical care gives a gen- eral view of effective private demand for health and medical care, although certainly not an accurate picture of needs or costs. Per- sonal consumption spending for medical care rose 140 percent between 1947 and 1958, from $6.8 billion to $16.4 billion.5 During this same period, gross national product rose 88 percent, from $234 billion to $441 billion. Private medical care spending rose from 2.9 percent of G-NP in 1947 to 3.7 percent in 1958. In terms of total personal consumption spending, private medical care spending rose from 4.1 percent in 1947 to 5.5 percent in 1958, hut remained about .15 percent of consumer spending on all services. The greatest increases in private medical care spending were for hospital costs and health insurance payments. Spending for hospital- ization rose a.bout 280 percent, from $1.1 billion in 1946 to $4.3 billion in 1958. Insurance payments rose 205 percent, from $444 million to $1,359 million. Payments for drugs and prescriptions rose more than 150 percent, from $1.2 billion to $3.2 billion. Payments to doctors rose about 130 percent, from $1.7 billion in 1947 to $3.9 billion in 1958 and payments to dentists rose about 115 percent, from $772 million to $1,674 million. (See chart 2 and table 2.) 5 U.S. Commerce Department, "U.S. Income and Output, Supplement to Survey of Current Business, 1958." Also "Survey of Current BusIness, July 1959," tables 11-4. PAGENO="0065" TRENDS IN SUPPLY AND DEMAND OF MEDICAL CARE 53 Rising national income and new health consciousness provide partial explanations for increases in private demand for medical care. How- ever, increases in private spending for hospitalization and health in- surance indicate the nature of this demand. The increase ill spend- ing for hospitalization reflects the central role of hospitals in supply- ing medical care. The growth of prepayment and health insurance spending represents a new approach to the problem of financing medi- cal care, a new approach to adjust supply and demand. This new approach has created a new component in total private demand for medical care since the end of World War II. Some aspects of this component will be discussed below. 4,000 3,000 2,000 1,000 CHART 2. Personal consumption spending (for medical care). M~lIions of Dollars 5,000 0 1946 47 ~48 `49 `50 `51 `52 `53 `54 `55 56 `57 58 `Source: Commerce Deportment 47761-59-5 PAGENO="0066" L'i (11 10,384 3, 001 1.674 769 ~ t?~J 4, 310 1,359 3,261 ~ 1;iol ~ C) * C) TABLE 2.-Personal consumption spending for medical care [Millions of dollars] 1946 1047 1048 1049 1050 1051 1052 1953 1054 1955 11)56 1957 1958 Total Physicians :DentisLs Other professional services Privately controlled hospitals, nursing homes, and sanitariums Mcdicd care aini hospitalization insur- anco i)rug preparations and sundries Ophthalmic products and orthopedic ap- pliances 6, 104 6,817 7,749 8,051 8,741 9,440 10,172 11,072 11,925 12,827 ]~4, 048 15, 051 1, 720 772 338 1, 163 444 1, 271 306 2, 020 784 390 1, 397 513 1, 313 400 2, 327 900 445 1, 621 559 1, 466 431 2, 338 920 455 1, 772 557 1, 555 454 2, 427 001 482 2, 037 629 1, 719 486 2, 519 997 510 2, 248 641 1, 979 546 2, 657 1, 098 544 2, 486 749 2, 058 580 2, 840 1, 234 586 2, 729 942 2, 137 604 3, 109 1, 406 634 2, 962 1, 056 2, 163 595 3, 189 1, 508 653 ~ 3,229 1,090 2, 473 * 685 3; 470 1, 625 697 * 3, 518 1,055 2, 869 814 * 3, 693 1 705 734 * 3,884 * 1, 064 3, 098 * 873 Source: Commerce Department, ITS. Income and Output. PAGENO="0067" TRENDS IN SUPPLY AND DEMAND O~ MEDICAL CARE 55 Family medical care 8pending The family is the typical consumer unit in private demand for general health and medical care. The "average" family spends about 5 percent of its income on such care. However, the real impact of medical care needs on the family budget appears not in average family spending but in the uneven incidence of. sickness and disability, in the uneven distribution of medical charges and in the uneven ability of families to pay the costs incurred. A nationwide survey, sponsored by the Health Information Founda- tion,6 indicated that 53 percent of all families spent less than 5 percent of their income and 8 percent had no health or . medical expenses. On the other hand, 2 percent of all families incurred expenses exceed- ing 50 percent of their income. More than 40 percent of the total medical costs of all families were incurred by 11 percent of the families, which had medical expenses greater than $495 a year. The 1950 Bureau of Labor Statistics study of consumer spending ~ found an average medical care expenditure by all families of $197 a year, of which $163 was paid directly for medical service and $34 was paid for group health plans and insurance. Average medical care spending ranged from $91 for families with income less than $1,000 to $446 for families with income of $10,000 and over. The more recent 1953 survey of the Health Information Founda- tion reported an average cost incurred for all personal health services of $207 per family, although the median charge was $110. The aver- age charge incurred by families with some kind~ of health insurance was $237 and the average charge incurred by uninsured families was $154 The propoition of income paid out for personal health services ranged from a median of 6.1 percent for families with income under $2,000 to a median of 3.2 percent for families with income greater than $7,500. 8 However, average medical care outlays ranged from 11.8 per- cent for families with income under $2,000 to 3 percent for families with incomes of $7,500 and over. For families with incomes under $2,000 and without health insurance, medical care took an average of 13.4 percent of family income.9 (See tables 3 and 4.) Medical care needs clearly impose a major financial burden on low- income families. Large, unexpected, medical care expenses can have a catastrophic impact on family budgets. Some indication of unmet medical care needs can be found in the pattern of rising family spend- ing on medical care as family income rises. This increase in spending would indicate the high order of importance placed on medical care for the family. However, spending on medical care as a proportion of family income declines as family income increases. This proportional decline would appear to indicate that families achieve some degree of satisfaction of their medical care needs as family income goes up. 6 Odin ;W. Anderson, "Family Medical Costs and Voluntary Insurance: A Nationwide Survey." Mc6iraw-T{ill, 1956. p. xiii. `U.S. Bureau of Labor Statistics, "Study of Consumer Expenditures Incomes and Savings," vol. XVIII. University of Pennsylvania, 1957, p. 9. 8 Anderson, op. cit., p. xii. Ibid., p. 114. PAGENO="0068" 56 TRENDS IN SUPPLY AND DEMAND OF MEDICAL CARE TABLE 3.-Charges incurred by families for health services, by income and health insurance status Family income . Mean gross charges Median gross charges All families Insured Un- insured All families Insured Un- insured Total, all income groups Under $2,000 $2,000 to$3,499 $3,500to$4,999 $5,000 to $7,499 $7,500 and over 8207 130 152 207 259 353 $237 164 168 220 262 362 8154 115 132 167 247 312 $110 54 82 119 176 238 $145 82 103 134 187 255 $63 43 54 83 105 185 Source: 1953 survey by Health Information Foundation and National Opinion Research Center, table A-is in Odin W. Anderson, "Family Medical Costs and Voluntary Health Insurance," p. 113. Footnote omitted. TABLE 4.-Fasniiy outlays for medical care as a percentage of family inconse by income and health insurance status Family income Aggregate family outlay as percent of aggregate family income Median of family total outlay as percent of aggregate family income All families Insured Unin- sured All families Insured Unin- sured Total, all income groups TJnder $2,000. $2,000to$3,499 $3,500 to $4,999 $5,000 to $7,499 $7,500 and over 4.8 11.8 6.1 5.4 4. 7 3.0 4.8 13.4 6.6 5.8 4. 9 3.1 4.8 11.0 5.3 4.2 4. 2 2.8 4.1 6.2 4.0 3.9 3. 6 3.3 4.4 10.2 5.1 4.4 3. 9 3.3 3.0 4.8 2.6 2.2 2. 2 3.0 Source: Table A-16 in Anderson, "Family Medical Costs and Voluntary Health Insurance," p. 114. Footnotes omitted. Growth of prepayment Prepaid health insurance programs cover an increasing proportion of the Nation's private medical care bills. The heavy burden of hos- pital and medical charges on low- and middle-income families, the un- predictability of individual family medical expenses, and the relative predictability of aggregate family medical care costs have contributed to the tremendous expansion during the past 12 years of voluntary health insurance coverage and prepayment programs which now cover more than 70 percent of the American people. (See chart 3.) The growth of these prepayment plans, enabling families to fintince the greater part of their medical care expenses on a regular budget basis, is a major feature of the new pattern of private demand for med- ical care. These programs reflect rising demand for medical care, but they also contribute to rising demand by financing medical care needs, conve.rtrng need and health consciousness into effective demand. PAGENO="0069" 120 100 80 60 40 20 TRENDS IN SUPPLY AND DEMAND OF MEDICAL CARE CHART 3. Growth of health insurance coverage. 57 In 1948, only 60 million Americans were protected by insurance for hospitalization, the most expensive part of modern medical care. By the end of 1958, more than 123 million people were insured against hospital expenses, 111 million were insured against surgical expenses, and 75 million were insured against regular medical costs of nonsurgi- cal physician care.'° (See table 5.) 10 Health Insurance Council, "Greater Security for the American People." Preliminary report on 13th annual survey, voluntary health insurance in the United States as of Dec. 31, 1958. Millions of People Covered 40 Source: Health Insurance Council 0 1946 `47 `48 `49 `50 `51 `52 `53 `54 `55 `56 `57 4TTO1-59-6 PAGENO="0070" 58 TRENDS IN SUPPLY AND DEMAND OF MEDICAL CARE TABLE 5.-G'rowth of health insurance coverage, number of people protected against hospital expenses, surgical expenses, and regular snedical expenses [In thousands~ End of year Hospital expense Surgical expense Regular expense medical End of year Hospital expense Surgical expense Regular medical expense 1940 1941 1942 1943 1944 1945 1946..... 1947 1948 1949 12,312 16,349 19.695 24,160 29,232 32. 068 42,112 52,584 60.995 66, 044 5,350 6,775 8,140 10,069 11,713 12. 890 18,609 26,247 34,060 41, 143 3,000 3,100 3,200 3,411 3.840 4. 713 6.421 8,898 12,895 16, 892 1950 1951 1952 1953 1954 1955 1956 1957 1958 76,639 85,348 90,965 97,303 101,493 107, 662 115,949 121,432 123,038 54,156 64,892 72,459 80,982 85.890 91, 927 101,325 108,931 111,435 21,589 27,723 35,670 42,684 47,248 55. 506 64,891 71,813 75,395 Source: Health Insurance Council. Insurance benefits now pay almost a quarter of all private medical care bills. Direct "out of pocket" private spending for medical care dropped from 88.7 percent in 1948 to 72.5 percent of total private medical care spending in 1957. Insurance benefits for hospital serv- ices rose from 6 to 15.3 perce.nt of private medical care spending, and insurance benefits for physicians' services rose from 2 to 7.8 percent." Private medical care spending rose from 4 percent in 1948 to 4.9 percent of per capita disposable income in 1957. During this period, direct "out of pocket" per capita payments for medical care remained fairly stable, but insurance benefit payments for medical care rose more than 250 percent, from 0.3 percent in 1948 to 1.1 percent of per capita disposable income in 1957.12 Direct "out of pocket" payments for doctors' services as a percent- age of per capita disposable income dropped 26 percent from 1948 to 1957, but insurance payiiients to doctors rose 375 percent. In 1948 insurance benefits were only 6.4 percent of private spending for doc- tors' services; by 1957 insurance benefits were paying 30 percent.'3 Voluntary health insurance, which accounted for about one-quarter of private hospital service spending in 1948, covered more than half of private spending for hospital care in 1957. Insurance benefits paid for more than 40 percent of private spending for hospital and doctor services in 1957 compared to 15 percent in 1948. One dramatic indication of tile growth of prepayment programs appears in the 470-percent increase in benefit payments made by vol- untary health insurance, from $606 million in 1948 to $3,474 million in 1957. Payments into voluntary health insurance programs rose 380 percent, from $862 million in 1948 to $4,144 million in 1957.'~ Tile declining rate of overhead expenses, which represents the cost of get- ting the insurance protection, reflects in part the savings achieved by enrollment of large groups in health insurance programs but pressures of rising utilization and opposition to increases in health insurance premiums have also forced insurance carriers to reduce the gap between earned income from prepayments and outgoing benefit payments. ~` Agnes W. Brewster, "Voluntary Health Insurance and Medical Care Expenditures: A 10-Year Re- view," Social Security Bulletin, December 1958, p. 9. 12 thid., table 2, p. 10. 13 Ibid., table 6, p. 13. 14 IbId., table 3, p. 11. PAGENO="0071" TRENDS IN SUPPLY AND DEMAND OF MEDICAL CARE 59 Collective bargaining is one of the m~Ljor factors contributing to the growth of health insurance coverage and medical care prepayments. In 1945 an estimated 500,000 workers were covered by prepaid health programs under collective-bargaining agreements. By 1957 more than 12 million workers were covered.'5 Nonwage fringe benefits negoti- ated during World War II and exemption of health benefits from Korean War wage freezing encouraged growth of these industrial health prepayment programs and expansion of their coverage. As the proportion of workers covered by these programs has increased, more attention is centered on broadening the coverage of existing plans to workers' families and to retired workers. Also, more atten- tion is given to problems of adding new benefits and raising the level of benefits. The growth of industrial health prepayment programs and other voluntary group enrollment for health insurance reflects the new awareness of the value of medical care and awareness of the need to budget unpredictable medical costs. The demand for health insur- ance is part of the total demand for medical care, and they rise to- gether as family income rises. These are underlying factors behind the rising proportions of total and individual medical care spending channeled into prepayment insurance programs, which pay an increas- ing part of the Nation's private medical care bill. Health needs of older people Older people need more health and medical care at the stage in life when they can least afford to pay for this care. Longer life spans, brought about by progress in preventive and therapeutic medicine, have changed the Nation's demand for health services, with emphasis shifting from control and treatment of communicable, infectious dis- eases to the chronic and degenerative diseases. The three leading causes of death at present are heart diseases, cancer, and vascular lesions of the brain-all characteristic health problems of older age groups. Nearly 16 million Americans, or about 9 percent of the population, are now over 65 years of age. By 1970 there will be an estimated 20 million persons over 65, and, if research in disease and health prob- lems of the aged progresses at a rapid rate, the estimate for 1970 may be raised closer to 30 million.'6 People over 65 are particularly af- fected by the cost of medical care, since the chronic and degenerative health problems cause more long-term disability, greater need for medi- cal services, and longer hospital stays at a time when income is sharply reduced by retirement. (See table 6.) Three-fifths of the men and women past 65 received less than $1,000 in cash income last year. Five major points on health needs of the elderly are listed in the summary of discussions this year before the Senate Subcommittee on Problems of the Aged and Aging: 1. Most older people in this country do not have hospital and surgi- cal insurance enabling them to get health services readily and at rea- sonable cost, although in recent years voluntary health insurance has increased among persons over 65. `5 U.S. Bureau of Labor Statistics, "Analysis of Health and Insurance Plans Under Collective Bargaining, Late 1955." Bulletin No. 1221, November 1957, p. iii. 16 U.S. Senate, Committee on Labor and Public Welfare, "The Aged and Aging in the United States, Summary of Expert views Before the Subcommittee on Problems of the Aged and Aging," 86th Cong. 1st sess., 1959, p. 6. `7 Ibid. PAGENO="0072" 60 TRENDS IN SUPPLY AND DEMAND OF MEDICAL CARE 2. Public and private efforts are needed to preserve health and functional ability in old age. 3. Older people spend more on health care and are less able to afford it than younger age groups. 4. Three basic principles in organization of medical care are (a) high quality of medical care, (b) a satisfactory payment system to support professional, hospital, and supplementary services, and (c) administration guaranteeing proper standards and economy of service. 5. Most older people in the future could readily acquire paid-up health insurance if it were provided through the social security mecha- nism. These general considerations lead to the conclusion that the present health care needs of our older citizens are not being met adequately. Furthermore, their health care needs are a. growing national problem. Although about 40 percent of the population aged 65 and over has some kind of health insurance, there is less opportunity for the "over 65" group to get private health insurance for a number of reasons. Elderly people are poor risks for private insurers because their health needs are greater than the needs of younger age groups. Higher pre- miums charged for greater risk are more difficult to pay when income is reduced by retirement. However, even when elderly people are will- ing and able to pay the premiums, they frequently are unable to get private health insurance coverage. Further consideration of these greater health care needs and reduced ability to get insurance protec- tion appear below in the discussion of hospital utilization by older people. TABLE 6.-Health problenw of people aged 65 and over, 1957-58 Experience Rate per 100 persons per year Persons aged 65 and over Persons under age 65 Persons of all ages Restricted activity days Bed disability days including hospital days Days in short-stay hospitals Incidence of acute conditions Persons with 1 or more chronic conditions Persons with activity limited by chronic conditions Persons injured Physician visits Dental visits 4, 730 1, evo 178 163 76 42 25 680 80 1 743 697 76 269 38 7 28 514 164 2 000 780 85 260 41 10 ~ 530 160 Source: Tabulation from U.S. National Health Survey in "Hospitalization Insurance for Old-Age Survivors, and Disability Insurance Beneficiaries," report submitted to the Committee on Ways and Means by the Secretary of Health, Education, and Welfare, Apr. 3, 1959, p. 13. PUBLIC SPENDING Public support for health and medical care A major component of total demand for health and medical care is public spending, which is about a. quarter of total spending for this purpose. Public funds from Federal, State, and community sources, almost $5.5 billion in fiscal 1958, are spent for veterans, military per- sonnel, hospital construction, medical research, maternal and child health services, medical vocational rehabilitation, public assistance medical payments, school health programs, and various other public PAGENO="0073" TRENDS IN SUPPLY AND DEMAND OF MEDICAL CARE 61 programs providing health care for such groups as Indians, merchant seamen, civilian residents in the Panama Canal Zone, narcotics ad- dicts, and the residents in prisons and penitentiaries.18 The largest public outlays are for State and local public hospitals and medical care, about $1.8 billion, and the next major outlay is Fed- eral spending for veterans and military personnel and their families, about $1.5 billion in fiscal 1958. Veterans' hospitals and medical care accounted for about $800 million, hospital and medical care adminis- tered by the Department of Defense accounted for about $580 million and Medicare assistance for medical services in community hospitals to eligible military dependents accounted for about $85 million. (See table 7.) TABLE 7.-Public spending for health and medical care, selected fiscal years, 1929-58 [In millions of dollars] 1929 1940 1950 1955 1956 1957 1958 Total Health and medical services General hospital and medical care Defense Department facilities - -- Medicare Veterans' hospital and medical care Public assistance (vendor medi- cal payments) Workmen's compensation medi- cal benefits Temporary disability insurance medical benefits Medical vocational rehabilitation Maternal and child health ser- vices School health programs Medical research Other public health activities- - - - Medical facilities construction 513. 0 891. 7 3,365. 1 4,309. 7 4, 509. 6 4, 975. 3 5, 443. 7 414.0 836.6 2, 780. 1 3, 923. 4 4, 148. 3 4,497. 3 4,918.5 215. 0 30. 0 30. 0 25. 0 5. 0 9. 0 100.0 415. 0 45. 0 72. 0 90. 0 4 13. 7 17.9 3. 1 179.5 1, 174.0 332.0 582. 2 193. 0 2. 5 7.4 29. 7 30. 6 55. 0 373. 7 1,449. 5 602. 6 722. 0 211. 9 315. 0 20. 6 9. 2 93. 4 66.3 105. 9 327. 0 1, 577.8 548. 0 723. 5 252. 6 335. 0 22. 7 11. 0 104. 8 74. 2 115. 9 382. 8 1, 707. 0 529. 3 24.7 732. 9 287. 6 355. 0 25. 8 12. 7 113. 8 81. 0 183. 0 444. 5 1,881. 5 584. 7 86.6 794. 1 320. 2 370.0 31.9 14. 9 122. 7 87. 8 237. 5 386.6 99. 0 55. 1 585. 0 386. 3 361. 3 478. 0 525. 2 Source: Social Security Administration, Research and Statistics Note No. 26-1959. Federal support for research Federal support for medical research is growing in amount and im- portance. Total medical research spending rose from $88 million in 1947 to $479 million in 1958. The Federal share rose from 32 percent in 1947 to 56 percent in 1957, but dropped to 48 percent in 1958 al- though the Federal funds for medical research increased from ~186 million in 1957 to $230 million in 1958.19 Medical research in uni- versities is particularly dependent on Federal funds. Two-thirds of university medical research funds came from the Federal Government in 1957.20 The 1958 report to the Secretary of Health, Education, and Wel- fare by a panel of consultants, headed by Dr. Stanhope Bayne-Jones, indicated that Federal support for research and specialized training provides a rising proportion of total budgets of medical schools. 18 Merriam, op. cit. Also, Federal Hospital and Medical Care Programs, Special Analysis 3, reprint of pp. 1004-1010 from the Budget of the U.S. Government for the fiscal year 1960. 19 U.S. Department of Health, Education, and Welfare, "The Advancement of Medical Research and Education." Final Report of the Secretary's Consultants on Medical Research and Education, June 27, 1958, p. 22. Hereafter cited as Bayne-Jones report. 29 IbId., p. 22. PAGENO="0074" 62 TRENDS IN SUPPLY AND DEMAND OF MEDICAL CARE A more recent study of 20 medical schools for the National Insti- tutes of Health indicates that Federal funds made up 9 percent of medical school budgets in 1948, about 29 Percent in 1958, and that by 1970 an estimated 44 percent of medical school budgets will come from Federal grants.~ These figures do not include construction grants, which would raise the proportion of Federal support for medical education. Assuming an annual growth in gross national product of only 3 percent, the Bayne-Jones report projects a desirable rise in national spending for medical research from $330 million in 1957 to $900 mil- lion or $1 billion a. year by 1970. Attainment of this level of spending for medical research will require private industry and private philan- thropy to triple private spending for medical research, the report states, and Federal spending for medical research should rise to about $500 million a. year in 1970. Although the report stresses the need for increasing Federal support., it also states clearly that non-Federal spending must continue to support about. half of the total medical research effort.22 Widespread awareness of the value of research has created strong public support for medical research. Discovery of an effective polio- myelitis vaccine has raised hopes for other major medical break- throughs. It is significant that. in 1959 Congress raised appropria.- tions for National Institutes of Health research activities by $106 mil- lion a.bove the Administration budget request to $400 million. This action reflects public support for medical research. This demand for progress in medical research can be expected to continue to grow. DEMAND FOR HOSPITAL SERVICES Trend to hospitals for 1~ecilth care Hospitals are focal points for advances in medical science, and the America.n people increasingly demand hospital services for diagnosis and treatment of disease, as well as for preventive medicine and com- munity health education. Increased utilization of hospitals stems in part from the social and teclmological role of hospitals in providing medical services, but rising income and prepaici health insurance programs provide additional incentives toward utilization. Hospital utilization is also determined by age, sex, marital status, education, a.nd a variety of less measurable socio-economic factors, in- cluding existing alternatives to hospital care such as family or home care or care by nonhospital institutions such as nursing homes and halfway homes for mental patients. Among the most significant factors, however, are the prevalence of hospitalization insurance, the increase in older age groups, and the incidence of long-term diseases and mental illness. Hospitalization insurance The most widespread benefit of health insurance is protection against costs of hospitalization. Evidence of higher hospital utiliza- tion by insured persons raises the question of abuse of this protection. It is difficult to establish abuse, which implies doctor-patient collusion, 21 U.s. Department of Health, Education, and Welfare, "A Study of 20 Medical Schools," a report to the Director, National Institutes of Health, from the staff committee on support for research and training, April 1959, p. 6. 22 Bayne-Jones report, pp. 29-30. PAGENO="0075" TRENDS IN SUPPLY AND DEMAND OF MEDICAL CARE 63 but it is possible to indicate some factors involved in utilization of hospital care by the insured and the uninsured. The 1953 Health Information Foundation survey 23 revealed that hospital admission rates vary little between different income groups, but that insured persons, regardless of income, had higher admission rates-14 admissions per 100 insured persons per year, compared with 9 admissions per 100 uninsured persons per year. The survey also indicated utilization of 100 hospital days per 100 insured persons pear year and 70 hospital days per 100 uninsured persons per year but with an average length of stay in hospitals of 7 days per insured person and 8.3 days per uninsured person. Among the uninsured, there was little variation in hospital utiliza- tion, regardless of family income. Among insured persons, however, individuals from low-income families (less than $2,000 a year) had a much higher hospital admission rate than any other income group-21 hospital admissions per 100 persons. This rate is far above the gen- eral average admission rate of 14 per 100 insured persons per year and 9 per 100 uninsured persons per year,24 but this high utilization by low income individuals may reflect the large number of older retired people in this income group. In all income groups, hospital utilization by insured persons ex- ceeded utilization by the uninsured, even at very high income levels. Apparently, families with health insurance go to hospitals more often and stay a shorter time than do those without such insurance. This higher utilization by the insured does not imply excessive use of hospitalization insurance protection, although it does indicate that insured persons (or their doctors) use hospitals for less serious con- ditions. If at least some of these less serious conditions are presumed to require hospital care, it would appear that there are unmet hospi- talization needs and lack of health consciousness among uninsured people. The lack of health consciousness is indicated by the failure of uninsured persons to demand more hospital care as family income rises. Whether insured persons get more hospital care because they have hospitalization insurance or whether they get hospitalization insur- ance because they want hospital care, it is clear that insured persons are more health conscious. The significance of health consciousness in hospital utilization is confirmed by the observation that insured persons, regardless of income or insurance benefits, spend more for "out of pocket" medical expenses than do the uninsured. As health insurance benefits become more comprehensive, covering "out of hospital" medical care, it is possible that the high rate of hospital utilization by insured persons will be reduced. If a patient must be hospitalized to get the benefit of health insurance, both patient and doctor may have ,some incentive to use hospital care to ease financial burdens on the patient. On the other hand, compre- hensive health insurance, which covers outpatient hospital visits and doctor visits to the patient's home or patient's visits to the doctor's office, may reduce hospital utilization, but will also raise questions concerning overutilization of nonhospital medical services. There exists already a growing trend toward greater comprehensiveness of 23 Anderson, op. cit., pp. xiv-xv, 55. 24 Ibid., p. 58. PAGENO="0076" 64 TRENDS tNT SUPPLY AND DEMAND OF MEDICAL CARE benefits in health insurance coverage, including home calls, visits to doctors' offices, and diagnostic X-ray and laboratory fees. However, the effect of comprehensive health benefits on hospital utilization has not yet been clearly established. Utilization by older people Older people need more hospital care than younger age groups. Those past 65 use more than twice as many days of hospital care as those under 65. But the surge in demand for hospital care begins before 65, affecting those in the final years of their active working lives. People. in the 55-to-64 age bracket use hospitals at a rate three times as great as those in the 35-to-44 age group.25 About 60 percent of the over-65 population do not have any hos- pital insurance. Insurance coverage for the aged increases with in- come, but the low incomes characteristic of this age group indicate that these people who need hospital insurance coverage the most have the lowest proportion of protection. According to Prof. Wilbur Cohen, an authority on welfare and problems of the aging: These are the people who, as they need hospital care, even though they are receiving social security, must ultimately apply for public assistance to supple- ment their social security to pay for their hospital care; or not receive hospital care; or borrow money; or receive it from relatives or friends; or get free care in some way from the communities.28 About half of the. uninsured persons in the over-65 age group reported that they could not afford or were refused health insurance, and half indicated that they had not consiclere.d or did not want such protection.27 It would appear, therefore, tha.t even apart from lack of financial ability, older people are less aware of their health care needs and potential benefits from modern medical science. There may also exist a reluctance to get medical care for what older people may regard as inevitable ilineses of old age.28 Since hospital and nursing home care imposes heavy medical ex- penses on older citizens, the lack of tax-supported health insurance and the failure of private health insurance programs to meet their needs for protection leaves this portion of the population in a particu- larly weak, vulnerable position in financing their hospitalization needs. Mental illness and c/ironic diseases Mental illness and chronic diseases are raising the demand for long-term hospital care. To a large extent, this trend in demand for hospital care reflects the conquest of communicable, infectious dis- eases, with consequently longer life expectancy and resulting increases in health problems of the older age groups. Mental and tuberculosis patients occupy almost half of all available hospital beds, but the demand for long-term-care facilities appears even more strikingly in the 94 percent average occupancy reported in 1958 by the long-term, non-Federal psychiatric hospitals, many of which have waiting lists. In contrast to this high occupancy, voluntary, short-term, general, 25 Ray E. Brown, "Forces Affecting the Community's Hospital Bill." Reprinted from "Hospitals," Journal of the American Hospital Association, Sept. 16, 1958, vol. 32, p. 3. 26 U.S. Senate, "Hearings Before the Subcommittee on Problems of the Aged and Aging of the Com- mittee on Labor and Public Welfare." 86th Cong., 1st sess., June 1959, p. 26. 27 "Progress in Health Services," Bulletin of the Health Information Foundation, January 1959. 26 "Progress in Health Services," Apr11 1959. PAGENO="0077" TRENDS IN SUPPLY AND DEMAND OF MEDICAL CARE 65 and special hospitals reported an average occupancy of only 76 per- cent.29 Mental illness is a heavy financial burden on the Nation with mini- mum annual costs estimated at more than $2.4 billion a year.3° The average length of stay for a patient in a mental hospital is about 1 year and 10 months. More than 80 percent of all mental patients are in State mental hospitals and about 10 percent are in veterans' hospitals. Mental and tuberculosis hospitals have traditionally been operated at taxpayer expense, in part because of the high cost of the necessary long-term care and in part because of a public interest in protecting the community from mental and tuberculosis patients. However, medical progress is reducing the need for tuberculosis facilities and tranquilizing drugs are changing the pattern of long-term care for mental patients and are easing the need for facilities for confinement of mental patients. Nevertheless, long-term care will remain beyond the financial capacity of most families. The cost per patient-day in non-Federal psychiatric hospitals in 1958 was only $4.40 compared to $29.24 in voluntary short-term general hospitals, but despite this lower cost per day, a patient's length of stay would impose staggering burdens on most families. General hospital care for patients with long-term diseases is costly and clearly not an economic use of limited hospital resources. The increasing burden of chronic and degenerative diseases accentuates the need for alternative facilities, nursing homes, and long-term-care insti- tutions which can furnish skilled care for patients who need medical attention but who do not require the intensive care and treatment and expensive services included in the overhead costs of general short- term hospitals. CHAPTER 2. SUPPLY OF MEDICAL CARE ORGANIZATION OF MEDICAL SERVICES The visiting family doctor with his little black bag no longer holds the center of the stage in supplying medical services. The drama of medical care has shifted to the hospital operating room, where teams of specialized physicians, assisted by an ever-increasing number of highly trained assistants, technicians, and laboratory experts, co- operate in complex techniques of surgery, radiology, pathology, and many other specialized medical skills. In order to utilize the new techniques and new skills achieved by medical science, the supply of medical services must be organized. Changes in organization of these services is taking place not only in hospitals, but also in the private practice of physicians, particularly through the growth of group practice. The complexities of organization reflect the pluralistic nature of supply in the medical care field. Private physicians' services for fees are supplemented by free services to the indigent as charity, by industrial inplant services, by hospital outpatient clinics, by public 29 "Hospitals: GuideIssue,"J.A.H.A.,Aug. i,1959, p. 384. 3° Rashi Fein, "Economies of Mental Illness." Joint Commission on Mental Illness and Health, 1058, p.xu. PAGENO="0078" 66 TRENDS IN SUPPLY AND DEMAND OF MEDICAL CARE health services, and by a. variety of combinations of private, public, and semipublic voluntary, nonprofit meclica.l services and health facili- ties. Direct payment for medical services is supplemented by third- party payments through health insurance, public assistance. Govern- ment programs such as Medicare and by philanthropy. Similarly, the suppiy of hospitals. medical schools. and other health facilities are determined by a wide range of private, public and semipublic agencies. These agencies make decisions and perform functions vitally affecting the general welfare, at times without recognition of the public inter- est or without enlisting public support.. Group medical practice and specialization affect the general prob- lems of supply and efficiency of physicians. As a means of coordinat- ing specialized physicians' services, economizing on laboratory facili- ties and administrative expenses. group practice of medicine may well be a partial solution to the shortage of medical doctors. The trend to group practice brings another intervening step in the financial relations between doctors and patients, since partners in group prac- tice are likely to get a fixed income plus "profit sharing" from the group's income, rather than direct fees from patient to doctor. In the field of hospital organization, requirements for expensive equipment and additional skilled personnel have diminished the role of small, proprietary hospitals and have augmented the central role of the large public and voluntary general hospitals. Hospitals are developing progressive patient care with patients moving from inten- sive care units to less costly intermediate care units and home care ~rograms as rapidly as possible. Hospital organization requires coordination in supplying services to patients, as well as coordination in supplying educational opportunities for medical students, interns, and auxiliary medical personnel and coordination of research activi- ties. This coordination has become a specialized field with graduate training for hospital a dlministrators. The new trends in organization of medical services undoubtedly affect the supply of these services but there is great difficulty in meas- uring the effect. The Pressure of rising costs in supplying medical services encourages any trend to greater efficiency, and further devel- opments in organization of medical services can be expecte.d as a result of this pressure, in spite of resistance from traditionally conservative professiona.l organizations. QUALITY CHANGES IN MEDICAL CARE The quality of medical care improves as medical research provides new, more effective techniques for prevention, diagnosis, and treatment of disease. However, there is no simple, single index of quality changes in meclica.l care.3' The trainedi, competent Physician has a primary role in providlmg high quality medical care. Quality reflects the caliber of medical edu- cation and research, specialization in medical practice, andl progress in hospital administration. It also reflects health education of the pub- lic, consumer financing practices, community socia.l and economic pat- terns, aiid many other factors. The quality of medical service a~ physician is able to provide depends also on the drugs, equipment, and supplementary personnel available to him andi to the community lie serves. 31 U.S. Veterans' Administration, "Report of the Committee on Measurement of the Quality of Medical Care," Department of Medicine and Surgery, April 1959, p. 27. PAGENO="0079" TRENDS IN SUPPLY AND DEMAND OF MEDICAL CARE 67 Rising life expectancy, reduced mortality rates, and control of many communicable diseases are indications of rising quality of medical care, but they are difficult to disentangle from questions about the quantity or the availability of health and medical service or from improved diet and improved living conditions. Although judgments on quality of medical care are generally subjective and elusive, nevertheless, they are pertinent in considering the more measurable factors of supply, the supply of trained medical personnel, and the supply of hospital facilities. SUPPLY OF MEDICAL PERSONNEL Physicians (M.D.) Present trends indicate that population growth is outstripping the supply of medical doctors, the most important group supplying medical care to the American people. Since 1949, the ratio of M.D. physicians has declined from 135 to 132 per 100,000 population, al- though their total number has risen from 201,000 to more than 227,000. In 1930, the ratio was 125. (See chart 4 and table 8.) The national totals fail to reflect the number of physicians who are retired or partially retired, and thus understate the physician-popula- tion ratio. Furthermore, the national average covers wide geographic differences. In New England and the Middle Atlantic States, there are about 160 physicians per 100,000 population, but there are less than 100 physicians per 100,000 population in the South Central States. In 1959, about 6,900 medical students were graduated from the Nation's medical schools. The output of physicians will have to ex- pand to 9,100 a year from medical schools in the United States, plus another 750 from foreign medical schools, if the 1957 ratio of 132 physicians per 100,000 population is to be maintained in 1970.32 There are currently eighty-one 4-year medical schools and four 2- year medical schools. The graduating class of each 4-year school aver- ages about 90 students. Medical schools must increase in number and capacity to provide adequate personnel for medical research, educa- tion, and maintenance of current physician-population ratios. More physicians can be produced by existing medical schools. Since the attrition rate in these schools results in under-utilization of exist- ing clinical facilities during the final 2 years of medical education, 2-year preclinical medical schools can provide students to fill the existing clinical facilities. Also, experiments are underway to shorten the time required to produce physicians by strengthening premedical training and by speeding the process of medical training. However, the Bayne-Jones report points out that even with such trends underway, from 14 to 20 new medical schools will have to be built to maintain a ratio of 132 physicians per 100,000 population and to meet research and educational needs. Such a program, requiring $500 million to $1 billion for construction of new medical schools, may well require substantial Federal aid. The report states: Unless there is a marked change in sociai philosophy leading to private gifts or State appropriations on an unprecedented scale, large Federal appropriations will be required.~ Without the large number of foreign-trained physicians serving the American people, the physician-population ratio would have de- 32 Bayne-Jones report, p. 34. ~ Ibid., p. 36. PAGENO="0080" 68 TRENDS IN SUPPLY AND DEMAND OF MEDICAL CARE dined much more seriously in the postwar years. More than 10 per- cent of new physicians added to the total supply in this country since 1950 were educated outside the United States. About 400 of the 1,300 foreign medica.1 school graduates added to the supply in this country were American citizens studying abroad, but the rest were aliens. Thus the output of medical schools in the United States was supple- mented by the equivalent of 10 to 15 medical school graduating classes from abroad. Such dependence on foreign-trained medical doctors clearly indicates inadequate output of physicians by American medi- cal schools. Dentists The number of dentists in the United States increased from 81,000 in 1947 to about 100,000 in 1958, thus maintaining the ratio of 57 to 59 dentists per 100,000 population which has existed since 1949. How- ever, the national average fails to exclude an estimated 12,000 dentists who are retired or active outside the dental profession.3~ Also, the national average includes dentists in Federal service, almost 7,000 in 1958. As in the case of physicians, there is a wide range in distribution of dentists. New York had a 1958 ratio of 85 dentists per 100,000 population and South Carolina a ratio of 20 dentists per 100,000 population.35 Increased use of auxiliary dental personnel such as technicians and dental hygienists has resulted in a better supply of dental services, but needs for dental research and adequate dental services in the future will require a greatly increased number of dentists. The Nation could use 13,500 more dentists now, and unless more dentists per year are produced, the dentist-population ratio will continue to decline. While four new dental schools have been established during the past 2 years, more new schools are needed. Indeed, two new schools, each with a graduating class of 50, would be required each year between 1057 and 1070 to reestablish the 1055 ratio of 1 dentist for each 1,000 persons.n The American Dental Association has made a similar estimate. Pub- lic Health Service estimates, based on projected current levels of dental schools output, indicate a decline in the supply of dentists to 52.8 per 100,000 population in 1970 and 50.2 per 100,000 population in 1975.~~ Dental schools in the United States have graduated about 3,000 students a year since 1952. Nearly all of the 45 dental schools reported that they are operating at full capacity and would apply for Federal matching construction grants to expand teaching facilities, if such funds were available.38 (See table 8.) Nurses In contrast to continuing declines in the physician- and dentist- population ratios, the supply of active professional nurses has been rising and is expected to continue rising to provide higher nurse- population ratios in the future. (See table 8.) Nurses are the largest single group of health workers and nursing service is fundamental in supplying adequate hospital care. About GO 34 U.S. House of Representatives. "Medical School Inquiry," staff report to the CommitLee on Inter. state and Foreign Commerce, 85th Cong., 1st sess., 1957, p. 435. 3~ Ibid., p. 436. 35 Bayne.Jones report, p. 37. 37 U.S. Department of Health, Education, and Welfare, Health Manpower Source Book, sec. 9, PhysI- cians, Dentists, Nurses. Public Health Service Publication No. 263, 1959, p. 58 34"Medlcal School Inquiry," p. 71. PAGENO="0081" CHART 4. Supply of physicians, dentists and nurses, in relation to population, with projections on the basis of currently predicted output of medical, dental, and nursing schools. Rates per 100,000 population ________________________________________________________________ 300 200 00 Physicians ~ ~, C - Denfists~ ~ I I___ = I~_~_1_____~_I I I (f2 0 1946 1950 1955 960 Sources: Health, Education and Welfare Indicators. Projections from Health Manpower Source Book, Section 9. 1965 970 1975 PAGENO="0082" 70 TRENDS IN SUPPLY AND DEMAND OF MEDICIAL CARE percent of active professional nurses are employed in hospitals. Hos- pitals also supplement the relatively expensive services of gradua.te professional nurses, who have 3 or 4 years of professional training, with the less expensive and less highly trained services of practical nurses, nursing aids, and other auxiliary nursing personnel. Estimates of the number of graduate professional nurses range above 800,000, but many are retired permanently or temporarily from active employment. Active professional nurses totaled 460,000 in 1958, a ratio of 268 per 100,000 population. On the basis of cur- rently predicted output of nursing schools, the supply of active pro- fessional nurses will rise to 525,810 for a projected ratio of 269 per 100,000 population in 1965, and to 608,050 for a projected 1970 ratio of 284 per 100,000.~~ In 1958, about 1,100 professional nursing schools graduated 30,000 students. Increasingly, nursing education includes a 4-year collegiate degree program, although the present pattern is still j~redominantly a 3-year diploma program. Hospitals bear the major burden of financing nursing education, a burden estimated at two-thirds to three-fourths of the total cost of nursing education. Therefore, tuition fees must be supplemented by general university funds in collegiate nursing schools and by the institutions which operate nurs- ing schools as part of a general hospital program. The 2-yea.r 30,000 increase in the number of active professional graduate nurses from 1956 to 1958 tends to confirm the observation that almost half of nursing school graduates drol? out of professional life within a. few years after graduation. In spite of the preference for matrimony to employment, large increases in the number of 17- year-old girls, who comprise most of the students admitted to nursing schools, will maintain a rising supply of active professional nurses, according to Public Health Service proj ections. TABLE 8.-Physicians, dentists, and nurses: Trends and projections of supply Number of persons Rate per 100,000 population Physicians Dentists Active professional graduate nurses Physicians Dentists Active professional graduate nurses 1910 1920 1930._ 1940 1947 1948 1949 1950 - 1951 1952 1953 1954 1955 1956 1957 1958 19601 1965 1 1970 1 19751 135,000 144,977 153,803 175,163 201,277 203,400 205,500 207,900 210,900 214,200 218.061 221,700 226,625 239,350 259. 950 279, 000 296,100 37,684 56,152 71,055 69,921 82, 990 89,441 91.638 93,726 95.88.3 97, 529 99,227 100. 534 101,623 100,266 106. 735 112, 881 118,142 50,500 103,900 214,300 284,200 375,000 401,600 430, 000 460. 000 475,190 525, 810 608, 050 146.0 137.0 125.0 133.0 135.0 134.0 133.0 132.0 132.0 132.0 132.0 132.0 132.0 132.9 132. 8 130. 5 125.9 41.0 53.0 58.0 54.0 58.0 59.0 58.0 59.0 59.0 59. 0 59.0 59.0 57.0 55.6 54. 5 52. 8 50.2 55 98 174 216 251 249 258 268 264 269 284 1 Projections based on currently predicted output of U.S. medical, dental, and nursing schools plus grad- uates from foreign medical schools. Source: Health, Education, and Welfare Indicators, August 1958; Health Manpower Source Book Section 9-Physicians, Dentists, Nurses, Public Health Service Publication No. 263. 3~ Health Manpower Source Book, sec. 9, p. 73. PAGENO="0083" TRENDS IN SUPPLY AND DEMAND OF MEDICAL CARE 71 Other health workers As the functions of medical care become increasingly specialized and expensive, there has been great expansion of the professional and technical occupations in the health field to supplement and often to take over functions previously performed by physicians, dentists, and nurses. To some extent, this trend reflects the substitution of less expensive medical care for the more expensive services of physicians, dentists, and nurses. However, the growing complexity of medical care and equipment requires more workers with specialized profes- sional and technical training. More than 1.8 million workers were employed in health-related occupations in 1955. Of this total, about 750,000 were medical doc- tors, dentists, and professional nurses. The remaining auxiliary med- ical workers providing health services included: 40 Student nurses 113, 000 Practical nurses 175, 000 Hospital attendants 146, 000 Nursing aids 138, 000 Pharmacists 111, 000 Medical laboratory technicians 50, 000 Dental technicians 21, 000 X-ray technicians 50, 000 Dietitians, nutritionists 22, 000 Chiropractors 25, 000 Osteopathic physicians 12, 000 Veterinarians 17, 000 Medical record librarians 7, 000 Medical social workers 5, 000 Psychiatric social workers 5, 000 Physical therapists 7, 000 Occupational therapists 6, 000 Speech, hearing therapists 4, 000 Rehabilitation counselors 2, 000 Hospital and medical program directors 9, 000 In addition to the occupations listed above, office assistants to physi- cians and dentists totaled about 130,000. Nonprofessional attendants, orderlies and ward maids in hospitals and other health institutions numbered about 200,000. New paramedical occupations, requiring additional training and new skills, tend to appear with advances in medical science. The major training field is necessarily the hospital, which is a focus for all health training, but increasing emphasis by professional organiza- tions on maintenance of high standards within their specialty has resulted in longer training and greater reliance on university graduate degree programs for training. Lengthening of the training period tends to discourage people from entering these occupations and also raises the income reqUirements of those completing the training. Thus, at the same time that rising professional standards push up the cost of paramedical services, relatively low salaries in paramedical occupations continue to deter recruitment of enough trained workers to overcome persistent shortages in these occupations. Apart from doctors and dentists, about 9 out of 10 professional health workers are women.4' Thus, a characteristic employment prob- 40 U.s. Department of Health, Education, and welfare. Health Manpower Chart Book, Public Health Service Publication No. 511, 1957, p. 1. 41 Walter L. Johnson, "Personnel Shortages in the Health Field and working Patterns of women." Public Health Reports, January 1957, vol. 72, No. 1, p. 61. PAGENO="0084" ~72 TRENDS IN S13P?LY AND DEMAND OF MEDICAL CARE lem in the nursing profession, early retirement from professional ac- tivity, is a. significant factor in the supply of other professional health personnel. The high proportion of women in these professions helps to explain the persistent personnel turnover imposed by marriage, child raising, and female attitudes toward employment. The predominant number of women may also help to account for the relatively low pay scales in hospitals and health service employment. However, low wages and salaries are also determined by the institutional arrange- ments and organization in the health industry, where public institu- tions depend on tax funds and nonprofit agencies depend on philan- throphy and voluntary donation of services. The voluntary nonprofit hospitals and the general public community hospitals are usually un- able to pay fully competitive salaries, even for the professionally trained auxiliary medical personnel Therefore, these occupations are filled with women whose salary expectations are less optimistic than the expectations of men with similar professional training. Although the supply of personal health services could be increased by greater use of family services for patients in hospitals and by greater utilization of volunteer workers, there is no clear trend toward increased charitable or volunteer service to ease the foreseeable demand for professional and auxiliary health workers. The supply of such workers is inadequate in almost all fields, and shortages of profes- sional auxiliary health workers appear likely to continue for the fore- seeable future. SUPPLY OF HOSPITAL FACILIT~S Increases and shortages The supply of hospital facilities has grown by more than 10 per- cent in the last decade, but rising population and rising hospital uti- lization continue to put severe pressures on the available supply of hospital facilities. The American Hospital Association reported 6,786 hospitals in the continental United States with a total of 1,572,000 hospital beds in 1958. In 1948, there were 6,160 hospitals with 1,411,000 beds. In spite of these increases, however, the number of hospital beds per 1,000 population dropped from 9.7 in 1948 to 9.1 in 1958, and hospital utilization rose from 115 to 137 hospital admissions annually per 1,000 population.42 The decline in the average length of patient stay in general and other special hospitals tends to counteract pressures on the supply of hospital facilities. These pressures result in earlier discharge of patients and occasionally in refusal to admit patients. However, the average length of sta.y in general short-term hospitals fell from 15.4 days in 1932 to 11.1 days in 1948 and to 9.6 days in 1957. This would indicate that hospital care is changing so as to enable existing facilities to serve more people more quickly. Among the factors involved in declining length of patient stay are the concentration of expensive and effective diagnosis and treatment in the first few days of a hospital stay, tile trend to early ambulation of maternity and surgical patients, and tile development of "progressive patient care" which moves pa- tients from intensive care units to intermediate, less intensive and less expensive hospital care units and early transfer to home care status. Increasing reliance on tranquilizing drugs for mental patients enables 42 Hospitals: Guide Issue, I.A.H.A., Aug. 1, 1959, p. 384. PAGENO="0085" TRENDS IN SUPPLY AND DEMAND OF MEDICAL CARE 73 hospitals to give "out patieiit" analysis and treatment to cases which formerly had to be institutionalized. However, apart from these trends which reflect progress in medical science and reduce pressures on the supply of hospital facilities, con- siderations of supply also involve problems of distribution of facili- ties, as well as the nature and cost of hospital services. Types of hospitals The predominant type of hospital supplying medical care to the general public is the non-Federal, short-term general hospital. In 1958, there were 5,290 hospitals in this category with 610,000 beds. Voluntary, proprietary and State and local public hospitals are in- cluded in this category. Long-term hospitals include mental institu- tions, tuberculosis hospitals, and hospitals intended primarily for care of aged or chronic disease patients. However, in terms of all hospital beds, about two-thirds are owned by Federal, State, or local governments. The rest are under volun- tary or proprietary ownership, although proprietary ownership is declining. Three-fourths of the governmentally owned beds are in mental, tuberculosis, and veterans' hospitals.~3 Thus, Federal and State hospitals tend to be large, long-term care institutions, whereas general, short-term hospital care is provided by the private, nonprofit or community hospitals. Of the 5,290 non-Federal short-term general hospitals in 1958, 3,203 were voluntary, 896 were proprietary, and 1,191 were owned by State and local governruents. (See chartS and table 9.) 43 Health Education and Welfare Indicators, August 1958, p. 19. 47761-~9-'~ PAGENO="0086" 74 TRENDS ]~ SUPPLY AND DEMAND OP MEDICAL CARE CHART 5. Ownership of hospital beds. NOn -Pro fit -~--- Federal / - Proprietary -~ I ~I I I i~_ Source: Health, Education and Welfare Indicators TABLE 0.-Ownership of hospital beds [In thousands] Total Governmental Nonprofit Proprietary Federal State Local 1930 1940 1946 1947 1948 1949 1950 1951 1952 1953 1954 1955 1956 1957 956 1,226 1,436 1,400 1,411 1,435 1, 456 1,522 1, 562 1,581 1,578 1,604 1, 608 1,559 64 109 236 200 186 187 189 215 213 203 189 183 184 183 405 572 812 808 826 842 844 871 897 711 718 739 728 686 151 193 (2) (2) (2) (2) (2) (2) (2) 204 202 203 202 195 336 298 335 342 349 355 368 353 399 409 417 427 443 448 (1) 54 53 51 50 51 55 54 54 54 52 52 50 47 1 Included with nonprofit hospital beds. 2 Included with State governmental beds. Source: Health, Education, and Welfare Indicators, August 1958. Thousands of Hospital Beds 1,000 State and Local Governments 800 600 400 200 0- 1947 1950 955 957 PAGENO="0087" ThE~ThS IN SIJPPLY AND DEMAND OF MEDICAL CARE 75 When function rather than ownership is considered, the impact of mental illness appears startling in relation to the total supply of hospital beds. Almost half of existing hospital beds are occupied by mental patients. In 1957, there were 4.6 beds in general and other special hospitals per 1,000 population and 4.2 mental hospital beds per 1,000 population. The 4.6 rate of general hospital beds per 1,000 has ic- mained about the same since 1947, but the 4.2 rate of mental hospital beds per 1,000 in 1957 represents a drop from 4.7 beds per 1,000 in 1947. Non-Federal psychiatric hospitals contained 646,000 beds for mental patients and Federal psychiatric hospitals another 67,000 beds for mental patients. On the basis of the Public Health Service standard of 5 beds for the mentally ill per 1,000 population, there is clearly a shortage of mental beds, and the shortage would appear even greater if "nonacceptable" mental beds were excluded from the existing total. However, new emphasis on community and home treatment, out- patient mental health clinics and centers are easing the pressures on existing mental hospitals, although shortages and overcrowding of mental facilities continue. According to estimates reported to the Public Health Service, only 55 percent of total needs are met by exist- ing acceptable mental beds.44 The need for long-term care facilities to supply medical services for the rising aged population has brought new emphasis on the role of nursing homes. The best of these nursing homes provide skilled nurs- ing care and related medical services to patients who do not require the expensive, intensive care provided in general hospitals. The worst are venal and dangerous, extracting profit from the small pensions of elderly people crowded together in obsolete firetraps. A 1954 survey of nursing homes and related facilities by the Public Health Service indicated a total of about 25,000 nursing homes with about 450,000 beds.45 Only 180,000 beds, however, were provided with skilled nursing care. On the basis of needs reported to the Public Health Service, existing acceptable skilled nursing home beds are meeting only one-fourth of the needs. Thus, existing needs and future needs of the expanding aged popu- lation indicate an inadequate supply of long-term care facilities, but a national effort to increase the supply of nursing home facilities can alleviate the need for general short- and long-term hospital care. Distribution of hospitals Wide regional differences are included in the national average of 4.6 g~neral hospital beds per 1,000 population. New England and Northwest States have about 6 beds per 1,000 in their central cities; Southeast States, 5.7; Central States, 5.2; and Middle East, South- west and Far West central cities have about 4.5 beds per 1,000 popu- lation.4~ In the 375 hospital service regions reported in Hill-Burton State plans, general hospital bed availability varies from 1.6 beds per 1,000 population for 31 regions to 6.2 acceptable beds per 1,000 population for 20 regions. In rural States, where population density is low, more small hospitals with relatively low occupancy are needed, 44 U.S. Department of Health, Education, and Welfare, "The Nation's Health Facilities: 10 Years of the Hill-Burton Hospital and Medical Facilities Program, 1940-56." Public Health Service Publication No. 616, 1958, p. 77. 45 Ibid., pp. 83-85. 46 Ibid., pp. 33-34. PAGENO="0088" 76 TRENDS IN SUPPLY AND DEMAND OF MEDICAL CARE and adequate hospital care would require 5.1 to 5.5 beds per 1,000 population.47 Acceptable beds in areas outside central cities vary from 0.51 general beds per 1,000 in the Southwest to 1.86 in New England, with a national average of 1.42 beds per 1,000 population outside central cities.48 Hospital services and costs A vast range of services are provided by hospitals. Primarily, these are personal services, but they also require an arra.y of expensive equipment. Payroll expenses are the major hospital operating expense, taking about two-thirds of total costs and about two-thirds of the average cost per patient da.y. In the psychiatric hospitals, payroll expenses account for $3.08 of the $4.40 expenses per patient day. In the volun- tary, short-term hospitals, payroll expenses account for $17.71 of the $29.24 expenses per patient day. The American Hospital Associa- tion survey for 1958 listed 1.5 million full-time hospital workers with an overall average of 111 hospital workers per 100 patients and a range from 2~4 workers per 100 patients in voluntary short-term general hospitals to 34 workers per 100 patients in non-Federal psychiatric hospitals. (See table 10.) More than 90 percent. of short-term non-Federal hospitals have emergency rooms and operating rooms, clinical laboratories, basal metabolism apparatus, diagnostic X-ray equipment and electrocardio- graphs, the hospital survey indicated. Other expensive equipment such as electroencephalographs, radioactive isotope therapy, X-ray therapy, intensive care units, postoperative recovery rooms, and pre- mature nursery facilities are increasingly prevalent, as are such hospi- tal services as physical and occupational therapy as well as various skilled services, including medical and psychiatric social service. The addition of new, expensive equipment does not result in sav- ings on hospital payrolls. Instead, this equipment creates new pay- roll expenses. As newly developed diagnostic and treatment equipment is added to hospitals, more-not fewer-people are required to operate it. Hospital equipment is ex- pensive, its cost is impressive, but the enduring element of cost for these new services is the new trained personnel who must accompany it.49 In addition to the rising expense of professional and other highly trained workers, hospitals are faced with payroll pressures from the general wage level on costs of lower paid, unskilled workers. Al- though there have been significant increases in hospital wage scales and reduction of hours worked, there is still a considerable lag be- hind the general wage level of industry. WTages averaging $1,330 a year in 1946 for a 48-hour week rose to $2,873 in 1957 for a 42-hour week, an increase of 116 percent in pay and a decrease of about 14 percent in hours.5° But continuing requirements for hospitals -to match general wage and salary levels will put a long-run pressure on hospitals costs. 4' U.S. Department of Health, Education, and Welfare. "How Many General Hospital Beds Are Needed?" Public Health ServIce Publication No. 309, 1953, p. 62. "The Nation's Health Facilities," p. 49. 4' Russell A. Nelson, Maryland-District of Columbia-Delaware Hospital Association. Statement before Maryland Insurance Commissioner, May 26, 1958, (mimeo) p. 16. `° Frank Groner, American Hospital Association. Statement prepared for U.S. House of Representa. tives, Ways and Means Committee, hearings on Federal Unemployment Tax Act, Apr. 10, 1959, (mimeo) p.4. PAGENO="0089" TRENDS IN SUPPLY AND DEMAND OF MEDICAL CARE 77 This pressure, reinforced by the difficulty of introducing machinery and automation to replace personal services, will increase costs of hos- pital care by an estimated 5 percent annually for many years in the future.5' TABLE 1O.-Hospital personnel and payroll costs Number 1 Number 1 per 100 patients Hospital payroll per patient-day Hospital expense per patient-day All hospitals, total: 1946 1948 1950 829,000 938, 000 1,057,000 1,118,000 1,245,000 1,374, 000 1,401, 000 1 465,000 361,000 427,000 473, 000 73 76 84 84 93 101 107 111 156 173 191 $2.93 3.60 4. 79 5.63 6.83 7.98 8. 72 9.63 $5.11 7. 57 9.40 $5.21 6. 35 7.98 9.14 10.67 12. 16 12.96 14.74 $10.04 14.06 16.89 1952 1954 1956 1957 1958 Voluntary, short term general hospitals: 1946 1948 1950 1952 485, 000 184 11.22 19.55 1954 567,000 638,000 207 213 13.67 15.23 22.78 24.99 1956 1957 1958 680,000 720, 000 218 224 16.14 17. 71 26.81 29.24 I Full-time personnel plus full-time equivalents of part-time personnel. Excludes resident physicians, interns and students. Source: "Hospitals: Guide Issue," JAHA, Aug. 1, 1959, vol. 33, pt. 2, hospital statistics table 1. Hospital construction The Hill-Burton Federal aid program to stimulate construction of hospitals and other medical facilities was approved by Congress in 1946. Since that time, spending for hospital construction has ranged from about $650 million to $950 million a year. Hill-Burton funds helped to finance about 25 percent of all non-Federal construc- tion of health facilities, particularly in rural Southern States pre- viously having serious shortages of hospital facilities. (See table it) Hospital construction, slowed by the depression of the 1930's and by World War II, has been tremendously expanded by the Hill- Burton program, which aided construction of 135,000 new hospital and nursing home beds plus 750 units for outpatient care such as public health centers, clinics, and diagnostic centers during the first 10 years of the program.52 It is worth noting that this program, relying on State plans, provides Federal aid without regard to the public or private nature of the ownership of the projects to be aided, but requires that the operation of facilities be nonprofit. This is believed to be the first major example of Federal assistance to non- public groups, for public ends. Such action was found essential to a compre- hensive program, because of the dual natur~ of the existing hospital system, which has evolved to a large degree under private auspices.13 Such an approach is a frank recognition of the semipublic nature of privately owned medical facilities. Some indication of future hospital construction needs can be found in the estimates of State agencies concerning specific projects to be Ii Ray E. Brown, "The Nature of~Hospita1 Costs." Reprinted from Hospitals, J.A.H.A., Apr. 1, 1956, ~*12 "The Nation's Health Facilities," p. 19. 13 Ibid., p. 15. PAGENO="0090" 78 TRENDS IN SUPPLY AND DEMAND OF MEDICAL CARE undertaken if funds become available.. These estimates forecast. hos- pital construction amounting to $700 million a year and an additional $300 million a year for long-term care, including chronic care hos- pitals, nursing homes, rehabilitation facilities and outpatient diag- nostic or treatment centers.54 These estimates indicate unmet needs, and public, support appears to be strong for further Federal assist- ance in t.his program. In 1959 Congress, in the same appropriation bill which increased funds for medical research. also increased funds for Federal gra.nts for hospital construction to $186 million, $85 mil- lion more than requested by the achninistration. TABLE 11-Hospital construction, value put in place and value by source of funds [Millions of current doilarsj Value of new hospital construction put in place 1 Value of hospital construction by source of funds Total Public 2 Private Federal Non-Federal Without Federal aid Federally aided Federal share as percent of federally aided 1940 1949 1950 1951 1952 1953 1954 1955 1956 1957 87 679 840 947 867 682 697 673 626 858 54 477 496 528 473 365 360 322 298 333 33 202 344 419 394 317 337 351 328 525 5 169 146 132 113 66 35 22 37 45 82 386 466 569 532 434 529 553 467 560 124 228 246 222 182 133 98 122 253 31.8 39. 7 35.4 33.0 27. 1 28. 2 25.8 25.0 29. 7 1 Includes construction of health-related institutions, such as nursing homes. 2 Includes construction of federally owned hospitals. 3 For projects approved and construction begun under the Hill-Burton hospital and medical facilities con- struction program. Source: Health, Education, and Welfare Indicators, August 1958. CHAPTER 3. THE PRICE INDEX OF MEDICAL CARE TRENDS The medical care component of the Burea.u of Labor Sta.tistic.s- Consumer Price Index, currently weighted at. 5.3 percent of the total CPI (see table 12), gives an indication of price, reactions to de- mand and supply relations. During the period 1947-58, only two items in the medical care "market basket" increased more than the "all services" index: hospital room rates and group hospitalization insurance ~ All other services in the "market basket" of medical services increased less than the average for all services. (See chart 6 and table 13.) Medical care services generally rose about. 58 perc~nt from 1947 to 1958, but hospital room ra.tes rose more than 125 percent a.nd group hospitalization premiums rose ahnost 80 percent from 1951 to 1959. When hospital rates and hospitalization insurance prepayments are removed from the medical care price index (see table 13), tile rise in ~14-Ibid., p. 32. PAGENO="0091" TRENDS IN SUPPLY AND DEMAND OF MEDICAL CARE 79 the medical care price index from 1947 to 1958 is 35 percent, only slightly above the all items CPI increase of 29 percent during the same period. CHART 6.-Medical care price index and selected components. ________________ I947-'49~ 100 200- ___ Hasp/Ic! Rates \ ~0 *0~ ~ Physicians .000 Group Hospitalization Fees Dec. I950~l00 5~ ____ I L~I I 947 950 955 958 Source: Bureau of Labor Statistics The index of fees for general physician fees rose by only 44 percent, dentist fees by 38 percent, specialist fees by 35 percent and surgeon's fees by 27 percent. Expressing the increases as average annual rates without com- pounding, the rise in the price of hospital rooms was 7 percent from 1947 to 1958; for group hospitalization insurance premiums it was 6.5 percent. No other component of the medical care price index rose as fast. General physician and specialist fees rose by only 3 percent a year during this period. Thus, hospital expenses and hospitalization prepayments are clearly a. major factor in the rapid increase in the price index of medical care. Introduction of payments for group hospitalization among the items priced for the medical care price index in December 1950 was a justi- fied recognition of the rapid growth of voluntary health insurance, which in turn reflected a rechanneling of demand for increasingly expensive hospital care. However, as noted below, the rise in costs of prepayment is based not alone on hospital costs, but also on in creased hospital utilization. The trend in group hospitalization premiums since December 1950, when first included in the CPI, has closely paralleled the trend in the room rate-the major cost item covered in such plans. However the average annual increase in group hospitalization premiums during these 6 years (1950-56) was 7.5 percent corn- pared with 6.4 percent for hospital room rates, reflecting not only higher hospital costs but also greater utilization.55 55 Elizabeth Langford. "Medical care in the Consumer Price Index, 1936-56." Reprint from Monthly Labor Review, September 1957, p. 2. PAGENO="0092" 80 TRENDS IN SUPPLY AND DEMAND OF MEDICAL CARE Drugs and prescriptions, the nonservice component of the price index of medical care, account for about one-sixth of the total medical care weight in the CPI "market basket." The price experience for this component has been quite different from that of medical services. The drug price index went up about 26 percent, 2 percent a year, compared to the all-items rise of 30 percent, or 2.3 percent a year, and the medi- cal services price rise of 58 percent, or 4.1 percent a year, from 1947 to 1958. Therefore, it would appear that increasing prices of drugs and prescriptions, the commodity portion of the price index of medi- cal care, are relatively less significant in comparison to the increases in the prices of medical services. TABLE 12.-Relative importance of medical care price index components as percent of consumer price index all-items total 1947-49 average January 1950 December 1952 December 1957 All items Medical care Medical care (excluding drugs) General practitioner Officevisit Home visit Obstetrical care Surgeon: Appendectomy Tonsillectomy- Dentist Filling Extraction Optometric examination and eyeglasses-- - - Hospital services Men's pay ward Semiprivate room Private room Group hospitalization Accident and health insurance Prescriptions and drugs Prescriptions, narcotic and nonnarcotic~~ - - Penicillin tablets Multiple vitamin concentrates Aspirin Milk of magnesia Tincture of iodine 100. 0 100.0 100.0 100. 0 3. 3 5 2 5. 1 5. 3 2. 9 4. 4 4. 2 4. 4 1. 1 1. 3 1. 6 1. 6 .6 . . 1 .6 . . 2 .7 .` . 2 .7 `~ . 2 .1 . 1 .1 . 1 .2 .1 .2 .1 .7 1.2 .8 .8 .5 .2 .9 .3 .6 .2 .6 .2 . 1 . 6 . 2 . S .3 . 2 3 . 3 . 2 1. -~ .2 3 . 1 i . 1 f . 1 .2 .4 1. 0 . 8 1. 0 . 9 1. 1 . 9 . 2 . 1 . 1 .4 .1 . 2. . 1 . 3 . 1 .2 .2 . 1 .3 . 1 .2 .2 . 1 Source: "Relative importance of CPI components, 1957," Monthly Labor Review, July 1958, reprint No. 2287. LIMITATIONS The price index of medical care is useful for describing price trends and comparing medical care price trends with changes in prices of other goods and services. This index also can serve as a price deflator or an approximate indicator of the changing purchasing power of the medical care dollar. However, tile index does not show changes in the quality and quantity of medical care purchased. Consumer spending surveys may give an indication of quantity, but the index-making process precludes measurement of quality changes, even if these qual- ity changes were not extremely elusive. PAGENO="0093" TRENDS fl~ SUPPLY AND DEMAND OF MEDICAL CARE 81 Higher hospital rates reflected in the price index, for example, may indicate greater concentration of expensive hospital services during the first few days of a patient's hospital stay. Since 1946, the average length of stay in general hospitals has dropped from 13.5 to 9.5 days, a decline which might be considered a 30 percent increase in hospital efficiency or in the quality of hospital services. The index of hospital room rates must cover a variety of demand and supply factors, there- fore, which have a resultant upward pressure on the index. Some of the limitations of the medical care price index stem from the sampling techniques used. The CPI is determined by the prices of goods purchased by urban wage earners and clerical salaried work- ers and thus fails to indicate rural medical care spending costs. There is a presumption that urban medical care prices give an accurate pic- ture of total medical care cost trends, but such assumptions need further confirmation. A recent Health Information Foundation study lists four limita- tions of the medical care price index.56 First, the quality of priced services are not uniform. However, since the index measures price trends rather than specific prices, this limitation does not vitiate va- lidity of the index. A second limitation is that relative weights of items in the medical care index may not reflect changing patterns of medical care purchases. Thus, the items priced for the CPI may not be representative of the medical goods and services actually pur- chased by consumers. Third, the index does not measure changes in the product. Finally, the weight of the medical care component in the CPI may not reflect the overall importance of medical care within the CPI, since consumer purchasing patterns change over time. The Bureau of Labor Statistics makes surveys of consumer spending to revise and update the sources of the index, but the process is unend- ing since consumer purchasing patterns are continuously changing. Subject to its limitations, however, the price index of medical care does indicate price trends for the major components of medical care and shows the relation of these price trends to other consumer price trends. These trends may indicate changing demand and supply rela- tions and thus suggest the problem areas for public and private action. ~ Harry J. Greenfield and Odin W. Anderson, "The Medical Care Price Index." Health Information Foundation Research Series No. 7, 1959, pp. 7-8. PAGENO="0094" 82 TRENDS IN SUPPLY AND DEMAND OF MEDICAL CARE cc4 cc cc cc cc cc cc -~ .~ cc cc cc .4_c ~` cc~ cc cc H cc C -C cc C cc cc ccc cc c-cc cc ccc ccc ccc ccc ccc c-cc cc ~` cc cc cc ~ cc ~ ccc cc ccc c- ~ ~ ccc c- ccc ccc- c- ccc cc cc cc-cc c-cc ccc ccc cc - ccc ccc c-c cc cc cc ccc cc ~ ~ cc ccc cc ccc ccc ccc ccc - cc-cc cc ~- cc ccc ccc cc cc cc-cc cc cc c' - ccc- ~- cc cc ccc ccc- ccc cc cc-cc cc cc - ccc cc ~- c-cc-cc cc cc cc ~- ccc ccc ccc ccc cc ccc cc ~- cc ccc ccc ccc ct- cc ccc ccc cc ccc cc cc ~ - c'cc ccc cc ~ cc cc ccc ccc cc c-c-cc-cc cc cc cc ccc ccc-c-ccccccccccc-cccccc"' cc ccc cc cc cc ccc ccc - cc cc ccc ccc c-c cc- - ccc- cc ccc ccc- - ccc cc © cc c-cc cc cc cc ccc cc cc ccc ccc cc ccc ccc cc ccc cc cc ccc cc-c ccc cc cc c-cc cc cc cc ccc cc ccc cc ccc ccc cc ccc cccccc- ~ cccccc- ccc cccccccc c~~cc cccc~ ~_S cc cc ccc ccc ccc ccc cc ccc cc ccc ccc - cc cc ccc cc ccc cc ccc ccc cc cc ccc---- ccc cc cc cc--cc cc~ cccccccccc-c c-cccccc ~c~~c:c~ -~c3 c~~d ccccc ccccccccc ccc---c ccccc cc ccc cc cc ccc- cc ccc cc cc ccc - ccc- ccc- ccc ~ cccccc ccccccccccc~ccccccccccccc cccc- `cc `cc ~ ~ ~ Ccc c).C C cc cc cc :.~. k~ccJC ~: :~-~` IC ,~c~3c~cd' `-~`-~C -~ c~ ~ ~ ~ ~Ccc ccc cc,- .. ~ ~- cc~ ~ c.cccc~C-~c cc ~C'~-c cc~ ~ PAGENO="0095" TRENDS IN SUPPLY AND DEMAND OP MEDICAL CARE 83 CHAPTER 4. A LOOK AHEAD EXPANDING DEMAND FOR MEDICAL CARE 1. The increase of the population of the United States to an esti- mated 196 million in 1965 and 214 millinn in 1970 will raise demand for a greater quantity of medical care. Rising national income and rising per capita income, bringing rising standards of living, will also raise the demand for a greater quantity of medical care. 2. Increasing awareness of health and medical care needs will stimulate demand for a greater quantity of medical care and a higher quality of medical care. Rising expectations for improved medical care generate support for medical research and public action to increase the availability of medical care. 3. Rising demand for medical care generates private and public action to finance medical care. Private action appears in the growth of voluntary health insurance, which now pays about 25 percent of the private medical care bill and will pay an increasing proportion in the future. Public action appears in tax-supported programs to provide medical care for special segments of the population and in tax-supported medical research programs. INADEQUACIES OF SUPPLY 1. The supply of medical doctors in relation to the population has been falling. Since 1949, the ratio of M.D. physicians declined from 135 per 100,000 to an estimated 132.7 per 100,000 population in 1959. The Public Health Service has estimated ~ that the ratio will fall to 130.5 per 100,000 population in 1970 and to 125.9 medical doctors per 100,000 population in 1975, if the number of graduates of U.S. medical schools does not increase above the level currently predicted. 2. The supply of dentists in relation to population has been stable, but the Public Health *Service estimates the ratio will drop from 57 dentists per 100,000 population currently to 54.5 dentists per 100,000 population in 1965, to 52.8 in 1970, and to 50.2 in 1975, on the basis of currently predicted output of dental school graduates.58 3. The supply of health facilities and the supply of hospital beds judged acceptable by Public Health Service standards are inadequate in terms of rising population and increased demand. To meet the needs of the prospective population, 265,000 hospital beds will be required within the next 10 years in addition to the existing 1.5 mil- lion hospital beds. The supply of nonhospital health facilities offering alternative services for health and medical care, such as public health centers, nursing homes, mental health clinics, and physical therapy and diagnostic centers, can alleviate the shortage of hospital beds; at the present time, the backlog of demand for these alternative health facilities far exceeds their supply. CONTINUING PRICE INFLATION 1. Expanding demand for medical care and continuing inadequacies in supply of personnel and facilities providing medical care have con- tributed to rising prices of medical care services. This price inflation ~7' Health Manpower Source Book," sec. 9, p. 31. ~ Ibid., p. 58. 5~ "The Nation's Health Fadiities~" p. 31. PAGENO="0096" 84 TRENDS IN SUPPLY AND DEMAND OP MEDICAL CARE will continue even if supply expands simply to maintain current levels of availability of medical services. 2. Price inflation in the field of medical care is made more acute by longer training and increased specialization for medical personnel and by increasing utilization of expensive hospital equipment and proce- dures requiring additional personnel. Medical care is little suscep- tible to economies of mechanization and automation. Therefore, the rising quantity and quality of medical care will tend to create additional inflationary pressures on the price of medical care. PROVIDING AN ADEQUATE SUPPLY OF MEDICAL CARE Decisions of public policy on problems of medical care must recog- nize the primarily public service nature of medical services. These decisions must encourage and insure an adequate supply of resources available for medical care. The following observations stem from the paramount public interest in maintaining an adequate, rising level of medical care for the Nation in the future. 1. To maintain current physician-population ratios, the output of U.S. medical schools must be increased. This increase will require expansion of medical school facilities, construction of new medical schools, and greater educational oppor- tunities. A recent report to the Surgeon General GO includes the following recommendations to increase the output of medical schools: (a) Pri'~ate support for low-cost loans and scholarships and Federal grants-in-aid for needy medical students. (b) Establishment of new medical schools and expansion of existing medical schools. (c) Reconsideration of unreasonably restrictive medical school admissions policies. (d) Continuing appraisal of the length and content of medical training, including evaluation of experiments to shorten the training period. (e) Increased public and private support for basic operations of medical schools. U) Federal matching grants for construction and expansion of medical schools and teaching hospitals. (g) Public and private efforts to recruit candidates for careers in medicine and related health services. 2. The supply of health facilities must be increased. Shortages of health facilities and uneven geogTaphic distribution of hospitals hinder optimum organization and maximum efficiency in the supply of medical services. The existing Hifi-Burton program of Federal aid for construction of these facilities has been very successful in stimulating States and communities to increase the supply of hospital beds, public health centers, nursing homes, and other health facilities. In view of the rising needs for health care for older people, there is particular need for greater availability of nursing homes with skilled nursing care. Existing skilled nursing home beds are meeting less than half of the need for such care. The Hill-Burton program 60 Department of Health, Education, and Welfare, "Physicians for a Growing America," Report of the Surgeon General's Consultant Group on Medical Education (Bane Committee), Public Health Service Publication b~o. 709, October 1959. PAGENO="0097" TRENDS IN SUPPLY AND DEMAND OF MEDICAL CARE 85 should be expanded, with greater support for construction of nursing homes and other health facilities to supplement hospital care. 3. Organization of supply in the field of medical care must be improved. This can be achieved by public and private assistance and encour- agement for innovations in hospital administration, in the use of alternative health facilities such as nursing homes, and in more effi- cient use of available facilities and personnel through such programs as "progressive patient care," group medical practice, and increased utilization of voluntary workers in health care. 4. Public action is needed to protect particularly vulnerable groups against the burden -of heavy medical expenses. A Federal program of social health insurance for elderly citizens is justified by their larger health needs and smaller effective demand than other segments of the population. These elderly citizens are poor risks from the point of view of private insurers, but their health needs are important from a social and humanitarian perspective. Other possibilities are direct assistance or federally aided private insurance. Crippled and handicapped children create tremendous financial burdens for low- and middle-income families. Private voluntary and philanthropic assistance to ease the burden of such disabilities must be supplemented with greater public support. Low-income families have great difficulty in financing their medical care needs adequately. Continued private and public efforts to meet the needs of this group are necessary. The inability of low-income families to finance their medical care needs adequately imposes a heavy financial burden on private philanthropic sources of funds. To enable these families to get the medical care they need, public efforts by State a-nd local governments must be supplemented with greater Federal support. To summarize-Federal support and leadership is essential for cooperative action by States, communities, and private groups to insure adequacy in the supply of medical care. 0 PAGENO="0098" PAGENO="0099" PAGENO="0100"