PAGENO="0001"
90th Congress T JOINT COMMITTEE PRINT
1st Session
C
OLD AGE INCOME ASSURANCE
A COMPENDIUM OF PAPERS ON PROBLEMS AND POLICY ISSUES
IN THE PUBLIC AND PRIVATE PENSION SYSTEM
SUBMITTED TO THE
SUBCOMMITTEE ON FISCAL POLICY
OF THE
JOINT ECONOMIC COMMITTEE
CONGRESS OF THE UNITED STATES
Part V: Financial Aspects of Pension Plans
7~NM~NT DEPOSITORY
RTY OF RtJTGERS, THE SThTE &M~H~ERSITy
OLLEGE OF SOUTH JERSEY LIBRAR~ ~
~AMDE~, ~L J. 08102
JAN 2 319 S8CEMBER 1967
Printed for the use of the Joint Economic Committee
U.S. GOVERNMENT PRINTING OFFICE
WASHINGTON 1967
For sale by the Superintendent of Documents, U.S. Government Printing Office
Washington, D.C. 20402- Price 65 cents
PAGENO="0002"
JOINT ECONOMIC COMMITTEE;
[Created pursuant to sec. 5(a) of Public Law 304, 79th (`orig.
WILLIAM PROXMIRE, Wisconsin, Chairman
WRIGHT PATMAN, Texas, Vice Chairman
HOUSE OF REPRESENTATIVES
RICHARD BOLLING, Missouri
HALE BOGGS, Louisiana
HENRY S. REUSS, Wisconsin
MARTHA W. GRIFFITHS, Michigan
WILLIAM S. MOORHEAD, Pennsylvania
THOMAS B. CURTIS, Missouri
WILLIAM B. WIDNALL, New Jersey
DONALD RUMSFELD, Illinois
W. B. BROCK 3D, Tennessee
JOHN R. STARK, Executive Director
JAMES W. KNOWLES, Director of Research
EooNoMIsTs
WILLIAM H. MooRE JOHN B. HENDERSON GEO~GE B. ThEN
DONALD A. WEBSTER (Minority)
SUBCOMMITTEE ON FIsCAL Porao~
MARTHA W. GRIFFITHS, Michigan, Chairman
HOUSE OF REPRESENTATIVES SENATE
HALE BOGGS, Louisiana WILLIAM PROXMIRE, Wisconsin
WILLIAM S. MOORHEAD, Pennsylvania HERMAN E. TALMADGE, Georgia
WILLIAM B. WIDNALL, New Jersey STUART SYMU~GTON, Missouri~ ~ ~c
DONALD RUMSFELD, Illinois ~ JAVITS,~Ne*.,York ft.~" IL.4 k
~
NELSON D. McCLUNG, Econon~ ~t~i?~3~ hTiJ~ 1~') 3J31~
II ~
SENATE
JOHN SPARKMAN, Alabama
J. W. FULBRIGHT, Arkansas
HERMAN E. TALMADGE, Georgia
STUART SYMINGTON, Missouri
ABRAHAM RIBICOFF, Connecficut
JACOB K. JAVITS, New York
JACK MILLER, Iowa
LEN B. JORDAN, Idaho
CHARLES H. PERCY, Illinois
PAGENO="0003"
LETTER OF TRANSMITTAL
DECEMBER 20, 1967.
To the Members of the Joint Economic Committee:
Transmitted herewith for the use of the members of the Joint
Economic Committee and other Members of Congress is part V, "Fi-
nancial Aspects of Pension Plans," of the compendium of papers en-
titled "Old Age Income Assurance," prepared for the Subcommittee
on Fiscal Policy.
The views expressed in this document do not necessarily represent
the views of members of the committee or the committee staff, but are
statements of issues and alternatives intended to provide a focus for
hearings and debate.
WILLIAM PROXMIRE,
Chai~riman, Joint Economic Committee.
DECEMBER 19, 1967.
Hon. WILLIAM PRoxMnu~,
Chairman, Joint Economic Convn-tittee,
Congress of the (JnitedaS'tates, WashingtonD.C.
DEAR Mr~. CHAIRMAN: Transmitted herewith is part V, "Financial
Aspects of Pension Plans," of the compendium of papers on prob-
lems and policy issues in the public and private pension system, en-
titled "Old Age Income Assurance."
Part V treats the effects on saving and investment of pension pro-
grams, a~id contains nine papers contributed by invited specialists.
The subcommittee is indebted to these authors for their excellent con-
tributions which we believe will add much to a general awareness of
the issues in retirement income policy, particularly as these relate to
the impact of programs on capital markets. The time and learning
devoted to the preparation of these papers should do much to stimu-
late interest and to~ assist in policy decisions concerning future pro-
grams for old-age income assurance.
Dr. Nelson McClung, consultant to the subcommittee is responsible
for the planning and preparation of the compendium, with the edi-
torial assistance of Aiine McAfee, and the advice and suggestions of
other members of the committee's professional staff.
As the executive director's letter indicates, the compendium should
not be viewed as an expression of views or conclusions of the commit-
tee staff, nor should it be viewed as an expression of views of the sub-
committee or individual members.
MARTHA IV. (iRIFFIT[TS,
Uhai'rinan, Subcommittee on Fiscal Policy.
m
PAGENO="0004"
DECEMBER 18, 1967.
Hon. MARTHA W. GRIFFrnis,
Chairman, Subcommittee on Fiscal Policy, Joint Economic Com-
mittee, U.S. Congress, TVas/ungton, D.C.
DEAR MADAM CI-IAIn~rAx: Transmitted herewith is part V, "Finan-
cial Aspects of Pension Plans," of the compendium of papers entitled
`COld Age Income Assurance." This study was prepared at your re-
quest in order to bring together current thinking on the questions
of retirement income programs and thereby cont.ribute to policy de-
cisions by focusing attention on the more promising solutions to the
income problems of older people.
The coiupendium, which is being issued in five parts, confirms the
fact that programs to aid older people have grown in number, size,
and complexity, and that the coordination of these programs and their
combined impact on the income of older people have received very
little attention. Clearly, public policy issues exist with respect to
coordinating these programs, appraising their effects on the economy,
and improving their equity.
Part V contains contributions by the authors listed below. The
committee is indebted to these contributors who have given generously
of their time and expertise to provide the latest available informa-
tion and competent analytical perspective to this important subject.
Prof. Henry Aaron Dr. Arthur S. Fefferman
Prof. H. Robert Bartell, Jr. Prof. Dan McGill
Prof. John 0. Blackburn Dr. Roger F. Murray
Prof. David Cass Dr. James J. O'Leary
Prof. Peter 0. Dietz Elizabeth Simpson
Mr. C. Wadsworth Farnum Prof. M. E. Yaari
The major work in planning and compiling this compendium was
undertaken by Dr. Nelson McClung, consultant to the subcommittee,
with the advice and suggestions of other members of the staff. He
was assisted in the editorial work by Anne McAfee. Nothing herein
should be interpreted as representing either the opinions of the staff
or the members of the committee on any of the matters discussed.
JOHN R. STARK,
Executive Director, Joint Economic Committee.
Iv
PAGENO="0005"
OLD AGE INCOME ASSURANCE
Part V: Financial Aspects of Pension Plans
CONTENTS
Page
Transmittal letters III
THE MACROECONOMICS OF PENSION FUNDS, by Prof. John 0. Blackburn,
Duke University 1
THE SOCIAL INSURANCE PARADOX, by Prof. Henry Aaron, University of
Maryland 15
A RE-EXAMINATION OF THE PURE CONSUMPTION LOANS MODEL, by
Prof. David Cass and Prof. M. E. Yaari, Yale University 19
ECONOMIC ASPECTS OF PENSIoNS: A SUMMARY REPORT, by Dr. Roger F.
Murray, National Bureau of Economic Research 36
PENSION FUNDS OF MULTIEMPLOYER INDUSTRIAL GROUPS, TJNIONS, AND
NON-PROFIT ORGANIZATIONS, by Prof. H. Robert Bartell, Jr., North-
western University, and Elizabeth Simpson, National Bureau of Econ-
nomic Research 115
COMMENTS OF AMERICAN LIFE CONVENTION, AND LIFE INSURANCE Asso-
CIATION OF AMERICA, by Dr. Arthur S. Fefferman and Dr. James J.
O'Leary 171
GUARANTY FUND FOR PRIVATE PENSION OBLIGATIONS, by Prof. Dan
Mc Gill, Wharton School of Finance and Commerce, University of Penn-
sylvania. 199
AN ANALYSIS OF PROPOSALS FOR IMPROVING THE FUNDING AND FINANCIAL
MANAGEMENT OF PRIVATE PENSION FUNDS, by Prof. Peter 0. Dietz,
Northwestern University, and Prof. H. Robert Bartell, Jr., Northwestern
University 248
CORPORATE FIDUCIARIES OF EMPLOYEE BENEFIT FUNDS-AN OUTLINE OF
RESPONSIBILITIES AND SUPERVISION, by C. Wadsworth Farnum, senior
vice president, Bankers Trust Company 257
AUTHORS LISTED ALPHABETICALLY
AARON, Prof. Henry 15
BARTELL, Prof. H. Robert, Jr 115, 248
BLACKBURN, Prof. John 0 1
CASS, Prof. David 19
DIETZ, Prof. Peter 0 248
FARNUM, C. Wadsworth 257
FEFFERMAN, Dr. Arthur S 171
MCGILL, Prof. Dan 199
MURRAY, Dr. Roger F 36
O'LEARY, Dr. James J 171
SIMPSON, Elizabeth 115
YAARI, Prof. M. E 19
V
PAGENO="0006"
PAGENO="0007"
THE MACROECONOMICS OF PENSION FUNDS
BY JOHN 0. BLAOKBtIRN*
The provision of income for the elderly has become an increasingly
institutionalized undertaking in modern, urban societies. This develop-
ment suggests a number of issues, many of which are treated in this
Symposium. Like other economic activities, income maintenance de-
vices, if sufficiently large and widespread, have an impact on aggregate
saving, investment, output, and growth. It is the purpose of this paper
to examine pension plans in the light of their aggregate impact on the
economy. Part I discusses some concepts and definitions with respect to
pension assets, liabilities, and the structure of claims on national
wealth. Part II discusses such variables as length of working life, life
expectancy at retirement, and retirement income relative to working in-
come as they relate to saving, investment, income, and wealth. The
response of the economy to changes in saving patterns induced by pen-
sions is also considered under various assumptions. It is there shown
that pension saving, together with Federal deficits to maintain full em-
ployment, produce a public-private wealth structure which is analyti-
cally similar to an underfunded public retirement system.
Part III examines the growth of pension funds, 1946-1965 and notes
the growing share of personal saving which is thus institutionalized.
Part IV discusses some recent projections of pension funds to 1981 and
conclude that they are prcthably too low. If this is the case, then the
impact of pension funds on the economy is likely to be even greater
than previously indicated.
The paper has few conclusions as such, but it does throw additional
light on a number of important questions. Perhaps the major issue has
to do with the likely course of full employment saving in relation to
private investment demand. There may well develop a situation in
which private saving in pension funds must be partly offset by dis-
saving on the part of the Federal Government if relatively high un-
employment is to be avoided. The use of tax policy to encourage a kind
of private saving which requires offsetting public deficits would appear
to be curious indeed. There would seem to be an urgent need to explore
reinsurance and other devices for reducing the degree of private fund-
ing, if such an outcome of our present arrangement is indicated. Be-
yond that, liberal tax inducements toward pension saving might be
reevaluated, if there is in prospect a saving-investment relation which
is more often contractionary than not.
I. So~ CONCEPTUAL AND DEFINITIONAL CONSIDERATIONS
An economy in which the extended family is the dominant institu-
tion provides goods and services to the elderly via intergenerational
transfers within the family. Saving and investment may and do take
*Duke University.
1
PAGENO="0008"
2 OLD AGE INCOME ASSURANCE-PART V
place, but they are not necessarily related to the process of providing
for the aged. The physical separation of generations, frequently ac-
companied by a reduction in the number of children per family,
seems to be associated with the rise of more formalized financial pro-
visions for retirement, both as to public, and private arrangements.1
A striking feature of recent history in the United States is the
rapid growth and spread of pension arrangements which involve a
degree of compulsion-both with respect to public and private insti-
tutional arrangements. There are some problems of concept and defi-
nition in these arrangements that have to do with public and private
wealth, and the st.ructure of claims by which wealth is owned.
The desire to allocate some portion of current real output under
one's control to future consumption is a desire to save. In economies
using productive capital goods, resources released by saving for future
consumption (or for any other purpose) may be accumulated in the
form of capital goods. Since saving and investment are acts typically
performed by different economic transactors, the act of saving usually
results in the accumulation of direct or indirect financial claims on
investing sectors. In fact, the development of financial claims which
permit the intersectoral flow of saved resources may be regarded as a
major factor in the productivity of modern, high-income economies.
Until the advent of the social security program in 1935, financial
provision for retirement (if any) in the United States was mostly
an individual matter. Some private pension plans had been in exist-
ence for years. and some of these had features of compulsion. For
the vast majority of Americans, however, saving for retirement took
the same forms as for any ot.he.r purposes. Aside from equities in
the few existing pension plans, the major financial instrument tailored
especially for retirement was t.he annuity offered by insurance com-
panies. This instrument has never been of great importance in the
American financial structure.
By way of constrast., pension claims by households against public
and private institutions now represent a major financial instrument.
a form of financial saving, or as a. means of channeling house-
hold savin~s into the hands of investing sectors, pension equities al-
ready funded amount to some ~175 billion. If unfunded pension claims
were included, the sum would be larger by several hundred billion
dollars. Indeed, by some ways of reckoning the total pension obliga.-
tions, they may exceed in the aggregate all other types of financial
cla~ms except common stocks.
This observation raises the question as to the appropriate definition
of pension assets and liabilities. From the standpoint, of the. house-
hold, a claim to future payments constitutes an asset which might be
measured as the discounted present. value of future payments al-
ready earned. The corresponding liability, from the standpoint of the
paying sectors, is the discounted present value of future payments
likely to be made on the basis of work already performed by each em-
ployee. Viewed in this light, the a.ggregate "asset" of households and
"liability" of pension-paying sectors is enormous-much bigger than
reportecT reserves of the various public and private retirement funds.
1 For a discussion of these matters, see Juanita 1sf. Kreps. "The Economics of Inter-
cenerational Relationships". in Ethel Shanas and Gordon Strnih. Social Structure and the
Family: Generational Relations. Prentice Hall (Englewooci Cliffs, 1065).
PAGENO="0009"
OLD AGE INCOME ASSURANCE-PART V 3
For purposes of speculation, let us suppose that the unfunded lia-
bility of the OASDIEJII funds with respect to retirement benefits is
$400 billion.2 Further suppose that the various public and private
funds which are run on a more or less actuarially funded basis are
funded to the extent of 50 percent of their liabilities. These funds
amounted to approximately $150 billion at the end of 1966. This im-
plies a roughly calculated $720 billion as the value of household
"assets" with respect to retirement benefits already "earned", and
further implies a corresponding liability for the paying sectors. Pen-
sion claims would then indeed represent the second largest single
financial instrument of the U.S. financial structure.4 We could further
note, without taking the whole matter too seriously, that accrued
pension obligations of the Federal Government already exceed the
reported Federal debt, the upper limit of which is the subject of
annual Congressional shadowboxing.
Some observers would, no doubt, object to showing an additional
$150 billion of non-OASD}il pension obligations as anyone's asset or
anyone's liability. Indeed, trustee plans are not obligated to make any
payments beyond those which would exhaust their assets. How, it may
be asked, can they be said to have any further liability? There is, how-
ever, a growing disposition among accountants to measure liabilities as
the present value of probability-weighted future payments, just as
assets are treated as the present value of probability-weighted future
receipts.~ We can hardly make any assumption about business enter-
prise in the aggregate other than its continued existence. We may fur-
ther assume, in the aggregate, that pension promises will be largely
met. We may therefore argue for the broader definition of pension
liabilities since, legal obligations or no, the probability is high that
pensions will be paid. Indeed, the whole tenor of much present dis-
cussion about non-OASDHI pensions is to require stricter vesting and
funding practices. These measures, if enacted, would move the legal
liabilities closer to the likely-future-payment concept of pension
liabilities.6
WTe should also consider the relative values placed on future pen-
sion payments by households and paying sectors. It is likely that house-
holds place a smaller present value on future pension payments than do
the sectors which will pay them. This is equivalent to saying that house-
holds discount them at a higher rate, or hold lower subjectives esti-
mates of the probability that they will be paid. The evidence in this
2 Unnublished memorandum of the Chief Actuary, Social Security Administration. Mar. 9,
1967. The estimate is $350 billion, .or $417 billion if the amendments before the Congress
are enacted. Of the latter figure, $10 billion would relate to the Railroad Retirement Funds,
which are omitted from this study.
Composed as follows: Billions
OASI funds $20
Unfunded liability of same 400
Private. State. and local government, and Federni Civil Service plans 150
Unfunded liability of same 150
Total 720
The financial instrument now quantitively the most important is common stock: the
estimated aggregate value at December 31, 1965 was $778 billion. Federal Reserve Bulletin,
vol. 52. p. 1536. (Ortoher. 1966).
This view is explicit In the initial opinion paragraph on basic accounting methods In the
American Institute of Certified Public Accountants' Exposure Draft of Tentative Opinion:
Account2n.q for the Cost of Pension Plans. See The Journal of Accountancy, September,
1961. a. 63.
~ Recommendations to the effect are found, inter aBa, ln~ a report by the President's Com-
mlttse on Corporate Pension Funds and Other Private Retirement nnd Welfare Programs,
Public Policy and Private Pension Programs, Washington, 1965, especIally pp. x-xvl.
PAGENO="0010"
4 OLD AGE INCOME ASSURANCE-PART V
connection is somewhat spotty, but is consistent with the above hy-
pothesis. Workers under vested plans with the right to withdraw accu-
mulated sums upon termination ahnost always do so when they change
)obs. Further, claims to future pension payments are apparently not
very substitutable for other forms of saving. The first observation
suggests that households have a higher rate of time preference than the
rate which relates present contributions to future benefits, or that
they attach some probability less than one to actually collecting the
benefits, or both. The second observation suggests that households
with pension coverage who save as much as or more than noncovered
households (comparable in other respects) do not really value pen-
sion claims as highly as other financial claims; otherwise, they would
be more highly substitutable.7
Before leaving these somewhat fanciful concepts of pension claims,
we should note that measuring them on the basis discussed would pro-
foundly alter the apparent financial structure of the U.S. economy.
Private household wealth would be greatly increased, while govern-
ment sector and business sector liabilities would also increase con-
siderably. The Federal Government would show a large negative net
worth, and business firms would show net assets considerably smaller
than that implied by their books, or by the market value of their out-
standing shares.
Since there are some conceptual and computational difficulties in
measuring pension claims on a present value basis, most compilers of
aggregate data simply measure funded claims and ignore the rest..8
This practice will be followed in the remainder of this paper; when we
speak of pension funds and household pension equities we shall be
referring only to claims already funded. This concept alone yields
large and rapidly growing aggregates, as the following pages will
make clear.
II. So~n~ MACROECONOMIC ANALYSES
The role of retirement parameters in determining important macro-
economic variables such as saving, investment, income, consumption,
and the capital-output ratio, can be illustrated by means of some simple
models of the economy. No economies are actually that simple, of
course, but the influences described do carry over into real economies.
First, consider a. static model in a position of longrun equilibrium.
Population, output, technology, and the stock of capital are constant
through time. Further assume that people save only for retirement,
through pension funds, with given and constant ages of beginning and
ending work, and a constant life expectancy at retirement. Then all
capital, indeed all nonhuman wealth, is owned by the pension funds. It
can be shown, in such an economy, that output, the capital-output
ratio, and the share in consumption by retired persons depend on the
length of working life and life expectancy at retirement as well as on
The authoritative study in this regard is that of Phillip Cagan, The Effect of Pension
Plans on. Aggregate 5aving~ Occasional Paper 95, National Bureau of Economic Research,
New York, lOGS. Cagan's findings that pension savers apparently do not save less in other
forms than comparable nonpension households is consistent with an alternative hypothesis
that pension saving has a kind of "demonstration effect" which encourages saving of all
kinds. His findings, however, are not inconsistent with the hypothesis stated in the text.
S Flow-of-funds data prepared by the Federal Reserve Board, for example, treat only
funded pension claims as household assets. The national Income and product accounts con-
cept is even narrower; only, contributions to and investment income of private pension
funds are treated, in effect, as personal saving.
PAGENO="0011"
OLD AGE INCOME ASSURANCE-pART v 5
the usual variables of capital and labor productivity and the rate of
interest.~ Further, the static stock supply function of capital, in which
the equilibrium captial-output ratio is related to the rate of interest,
is negatively sloped; the higher the rate of interest (productivity of
capital, for any given capital-output ratio) the lower is the capital-
output ratio. This result is easily explained: the higher the rate of in-
terest, the smaller is the fund which must be accumulated for each
worker (relative to his wage) to pay his retirement benefits.
Essentially the same results carry over into a model in which popula-
tion, the labor force, and factor productivity (via the level of tech-
riology) are all growing smoothly at constant rates. The capital-output
ratio is rising through time, as is output and income per capita. Tinder
certain assumptions about the aggregat.e production function, there
exists a longrun equilibrium growth path for which the rate of interest
(capital marginal productivity) is constant, as technological advance
offsets the tendency toward diminishing returns to capital (implied by
the rising capital-labor ratio).
If, in such a moving-equilibrium model, all saving is for retirement,
and retirement income bears a constant relation to working income, it
can again be shown that the equilibrium supply curve of capital (in a
stock sense) is negatively related to the rate of interest. As in the static
model, the longer is life expectancy at retirement, the larger will be the
capital-output ratio and the share of national income consumed by
retired persons.'° With a given life expectancy, as the retirement age
falls and the retirement period lengthens, the equilibrium capital-out-
put ratio rises, the rate of interest falls, and consumption per capita
falls."
Let us turn our attention now from equilibrium-type models to our
own economy. Under neoclassical assumptions of smooth full-employ-
ment adjustments in interest rate, prices, and capital-output ratios,
an increase in the share of income and output saved would pose no
problems. If pension plans increase the share of national income which
is saved, then we should expect the economy to respond with lower
interest rates, a higher capital-output ratio, and more output per cap-
ita during each time period. It does not follow that consumption per
For a more detailed discussion of such models, ~ee my "Pensions, the National Income,
and the National Wealth," in Juanita Kreps, ed., Employment Income and Retirement Prob-
lems of the Aged, Duke University Press, Durham, 1963.
10 For a demonstration of the growing population but constant productivity case, see
Blackburn, op. cit. The more complicated cases in which both population and output per
worker are rising but along a constant-Interest-rate long-run equilibrium growth path are
simply variants of neoclassical equilibrium growth models of the type treated by Solow,
Brems, Phelps, and others. See, for example, R. M. Solow, "A Contribution to the Theory
of Economic Growth", Quarterly Journal of Economics, November, 1956, vol. 70, pp. 537-
562; Hans Brems, "Growth, Distribution, Productivities, and Thrift in Cobb-Douglas
Models," Southern Economic Journal, January 1963, vol. 29, pp. 181-88. (See also my
comment and his reply. Southern Economic Journal, October 1963 vol. 30, pp. 175-179)
and Edmund S. Phelps, "The Golden Rule of Accumulation: A ~Pable for Growthmen,'~
American Economic Review, September 1961, vol. 51, pp. 38-43. The proof that changes
in the retirement age, life expectancy, or the fraction of working income desired as retire.
ment Income have the same effects on the interest rate and the capital-output ratio as in
the static case Is too lengthy to reproduce here. It is available from the author.
~` Ando and Modigliani davelop a model of aggregate saving based primarily on lifetime
consumption (retirement saving) considerations. Their basic assumption Is that all saving
is for retirement, though they indicate ways in which this assumption can be relaxed with-
out destroying the results. It is interesting that estimates of the stock of wealth implied by
a wide range of empirically plausible values for working life, life expectancy, retirement
income as a fraction of working income, the rate of interes~t, and age distributions of the
population give values of the capital-output ratio far lower than the existing U.S. ratio.
It would appear that nonretirement saving must be considered in the Ando-Modigliani
model. See Albert Ando and Franco Modigliani, "The `Life Cycle' Hypothesis of Saving;
Aggregate Implications and Tests," American Economic Review, March 1963, vol. 53, pp.
55-84.
PAGENO="0012"
6 OLD AGE INCOME ASSURANCE-PART V
capita would be raised; the increment to output (assuming positive re-
turns to capital) might fall short of the increment to saving (assuming
diminishing returns to capital, and a sufficiently high initial capital-
output ratio) 12
Under post-Keynesian assumptions, additional savings induced by
pension plans again need not threaten either price stability or full
employment. Under conditions of chronic excess demand, additional
private savings via pension plans would merely reduce the size of
Federal surpluses needed to restrain demand, or might permit greater
monetary ease with generally lower interest rates. Under conditions
of chronically inadequate private demand, the existence of additional
saving through pensions implies larger Federal deficits, and perhaps
lower private saving rates.
The latter possibility may be taken as a case at least worthy of con-
sideration. The years following 1957 were either years of unemployed
resources, or Federal budget deficits, or both. The only post-1957 years
in which private investment absorbed private saving forthcoming
at more or less full employment were 1965 and 1966. The Federal
budget., on a national income account.s basis, had a slight surplus in
those years. Yet that volume of private investment, abetted until late
1965 by monetary ease and longer still by the investment credit and
liberalized depreciation rules, may well turn out to be unsustainable.
Unsustainable, that is, in the sense that productive capacity grew in
1965 and 1966 more rapidly than could full-employment output, con-
strained by the growth rate in the labor force and in labor productiv-
ity. Data now available do not permit any conclusive statements; the
question would a.ppear to be sufficiently open to warrant an examina-
tion of the "excessive saving" case.
If pension saving does raise the implied full-employment growth
rate in output beyond that consistent with the growth in the labor
force and labor productivity, and if the capital-output ratio, interest
rate, profit. rate, and technological developments do not easily reconcile
these divergent rates, then a balanced-budget economy would tend
toward stagnation and chronic unemployment.'3 The use of monetary
and especially fiscal policy to offset pension saving would result in a
growing public debt.
Thus private pension saving would take place at the expense of
public dissaving. Apparent private wealth would increase, if we make
the common assumption that households do not take into account any
share of their future tax liabilities with respect to the public debt.
~ This is a major point of Phelps' paper (op. cit.). He finds that consumption is maxi-
mized where the growth rate equals the Interest rate. The noint is pursued~ in my "The
Social Insur'ince Paradox: A Comment." Canadiaa Journal of Economics and Political
Science, forthcoming. It is there shown that in the context of neoclassical growth models.
pension claims are overfunded (reduce aggregate consumption) if the interest rate lies
below the rate of crowth in output.
~ My version of the chronic oversaving case is, I hope, a bit more sophisticated than
versions discredited by events In the early postwar years. I am thinking explicitly In terms
of long-run growth considerations. Record nrlvate Investment outlays in 1965 (15.6 per-
cent of GNP) and 1966 (15.8 percent of GNP) are estimated to have raised productive
capacity in manufacturing by 5 and 7 percent, respectively. (Council of Economic Advisers.
Annual Report, 1967. p. 221 and p. 253. Capacity in other sectors. of course, does not neces-
sarily follow manufacturing capacity.) No one has seriously proposed that total output in
the United States might grow for long periods at such rates. Given estimates of the labor
force In future years, such growth implies annual increases In labor productivity rarely
attained and never sustained in the United States. My view also implies some pessimism
as to smooth (even in the long run) adjustments in the economy toward higher capital-
output ratios and lower interest rates. Movement toward the latter, of course, is further
constrained at the moment by balance-of-payments considerations.
PAGENO="0013"
OLD AGE INCOME ASSURANCE-PART V 7
Private pension claims are then indirectly supported by a public lia-
bility-wholly, if the increase in public debt equals the accumulations
of private funds; partly, if "excess saving" is only part of pension
saving. The important result of analyzing this case is the following:
as to saving, investment, income, and private wealth (including the
present value of future pension payments) private pension funds plus
public deficits are analogous to unfunded public pensions. When
OASDHI is underfunded, private wealth (including the expected
value of future pension benefits) exceeds national wealth. The Gov-
ernment has a liability equal to the underfunding of OASDHI.
Future tax revenues will, in effect, pay the pensions. In the private
pension fund, public-debt case, private wealth again exceeds national
wealth, but the public liability to be tax financed (at least as to inter-
est) appears not as OASDHI underfunding but as explicit debt.
Though the net effects are similar for some variables, they are quite
different with respect to other variables. Since Government securities
are not very important in private pension portfolios, and common
stocks are, the ceteris paribus effect of the pension saving-chronic
excess saving-Federal deficit case is to raise stock prices relative to
bond prices, to lower stock yields relative to bond yields, and to alter
the cost of equity capital relative to debt capital.
If, on the other hand, the correct characterization of the postwar
period in the United States is one of chronic undersaving and excess-
demand inflation, the effects of pension saving are quite different.
Price inflation has been less severe than it otherwise would have been.
Federal deficits would have been larger (occasional surpluses smaller)
than would have been required to maintain the same degree of demand
inflation in the absence of pension saving.
There is yet another aspect of pension saving which needs our con-
sideration; namely, the influence of tax concessions on aggregate sav-
ing. Under present arrangements, some 40 percent of personal saving,
and subsequent investment earnings thereon, escape current taxation.
Various assumptions as to deferment periods, discount rates, tax rates
during retirement, and the like, produce widely varying estimates of
the effective tax benefit. Nevertheless, the implied rate of taxation on
pension saving is materially lower than on income in general. We are
thus taxing all personal saving on the average, and pension saving in
particular, at a lower rate than we tax consumption. In the oversaving
case discussed above, we are subsidizing saving through tax policy in
such a way as to require even larger offsetting public deficits than
would be required without the subsidy.
III. PENSION FUNDS, 1946-65
The explosive postwar growth of pension funds in the United States
is a widely documented phenomenon, so that it need not be discussed
in detail here. We shall briefly trace out the course of accumulated
pension funds since 1946, and compare annual increments (saving)
to the total of household saving in the same period. Figure 1 traces, on
logarithmic scale, the growth of pension funds (other than OASDHI
and the Railroad Retirement Funds) ,14 from approximately $11 bil-
14 Corporate pension funds, insured funds, State and local government funds, and Federal
Civil Service Retirement Fund. Railroad Retirement Funds are omitted since they are
relatively small and hardly growing.
PAGENO="0014"
8 OLD AGE INCOME ASSURANCE-PART \T
lion in 1946 to $135 billion (book value) at the end of 1965. The gradu-
ally decreasing slope of the total funds graph in Figure 1 indicates a
high but declining growth rate. From around 16 percent in the early
years, the annual growth rate has fallen to~ about 10.5 percent in the
last few years.
That the growth rate will fall still further (but not necessarily
indefinitely) is suggested by the growth of annual increments to pen-
sion funds, also shown in Figure 1. They fluctuate somewhat, but have
grown since 1952 at an average rate of about 8.7 percent, with no ap-
parent declining trend. By way of illustration (but not by way of
prophecy) if annual increments continued to grow at 8.7 percent, the
growth rate in total funds would fall asymptotically toward 8.7 per-
cent also. These matters are discussed in more detail in Section IV
below.
The relationship of pension saving to total household saving is
shown in Table 1, as is the Personal Disposable Income-Personal Sav-
ing relationship. Both of these latter aggregates are adjusted slightly
from the concepts used by the National Income Division, Office of
Business Economics. Employer and employee contributions, and in-
vestment earnings of certain Government funds, are included here in
both personal disposal income and personal saving, whereas they are
excluded in official U.S. data. The adjustment is made here to treat
State and local government funds, and the Federal Civil Service Re-
tirement Fund, m a manner consistent with private funds. My figures
are slightly distorted by the inclusion of realized capital gains in pen-
sion fund increments. The amounts, however, are quite small.
TABLE 1.-PERSONABLE DISPOSABLE INCOME, PERSONAL SAVING, AND PENSION SAVING, 1946-65
[In billions of current dollarsi
Year
Personal
disposable
income
Person
al saving
Pension
saving
Amount
Percent of
personal
disposable
Amount
Percent of
personal
saving
income
1946
1947
1948
1949
1950
1951
1952
1953
1954
1955
1956
1957
1958
1959
1960
1961
1962
1963
1964
1965
160.7
170. 6
190.0
189. 6
208. 2
227.8
239. 9
254. 1
258. 8
277. 1
295. 4
310.5
321. 5
340. 1
353. 1
367.9
389. 0
407. 7
440.1
469. 9
15. 9
8. 1
14.3
10. 4
14. 4
18.5
19. 8
19. 8
17. 8
17. 6
22. 8
22.8
25. 0
21. 9
20. 1
24.7
25. 3
24. 3
30.6
29.7
9.9
4. 7
7.5
5. 5
6. 9
8.1
8. 3
7. 8
6. 9
6. 4
7. 7
7.3
7. 8
6. 4
5. 7
6.7
6. 5
6. 0
7.0
6. 3
1.6
1. 8
2.1
2. 9
3. 0
3.4
4. 3
4. 5
4. 8
5. 4
6. 0
6.5
7. 3
8. 2
8. 1
11.4
9. 4
10. 3
11.6
12. 8
10. 1
22. 2
14.7
27. 9
20. 8
18.4
21. 7
22. 7
27. 0
30. 7
26. 3
28.5
29. 2
37. 4
40. 3
46.2
37. 2
42. 4
37.9
43. 1
Source: Author's estimates, prepared from data by Department of Commerce Securities and Exchange Commission; dato
adjusted for saving through Government funds.
For private pension funds, contributions and fund earnings appear
in national income and, not being subtracted, remain in personal dis-
PAGENO="0015"
8A
OLD AGE INCOME ASSURANCE-PART V
Figure 1
U.S. Pension Funds, Book Value, and Annual Increments
$100 billion
90
80
70
60
50
40
30
$10 billion
9
8
7
6
5
$1 billion
Year 1946 `48 `50 `52 `54 `56 `58 `60 `62 `64 `66
PAGENO="0016"
PAGENO="0017"
OLD AGE INCOME ASSURANCE-PART V 9
posable income. They remain as an element of personal saving, since
that is defined as personal disposable income less personal outlays
(mostly for consumption). The two kinds of government funds in
question, namely, funds of State and local governments, and the Fed-
eral Civil Service Retirement Fund, are treated differently in our na-
tional accounts. Contributions to them are deducted as social insurance
contributions in moving from national income to personal disposable
income. Their interest earnings have already been excluded from na-
tional income and product in the calculation of net government interest
paid (all of which is excluded `from national income). These funds,
like private funds, and unlike the OASDHI fund, are run on a more
or less actuarially funded basis; increments are, therefore, treated here
as part of personal saving just as increments to private funds are so
treated in the national accounts.
Table 1 suggests a rising trend in the share of personal saving taking
place through pension plans. T'his view is borne out by regressing that
share on time as a variable; `time is significant at the .001 level.15 This
result is not surprising in view of the fact that pension saving is grow-
ing more rapidly than output. and income, and the roughly `constant
share of personal disposable income which is saved. The latter variable,
personal saving as a percent of personal disposable income, has fluc-
tuated considerably, but shows no significant time `trend.'~
These findings are somewhat puzzling in the light of evidence that
pension saving is likely to be added to other saving, thus increasing
the share of income saved. The authoritative study in this matter is
that of Phillip Caga.n.'~ His findings from a sample of households
suggests that the saving ratio should rise, cetern~s paribus, as pension
saving grows in volume and spreads throughout the economy. He, of
course, discusses the apparent inconsistency in household sample and
aggregate time series data, and concludes that other forces making for
smaller saving ratios must have offset the influence of pensions during
the postwar period.
In any event, since the share of disposable income saved by house-
holds has not systematically risen, and since pension saving has grown
more rapidly than output or income, the share of pension saving in
personal saving has assumed major significance, exceeding 40 percent
in 3 of the last 5 years.
If pension saving performs in the manner which is projected in sec-
tion IV below, it will continue to rise as a share of personal saving,
unless personal saving begins to rise as a share of disposable income.
This latter possibility `should not be ruled out; whatever forces may
have offset the tendency of pension saving to increase total saving in
the recent past may not continue to operate. We should be prepared to
face the policy implications of a small upward shift in the share of
disposable personal income saved.
~ For the period 1946-65, regressing the share of pension saving in total personal
saving on time yields, w=13.06+1.54t where ~v is the share of pension saving in personal
saving. The standard error of the t coefficient Is .167.
1~ For the perio'd 1946-65, regres'sing t'he share of personal disposable income saved
(11) on time (t) yields
y=7.55-.055t
The standard error of the t coefficient is .044, and is not significant at any reasonable level.
`~ Cagan, op. cit.
83-200-67-pt. 5-2
PAGENO="0018"
10 OLD AGE INCOME ASSURANCE-PART V
IV PRo,j~cTIoNs OF PENSION FUNDS
The explosive growth of pension funds, particularly that of cor-
porate noninsured funds, has long since attracted much attention,
and has given rise to various projections of growth to 1975 or 1980.
Perhaps the most widely quoted are those of Daniel Holland, pub-
lished by the National Bureau of Economic Research.'8 These esti-
mates suggest a rise of private pension funds from about $52 billion
in 1960 to around $200 billion in 1981, or from $72 billion in 1960 to
$325 billion in 1981, if State and local government funds, treated
separately by Holland, are added in.'°
In the light of experience since the estimates were prepared, and
on theoretical grounds as well, these estimates may well be too low,
perhaps considerably too low.20 in my view, which is not nearly so
carefully supported as is Holland's, it appears that $450-$470 billion
is a more likely estimate; the derivation of this estimate is given below.
Holland's group of estimates begins with four projected series of
covered employment, and further projections of numbers of retired
beneficiaries. Contributions per covered employee, and mean benefit
per ret.ired beneficiary, are also estimated. These, with assumed fund
earnings, are used to build up a. year-by-year set of estimates for in-
flows, outflows, and fund balances from 1962 to 1981. Numerous
combinations of assumptions are used, but the sets which Holland
regards as most likely produce the results cited above.
Most of the sets produce a pattern of rising fund balances, though
with absolute amounts of annual increments declining in the later
years. This implies, of course, growth at declining rates; in fact, some
of the sets produce fund decurnulation. The most likely results are
heavily influenced by the use of a constant annual contribution per
employee throughout the 20-year period, while benefits per retired
beneficiary are permitted to rise. The assumption of a constant con-
tribution per worker is based on the 1950-61 experience, supported by
an analysis of likely developments in coverage, fund earnings, and
benefit levels.2'
Holland also prepares several alternative estimates, in one of which
contributions and benefits rise at 4 percent per year (the average rise
in money wages, 1950-61). This, of course, leads to higher projections
(about 15-20 percent higher in 1981) ~22 This is the highest of all his
estimates, but appears still to be biased downward in that fund earn-
ings are estimated to rise only 4 percent per year whereas fund totals
in the earlier years are rising much more than 4 percent. One would
suppose that earnings would be related to fund totals.
All of Holland's estimates implicitly assume a decline in the aggre-
gate funding percentage; that is, the extent to which accumulated lia-
bilities are covered by segregated pension fund assets.23 This may
`~ Daniel M. Holland, Private Pension Funds: Projected Growth, Occasional Paper 97,
National Bureau of Economic Research (New York, 1966). Holland includes State and
local government funds In his study.
`°Ibid., p. 143.
20 As a result of time elapsed between the preparation of the estimates and publication,
Holland was able to compare actual with predicted results for 1962-65. (Ibid., p. 87.) After
adjusting for an upward revision of $2.5 billion in 1961 (Holland's base year), his esti-
mates fall short of realized amounts by an increasing sum. Annual Increments for 1964
and 1965 are underestimated considerably.
~` Ibid., pp. 57-58.
22 Ibid., p. 118.
~ Holland has taken the position that he Is implicitly assuming that past funding prac-
tices will continue, Ibid., p. 146. I believe that he has, perhaps unawares, implicitly
assumed a declining funding percentage. My argument follows In the text.
PAGENO="0019"
OLD AGE INCOME ASSURANCE-PART V
indeed be the most realistic assumption; individual firms and plans
may shrink or terminate, but the entire business sector, and pension
funds in the aggregate, are likely to grow. Thus a lower funding per-
centage might be consistent with reasonable employee benefit pro-
tection. However, it would seem desirable to make this assumption
explicit, and to prepare estimates on the alternative assumptions that
the funding percentage will remain constant or even rise. These devel-
-opments may not occur, but one should consider the possibility that
they might, particularly in view of recent recommendations for earlier
vesting and better protection of beneficiaries' expectations.24
The assertion that a falling funding percentage is implied by all
-of Holland's estimates may be demonstrated as follows:
(1) With respect to those models in which the mean contributions
per covered employee are level for 20 years, but in which the mean
benefit paid rises, the funding percentage must eventually fall unless
fund earnings rise in an offsetting fashion, or life expectancy at retire-
ment declines, or vesting grows generally less liberal. It is not suffi-
*cient that fund earnings rise in absolute amount, but as a percent of
fund principal as well. It can be shown that fund earnings have not
risen sufficiently, but that observation is irrelevant here; it suffices to
point out that Holland's calculations do not assume a rising investment
return on pension funds. None of the other required developments have
occurred, or are assumed in his model. Therefore, in Holland's con-
stant contribution, rising benefit model, a falling funding percentage
is implied.
(2) With respect to his alternative calculation in which contribu-
tions, benefits, and fund earnings rise at 4 percent (and which calcu-
lation produces the highest estimate of fund levels) the demonstration
is a bit lengthier. My proposition holds, nonetheless; a falling fund-
ing percentage is implied. Let us approach the matter by considering
a fully vested, fully funded pension system. There would then be for
each worker a "fund" which would rise from zero, at entry into a
pension plan, to a maximum at retirement. This maximum would be
a fund sufficiently large to provide benefits (allowing for investment
income on the declining balance), through the beneficiary's remaining
life expectancy. If each age cohort contained the same number of
covered workers (or beneficiaries), and if benefits remained the same,
the average ftmd for covered workers (halfway builtup) would
roughly equal the average fund per retired worker (halfway run-
down). Shifting the population toward more beneficiaries and fewer
workers does not, therefore, lower the required fund as long as the
age distribution among the workers and retired remains approxi-
inately equal. In a growing population (with constant benefits) the
mean fund per worker is smaller, since it is pulled down by the larger
number of younger workers, and the mean fund per retired benefici-
aries is larger, since there -are relatively more newly retired workers
who have not run down their funds very far.
In a growing population with rising wage rates and benefits, the
average worker's fund is larger, and the average retired person's
fund smaller, than in the last example, since younger workers' funds
reflect more recent higher wages and benefits. Nevertheless, and this
is the crucial point, aggregate funds for both covered employees and
24 Such recommendations are cited In footnote 6.
PAGENO="0020"
12 OLD AGE INCOME ASSURANCE-PART V
retired persons are growing at rates which reflect both rising benefits
and rising numbers of persons. That is to say, one should not expect
aggregate pension funds in a fully funded situation to grow less
rapidly simply because more and more persons are retired. Retired
persons require accumulated funds as do covered workers.
Holland's highest alternative calculation, as we have said, is based
on the assumption that. wages, contributions, and benefits per worker
are rising steadily at 4 percent. It. is further based on estimates of
covered workers and beneficiaries in the period 1961-81 which grow
at about 3.2 and 8.1 percent per annum, respectively. WTe should, there-
fore, expect a fully funded system (reflecting his assumption) to grow
roughly at a rate which is the sum of the 4-percent growth rate in
benefits and the weighted mean rate of growth in numbers of peo.ple,
or some rate between 7 and 12 percent. Now the funds with which we
are dealing are not fully funded but they would grow a.t the same rate
as our fully funded example if the funding percentage remained the
same. Holland's highest projection, on the other hand, grows at a rate
with starts at 10.7 percent and declines to 4.7 percent in 1981. His esti-
mate, therefore., implies a falling percentage of funding, taking pri-
va.te pension plans in the aggregate, since constant funding would
reqmre some minimum growth rate higher than 7.2 percent (4 percent
plus 3.2 percent).
My own estimates, as indicated, suggest that private, State and local
government., and Federal Civil Service funds will reach a level of $450
to $470 25 billion by 1981, if the percentage of funding does not fall.
This compares with Holland's $312 to $339 billion for Private (incluci-
in~ State and local government) p]ans.2G Only a. small part of the
difference relates to my inclusion of the Federal Civil Service Retire-
ment. Fund; most relates to differences in the estimates.
Holland's careful study may well be on target if funding percent-
ages do decline. The base from which he started was also revised up-
ward after his calculations were complete, but before his work was
published. Starting from a corrected 1961 base would have raised
most of Holland's 1981 estimates by $7 to $8 billion. My own esti-
mate is presented with some benefit of hindsight, and to broaden the
range of alternatives which policymakers face.
V. IMPLICATIONS OF THE SIZE OF Nox-OASDHI FUNDS
The funds which are t.he subject of this inquiry now tot.al some $150
billion; in 1980 they will likely reach $350 to $470 billion, depending
on developments in coverage, funding practices, and like variables.
The implications of this development reach into many areas of the
economy, and raise many quest.ions. Most of t.hese are explored else-
where in this Symposium; only one or two will be pursued here.
The share of the national wealth which is owned through pension
fund intermediaries will rise from the present 6 t.o 7 percent to some
10 percent. The share of all common stocks which is owned by the
~ My estimate is prepared on the following assumptions: the growth rate in funds will
decline evenly from 10.5 percent (the 1965 rate) to 6 percent in 1980. 6 percent is
regarded as the minimum rate at which funds will grow, based on Holland's growth rate
in covered workers (3.1 percent) and a rather low estimate of the likely annual increase
in money wages. It is further assumed that benefits will, in the longrun, move roughly in
line with money wages.
~ Ibid., p. 143.
PAGENO="0021"
OLD AGE INCOME ASSIJRANCE-PART V 13
pension funds will likely rise from the present 5 percent, though too
many variables are involved to make any projections worth serious
consideration. Perhaps the major issue discussed here is the impact
of pension saving on aggregate saving, and the resulting public deficits
which might be required to keep the economy in the neighborhood of
"full employment," somehow defined.
If one takes the view that our major problem has been that of too
little saving, then there is no policy problem at all. If one takes the
other view that private planned saving, including pension saving,
is likely to exceed planned investment in most years at recent rates
of growth, then a policy problem emerges. A problem emerges; that
is, unless one takes the neoclassical view that the interest rate, the
profit rate, the capital-output ratio, and factor substitutability all
smoothly adjust to the higher implicit growth rate. In the "oversav-
ing" case, which I think, on balance, to be the most likely case, Fed-
eral policies might take two directions: one is to realize the growth
rate implied by a savings share and savings level which tends to out-
run "sustainable" investment ("sustainable" in the sense discussed
above). This would involve measures to accelerate technological
change, along with other measures to raise the rate of increase in labor
productivity such as basic education, retraining, elimination of bar-
riers to entry, relocation allowances, and the like. Lower interest rates~
along with measures to raise apparent capital productivity (such as
the investment credit) could also be part of the program.
To the extent that the "equilibrium" (in a rough sense) growth rate
cannot be accommodated to the (possibly rising) full-employment sav-
ing ratio, there is implied a second policy measure. This measure is a
fiscal policy resulting in a growing Federal debt (monetary ease as a
stimulant is really part of the growth-accelerating policy set mentioned
above). Thus, as we have suggested, pension funds might be accumu-
lating financial claims while the Federal debt is growing. Relative
yields on stock and Government securities would have to adjust in such
a way that pensions funds would acquire at least some of the Govern-
ment securities. The lower interest rates implied by the "oversaving"
case, incidentally, will tend to raise the volume of pension funds re-
quired to maintain a given funding level.
As a final statement, one cannot reach normative judgments about
the future of pension funds and aggregate economic variables without
first forming some judgment about future aggregate demand prospects
for full employment. This is a hazardous territory indeed; the last 50
years are littered with the shattered reputations of competent econ-
omists who went astray at this point. My own view, which surely must
have come through clearly in the preceding pages, is that investment is
likely in most years to fall somewhat short of full-employment private
saving. The growth-accelerating policy set would be my own first
choice. To the extent that full-employment private saving is not there-
by absorbed, maintaining a roughly defined full employment growth
course will require Federal deficits and a rising debt.
An alternative policy measure which seems to merit careful consid-
eration is the development of a reinsurance plan which will assure em-
ploye benefits but require a lower level of funding for private plans.
This possibility is explored elsewhere in this symposium. It is im-
portant to realize that reinsurance has implications beyond benefit pro-
PAGENO="0022"
14 OLD AGE INCOME ASSIJRANCE-PART V
tection per Se. To the extent that it bears on funding practices, rein--
surance is related to all of the macroeconomic variables which have-
been examined in this study, and, in particular, represents a policy
substitute for Federal deficits.
The macroeconomic analysis of pension plans, then, raises a number-
of important issues. Policy implications are unavoidably present, par--
ticularly with respect to aggregate saving, growth, and full employ--
ment.
PAGENO="0023"
THE SOCIAL INSURANCE PARADOX*
BY HENRY AARON**
In 1958 Professor Samuelson demonstrated that social insurance;
could improve the lot of each person in society.' Each person could sup-
port a portion of the cost of maintaining the retired population in
return for which future generations would support him during his
retirement. The return on this form of saving would exceed the real
rate of interest which was in turn a function of population growth.
In this note I present a proof of a related theorem: that social insur-
ance can increase the welfare of each person if the sum of the rates
of growth of population and real wages exceeds the rate of interest.
I call this theorem, after Samuelson, the social insurance "paradox."
This theorem does not conflict with Samuelson's because nothing is
assumed here about the level of the interest rate or its determinants,
other than it equals both the marginal rate of time preference and the
marginal rate of transformation between present and future goods.
I. TIlE PARADOX STATED
Ideally, private insurers set premiums to equalize the present ex-
pected value of benefits and the present expected value of premiums.2
A reserve is accumulated for each person. As the number of people
covered by insurance increases, the total reserves for all people in-
crease. If the age composition of the insured population remains the
same, the rate of increase of total reserves is the same as the rate of
growth of the insured population. If the same procedure were followed
in social insurance, and if the covered population increased, then the
total reserves would likewise increase. In a mature system the amount
of reserves would increase at the same rate as population grows if the
age composition of the population remains constant. If social insur-
ance premiums were set below the level which would equalize the pres-
ent expected value of benefits and premiums, the rate of increase of
the social insurance trust fund could be reduced. In an extreme case,
no social insurance reserves at all might be accumulated. It follows,
that if the covered population increases and if no reserve is accumu-
lated, the present expected value of premiums is less than the present
expected value of benefits for each person, although for the group as
a whole viewed in perpetuity, the equality will be maintained.
The logical implication of the above remarks is that, if no reserve
is accumulated in the financing of old-age pensions, each person will
receive a larger pension than he has paid for. Or, put another way, the
* Reprinted from the Canadian Journal of Economics & Political Science, Vol.
XXXII, No.3, August 1066.
** Professor, University of Maryland.
2 A. Samuelson, "An Exact Consumption-Loan Model of Interest With or Without
the Social Contrivance of Money," Journal of Political Economy, LXvI (Demember 1958),
467-482.
2 ~ am abstracting here from profits, commissions, and other administrative expenses.
15
PAGENO="0024"
16 OLD AGE INCOME ASSURANCE-PART V
present value of his lifetime income, earmngs plus pension, will be
higher if no reserves are accumulated than if they are. This fact holds
for ever person. Consequently, for the Nation as ci whole, the present
expected value of the sum of real net lifetime receipts is greater `when
reserves are not accumulated than when they are, although the national
income at each point in time is unaffected. Has a fallacy crept in?
From an insurance equity stanclpomt it has. The fact that the same
pension is paid under the pay-as-you-go assessment system and the re-
serve system, although premniins are higher in the latter case than in
the former, means simply that a higher interest rate is implicitly being
paid on contributions under the "no reserve" system than under the
reserve system. If this higher interest rate is used in calculating each
person's implicit fund under assessment financing and in discounting
social insurance benefits to the present, the equality between expected
present values of premiums and benefits is restored.
II. THE PARADOX DEMONSTRATED
On the macroeconomic level, however, the paradox remains intact.
This section will be devoted to a rigorous demonstration of this ~r0p0-
sition. Assume that, upon reaching a particular age A, all people
enter the labour force and that the number reaching this age year
grows at. the rate of g percent. Assume that average real wages grow
at the rate of h percent and that the interest rate is i percent. Let
t=l+g, s=1+h, and r=l+i. Each person works vs years then retires.
He lives in retirement from n - vs years dying at the age of A + n.
During each year of retirement each person receives a pension equal
to the average wage then prevailing among active workers.
After Ic years (where lc>n), the total population, P, will be:
(1) P=P0t~ [ (tm ± t~-' ± . . . + t) ± (1 + t-' + t-2 + . . . + t'~÷~)],
where P0 is the size of the retiring cohort at time 0. The first paren-
thetical expression inside the brackets, multiplied by P0t0, gives the
total working population; the second parenthetical expression, multi-
pliedi by POt~, gives the total retiredi population.
The total wages any worker receives during his working life, w,
will be:
(2) w=w0sk(s_m+s_;71÷l+ . . . +8_i),
where the worker retires in year Ic, and W~ is the wage prevailing in
year 0.
During retirement the same worker will receive pensions, p, in-
dicated by:
(3) p=wos7~(l+s+s2+ . . . +s~-'~'-~).
During his working life, each pension recipient had paid taxes to
support pensions. Assuming that taxes were just sufficient to cover
pension costs: i.e.. that no reserve was acdumulatedl, each worker paidl
a fraction of his wage, f, equal to the ratio of the retired to the active
population. This fraction, f, is equal to the ratio of the second paren-
thetical expression in equation (1) to the first. Assuming that these
PAGENO="0025"
OLD AGE INCOME ASSIJRANCE-PART V 17
taxes had been invested, as they would be in private insurance or in
funded social insurance, they would have a present value, PVT~ at
the time the worker retires shown by equation (4).
(4) PJ7T = /w0sk(s-~mrm +s~1r1~ + . . . + S1r).
Assuming that at retirement the worker discounts future benefits
at the interest rate i, the present value of pensions, benefits PVB, is
shown by equation (5).
(5) PJ7B = W0S~ (1 + s~--' + ~2r-2 + . . . + sn-m_lr-n+rn+1)
The paradox presented in section I states that PVB>PVT when r is
below some value. I shall now show that the critical value of r occurs
approximately when i= g + h; i.e., where the rate of interest equals the
sum of population growth and the increase in real wages. To do so it
is necessary to establish the precise conditions under which PVB>PVT.
Writing both out in full, moving the numerator of f to the left of the
inequality sign, and canceling where possible, the inequality may be
written as follows:
~6 \ (1+ S?~1 - s2r2 +. . . + sn_m_lr_n+rn+1) (S_mTm + s_m+lrm_1 +.. . + s'r)
` ` (1+ t-1 + t-2 +.. . + ~n+rn÷1) > (t~ + tm-' +. . . + t)
Now, the right and left expressions are equal when rst. In that
case the numerator on each side becomes identical, term for term, with
the denominator. The derivative of the left (right) -hand expression
with respect to r is clearly negative (positive). Consequently for
smaller values of r, the inequality holds. For larger values, it does not.
When, r, s, and t differ only slightly from unity, the condition that
r=st can be approximated by the condition that i~g+ It.
III. WELFARE IMPLICATIONs
The results of section II show that if the sum of the rates of growth
of per capita wages and of population exceeds the rate of interest, and
if the rate of interest equals the marginal rate of time preference and
the marginal rate of transformation of present into future goods, then
the introduction of some social insurance pensions on a pay-as-you-go
basis will improve the welfare position of each person. If saving and.
hence, investment and, hence, the rate of growth of income are reduced
as the level of social insurance increases, this conclusion does not neces-
sarily follow. If the rate of growth is unaffected, the effective rate of
return on premiums paid for such social insurance will exceed the
marginal rate of time preference, and, consequently, people in the
active labour force would willingly forego some current consumption
in order to obtain such returns. Individually they are unable to do so;
collectively they can.
If a small trust fund is accumulated, the proceeds from which are
invested, the addition to welfare will be smaller than if no fund is
accumulated, and in the limiting case of a full reserve, no increase in
welfare will occur.
If the rate of interest exceeds the sum of the rate of growth of real
wages and the rate of growth of population, then introduction of social
PAGENO="0026"
18 OLD AGE INCOME ASSURANCE-PART V
insurance either on a pay-as-you-go or a funded basis will reduce wel-
fare, unless (a) market imperfections render the preexisting situation
suboptimal, (b) the social welfare function calls for income redistri-
bution, or (c) there are economies of scale in social insurance; none of
these factors will be discussed here.
PAGENO="0027"
A REEXAMINATION OF THE PURE CONSUMPTION
LOANS MODEL*
BY DAVID CASS and MENAHEM E. YAARI**
I. INTRODUCTION
Professor Samuelson's article (1958) on an exact consumption-loan
model of interest led to an interesting controversy. At issue were the
determination and properties of interest rates in a dynamic economic
system with no capital. One might have thought that these exchanges,
between Samuelson (1959) and Lerner (1959) on the one hand, and
between Samuelson (`1960) and Meckling (1960) on the other, would
result, if not in a complete resolution of the disagreements, at least in
the emergence of a clear picture of what the issues are and how they
might be treated. Unfortunately, such was not the case. The Samuel-
`son-Lerner and the Samuelson-Meckling dialogues leave the reader
rather perplexed, as though he had just watched a New Wave film-
`executed with brilliance, enjoyable while in progress, but not quite
`clear as to what is happening and never giving one a sense of resolu-
*tion.
`The 1960's have brought an upsurge of interest in capital theory
`and, more generally, in questions of allocation over time. Now Samuel-
son's model, even though it has no capital, is of interest to capital
`theorists because it has many of the features of a model of capital
~iccumulation with decentralized decisionmaking. Since the 1958-60
discussions of this model by Samuelson, Lerner, and Meckling left
some questions unanswered, we feel that a restudy of the model is in
order. The following is an attempt at such a restudy.
We shall take the liberty of deviating somewhat from Samuelson's
original notation.
II. THE MODEL
We shall concentrate throughout most of our discussion on the sim-
pler of Samuelson's two models; namely, the one in which people live
for two periods, earning a fixed income in the first and earning noth-
in~ in `the second. In the first period of his lifetime, a person earns one
unit of output, where "output" is something usable directly (and ex-
clusively) in consumption, and we do not inquire wherefrom it comes.
The generation which is born at time t will be called oeneration t,
and we shall assume that there are (1 +n)t people in it. `iI~ius, n is the
(relative) rate of growth of population, which is assumed constant.
Members of generation t are thought of as being alike in all respects,
*Reprinted from The Journal of Political Economy, August 1966, vol. LXXIV,
No.4.
**yale University. We are indebted `to Richard A'ttiyeh, Herbert Scarf, and
James Tobin for their comments. This essay is based on research which was sup-
ported by the National Science Foundation.
19
PAGENO="0028"
20 OLD AGE INCOME ASSURANCE-PART V
so one can speak of a member of generation t without specifying the
individual. Let the symbol C~ stand for the first-period consumption
of a member of generation t, and let C~ stand for his second-period
consumption. A member of generation t is assumed to value any given
consumption plan (C~, CD according to the value, U(C~, CD, taken
on by a "regularly shaped" utility function U at (C~, CD. The utility
function U is assumed to be the same for all generations.
Continuing in Samuelson's footsteps, we now proceed to assume
that output is nondurable and thus cannot be carried over from one
period to the next. This assumption reduces the production possibili-
ties in the model (that is, the possibilities of using output in one
period in the production of output in another period) to naught. It
is clear, furthermore, that the assumptions made so far are so restric-
tive as to rule out from the outset any possibility of trades, markets,
or prices. A member of generation t who wishes to engage in a trans-
action cannot find anyone willing and able to participate in the trans-
action on the opposite side.
Given that production and trade have both been dispensed with,
there remains only one other economic activity to be considered-
distribution. This function is still open in our economy, for output
can be taken from the young who earn it and given to the old who
do not. Thus, our first task will be the examination of alternative dis-
tribution schemes. However, before proceeding with this examina-
tion, we must define the notion of "the rate of interest." Writing rt
for the rate of interest in period t, we define:
_C~_1+ C~~1-l
1 rn
`-` t -1
or ~ ± r~= 1
Two things should be noted: (a) 1 +Tt is not a price; that is, no
transactions are ever held using it as a rate of exchange. It is not
even an "implicit price." in the sense of a price which emerges as a
byproduct of efficient allocation of resources. It is, rather, an ex post
rate of exchange which is inferred from observation of the consump-
tion pattern of a. member of generation t- 1, and it has reference
neither to trade nor to efficiency.' (b) When C~_,= 1, r, is clearly
not defined. If O~2=l and C~_1=O, we shall say that r, can be any
real number, and if C~_~=l and C~_,>O, we shall say that rt=±o0.
III. Dismrnn~riox OP OUTPUT
As has already been remarked. the only economic function remain-
ing in the mod~l of the foregoing section is that of distribution of
output. A pair of sequences,
{ C~, t=O, ±1, ±2, . . .
{ C~, t=O, ±1, ±2, . . .
with nonnegative elements will be referred to as a distribution sc7~erne,
~ Note, however, that if a member of generation t-1 were to maximize utility subjeet to
a given r:. then 1+rt would, as usual, equal the marginal rate of substitution of second-
period consumption for first-period consumption.
PAGENO="0029"
OLD AGE INCOME ASSURANCE-PART V 21
It specifies how much each member of any given generation shall
consume in each period of his lifetime. A distribution scheme will be
called feasible if it does not use up more output than is available in
any period. In other words, the scheme ({ 0~ }, { 0~ }) is feasible if
and only if
(1+n)tU~+(1+n)t10~i~ (1+n)t
or
for all t, where the inequality arises from an assumption of free dis-
posal.
Suppose that in period t no disposal takes place. Then,
On the other hand, we have, by the definition of the rate of interest,
that
c~+~t=1.
Subtracting the latter from the former, we get
U' C' T~Th 02 -o
,t tm(l+T~(l+fl) i-l-
As an immediate consequence one now obtains that
ifG~_,=0~~1, thenr,=n.
If C~1 = C~ =1, then r, can be any real number, and we might as
well agree once again that r,=n.
This equality of r, and n in case C~, =C~ is the manifestation in
the present framework of what Samuelson (1958) calls "the biological
rate of interest," which he finds "paradoxical," even "astonishing" (pp.
471 and 473, respectively) and which Meckling finds very hard to
accept. The main cause for suspicion seems to be the fact that a rate
of interest has been determined without any reference to impatience
and time preference or, more generally, to the utility function U. Some-
how, the fact that rt is a completely mechanistic construct, having no
reference to markets, seems to have become blurred.
IV. STATIONARY SCHEMES
In order to analyze the relation between the rate of interest and the
rate of growth of population somewhat further, let us define a station-
ary distribution scheme by the requirement
g;Ig:~for all t,
which implies that
Ui + 02_i
rt=r= for all t.
PAGENO="0030"
22
OLD AGE INCOME ASSURANCE-PART V
We are assuming for the moment that 01<1, so r is well defined. Now
let us check the feasibility of a stationary distribution scheme. By di-~
rect substitution, we find that
~ + ~1 if and only if r~n.
Thus, the feasibility of a stationary distribution scheme is equivalent
to the statement that the rate of interest is no greater than the rate of
growth of population. Furthermore, the same algebraic operation
which yielded the equivalence of the above inequalities also yields the
equivalence of the strict inequality rn is not.4 If r m, C,~ tends to 0, and, therefore,
Similarly, if the rate of Interest is required to be constant up to some point, then
r>n is possible, but r 0 such that rt~n- c for t 0,1,2, . . . , then
urn U2=0.
t->co
As a particular instance, this discussion applies to paths along
which the rate of interest is always zero, which is what Lerner seems
to advocate.
Now let us drop the assumption that output is completely perishable
and substitute in its stead the assumption that output is completely
durable.
The first thing to notice is that trades are still impossible. A mem-
ber of generation t who wishes to give up a quantity of output at time
t in return for a quantity of output at t + 1 can never find anyone
who is offering to give up a quantity of output at time t + 1 in return
for a quantity of output at time t. On the other hand, a technology is
now available to a member of generation t with which he can trans-
form output at time t into output at time t + 1, unit for unit.
Suppose that with this new regime we now require our economy
to proceed in an entirely decentralized fashion, each individual acting
on his own. The opportunity set available to each individual is now
given by the triangle OAB in Figure 2. (Recall that, when output
was nondurable, the individual's opportunity set consisted only of the
line segment OA.) An individual of generation t will, therefore, pick
his consumption plan (G~,O~) so as to maximize U(G~,G~) subject
to C~± G~= 1, which will lead him to a point, say D, on the bound-
ary of his opportunity set. Clearly, D will not in general coincide with
A; that is, the individual will in general choose to consume positive
amounts in both periods. Not also that the point D will be optimal
for individuals of all generations (since we are assuming a common
utility function for all) so that decentralized behavior leads, once
again, to a stationary path.
Now let us look at distribution schemes which are not necessarily
decentralized. Obviously, any scheme which was feasible under the
assumption that output is nondurable is also feasible under the as-
I
PIG. 2
VI. Du1tM~ii~ OUTPUT
Sa-200-G7--pt. 5-3
PAGENO="0034"
26 OLD AGE INCOME ASSURANCE-PART V
sumption that output is completely durable. Hence, the set of feasible
distribution schemes in our new regime certainly contains the feasible
set of the old regime. In particular, the set of all feasible stationary
schemes must contain ~ the shaded set in Figure 1, which appears in
Figure 2 as the triangle OAU.
A brief look at Figure 2 tells the story: While in the old regime
decentralized behavior led to an efficient. but (in general) nonoptimal
distribution scheme, in the new regime decentralization leads (in gen-
eral) to an inefficient (to say nothing of optimality) distribution
scheme.
This phenomenon, the inefficiency of decentralized behavior, does
not disappear if we drop yet another assumption and permit capital
(which in this model is in the form of inert inventories) to become
productive. This has been shown recently by Diamond (1965) ~6 In
Diamond's model, the inefficiency appears only if people "want to
save too much" (in some well-defined sense). In the present model,
people in the decentralized scheme ahvays want to save too much, in
the sense that the efficient distribution schemes involve zero inven-
tories at every moment of time, while the decentralized scheme (in
general) involves positive inventories at every moment of time. (In~
Diamond's model this is not always the case because, roughly speak-
ing, as long as inventories reproduce faster than people, it is efficient
to hold them.)
VII. FINANCIAL INTERMEDIATION
Thus far, we have avoided the question of how an efficient distri-
bution scheme might be brought about. Is there an economic agent that
could be introduced into the model and whose activity would ensure
efficiency? Before attempting to answer this question we must probe
a little further into the nature of the inefficiency of decentralization
in the model of the foregoing section.
At the heart of the inefficiency in our new model (with durable out-
put) lies the fact that decentralization forces people to hoard output
in their first period so as to be able to consume what they had hoarded
in the second period. The result is that in every period a fraction of -
total output is put aside in the form of a savings fund, to be carried
over the next period. But, when the next period comes around and
the older generation consumes its savings, the younger generation
establishes its own savings fund and the economy ends up carrying a*
load of deadweight, in the form of output which is never consumed.
Indeed, under the assumption of the same utility function for all, this
load of deadweight keeps growing like a geometric progression, be- -
cause the savings fund of generation t is 1 +n the size of the savings
fund of generation t-1.
The only way to restore efficiency to our system is to find an arrange-
ment whereby the savings of generation t (when it is young) are used
to provide for the consumption of generation t- 1 (when it is old). In
~ Actually, the two sets coincide.
6 Note that the present model, with durable output. may be looked upon as a special case
of Diamond's model, with the intensive production function (that Is, the function relating -
ouput per man to capital per man) identically equal to unity.
PAGENO="0035"
OLD AGE INCOME ASSURANCE-PART V 27
this way, the situation in which positive amounts of real output are
always being carried over, from period to period, will not arise. But,
this cannot be done in a decentralized manner within the present
framework of assumptions. The first thing which comes to mind at
this point is that maybe the trouble lies in the lack of sufficient overlap
between generations, which prevents the right trades from being made.
In order to investigate this possibility, we introduce into the picture
a financial intermediary of some sort (say, a banking system or a sys-
tem of pension funds) which is assumed to exist concurrently with
all generations. People are now able to save by holding the liabilities
of this financial intermediary, which they will in fact prefer to do if
by holding such liabilities for one period they can earn a positive rate
of interest. Now, the intermediary can, in fact, offer holders of its
liabilities a rate of interest of n per period simply by using the output
deposited with it by generation t to redeem the liabilities which are
held by generation t 1. The result is that all output available in pe-
riod t is in fact consumed in period t, partly by generation t and partly
by generation t -1. Generation t -1 receives whatever generation t de-
cides to save as payment in full (principal as well as interest) of the
debt incurred to it one period earlier by the financial intermediary.
There will no longer be a carryover of output from period to period
and, as a result, efficiency will be restored. Indeed, optiinality will be
restored as well, because in response to a rate of interest of n per period
people will choose to save exactly the amount which leads to the opti-
mal distribution scheme.
The outlook seems rosy until one takes a brief look at the balance
sheet of our financial intermediary, where things are rather unfor-
tunate: The balance sheet as of the end of period t shows zero assets
and liabilities of s(l+n)t, where s is the (stationary) saving ratio of
people in their first period of life. This means that at the end of period
t the net worth of the intermediary is given by -s (1 +n) t. Now by not
doing anything (that is, by shutting down) the intermediary can
guarantee itself a net worth of zero, and so one might argue that it
will never choose to engage in the aforementioned transactions. On the
other hand, it might be argued that the intermediary should be looked
upon as a social security system which is not privately owned, so that
its networth position is of no concern (or should not be of any con-
cern). This is what Lerner seems to be saying when he decries (1959,
p. 517) "those of the accountants who insist on the `solvency' of the
Social Security Administration." Later, Lerner seems to be taking a
180° turn by insisting on the very accounting practice which lie had
previously decried: "`Business is fine,' said the optimistic contractor.
`It is true that I lose money on every contract, but I always start a
bigger one and get an advance that more than covers the loss on the
old one'" (1959, p. 523).
Be that as it may, it is certainly the case that a privately owned
financial intermediary will not rescue the economy from inefficiency.
In other words, decentralization and competitive behavior still fail
to result in efficient behavior.
PAGENO="0036"
28 OLD AGE INCOME ASSURANCE-PART V
VIII LACK OF BORROWING AS A SOURCE OF INEFFICIENCY
Let Wt be the amount of output carried over by society from period t
into period t+ 1. If we restrict our attention to stationary states, then
it is clear that a necessary and sufficient condition for efficiency is
.wt=O for all t. However, if we wish to consider the nonstationary
cases as well, the condition for efficiency becomes somewhat more com-
plicated. For it is possible, for instance, for generation t to under-
consume and for generation t +1 to overconsume, with the result that
~J~t will be positive a.nd yet the economy will remain efficient. More gen-
erally, t.he building up of inventories does not destroy efficiency so long
as these inventories are eventually consumed. The exact statement in
this respect is as follows: A necessa~~y and sufficient condition for the
economy to be efficient iB that there exist a subsequence {Xtk} of the
sequence {Xt} such that
lim Xtk=O
The proof of this assertion will be given in the Appendix. Roughly
speaking, efficiency requires that inventories return periodically to a
level which is "practically zero."
From the point of view of balance sheets, xt is clearly net a~ssets
(total assets minus total liabilities) in the economy at the end of period
C. If we concentrate our attention at the time periods t~, we find that
efficiency requires net assets at the end of these periods to be (prac-
tically) zero.7 In other words, in order to have efficiency, it must be
the case that for each outstanding asset in the economy at the end of
period tk there exists a corresponding liability outstanding at the end
of period t~. But, now let us recall the time structure of our model.
People live for two periods, they consume in both periods but earn
income only in the first. This forces individuals to become net lenders
(and never net borrowers) no matter what rate of interest prevails.8
In particular, members of generation tk will want to be net lenders,
that is, to have an end-of-period balance sheet which shows just assets
and net worth and no liabilities. But efficiency requires that aggregate
net worth in the economy be zero, so there must exist someone in the
economy holding liabilities in excess of assets, that is, having a nega-
tive net worth at the end of period tk. This puts the unfavorable net
worth position of the financial intermediary of section VII in per-
spective. By the same token, the decentralized economy of section VI
is inefficient precisely because no one can be a net borrower while
everyone wishes to be a net lender.
It is of some interest to investigate Samuelson's discussion concern-
ing "the contrivance of money" in the light of the foregoing remarks.
Clearly, if efficiency is to be attained, people must be dissuaded from
holding output as a. store of value and persuaded to hold another
asset instead. If this other asset bears a positive rate of interest, then
people will in fact make the desired shift. As Samuelson points out,
the role of this other asset can very well be fulfilled by money. Peo-
The word "practically" is intended to convey the notion that, strictly speaking, net
assets at the end of period tk might be given by some c~O, but by taking k large enough,
this c may be taken as small as we wish. From here on, we shall neglect to remind the
reader of this qualification.
As long as this rate of interest exceeds -1.
PAGENO="0037"
OLD AGE INCOME ASSURANCE-PART V 2~
pie would save by buying the existing money supply and dissave by
selling it to next period's savers. On a stationary path, the price of
money will rise by a factor of 1 +n in each period, which corresponds
to a rate of interest of n per period, so individuals will in fact prefer
holding money to holding output. Thus, at least with respect to sta-
tionary paths, the introduction of money leads to efficieucy. (More
precisely: Every stationary path in the money economy is efficient. )
Samuelson's interpretation of this phenomenon is a philosophical one:
An economy is inherently more than a mere mechanical system of
particles in motion; it is, in fact, such a system plus something cailedi
a "Hobbes-Rousseau social contract" (Samuelson, 1958, p. 479). A.
physical system can operate efficiently without this added aspect, but a
social system cannot.9 Now, it seems to us that the social contract is
no more involved in Sarnuelson's money economy than it is in any
other general equilibrium model. For this reason we feel that Samuel-
son's discussion in this area is liable to be misleading. In general equi-
librium analysis one thinks of a single market convention in which
prices are announced and economic agents determine the trades and
the productive activities in which they wish to engage at these prices.
If the totality of all trades clears all markets, then the announced
prices are said to be equilibrium prices. The question which general
equilibrium theory asks is the following: Under the assumption that
everybody at the market convention takes prices as given, is there
a schedule of prices which leads to the clearing of all markets? In
Samuelson's money economy, commodities are time-dated output and
time-dated money, and all that one asks is whether or not a given price
schedule is an equilibrium price schedule. It turns out that the sched-
ule which sets the price of output in all periods equal to 1 and the
price of money in period t equal to (1 +n) is, in fact, an equilibrium
price schedule. The element of public trust in the monetary unit is re-
flected by the fact that a person who buys money in period t, at a price
of (1 +n) t, assumes that he will be able to sell it in period t+ 1, at a
price of (1 +n) t+~. But this is precisely what is meant in general equi-
librium theory by the phrase "taking prices as given."
At the beginning of this section, we argued that, if efficiency is to
be attained, someone will have to have a balance sheet showing
liabilities in excess of assets. For this reason it seems appropriate to
look upon money as a liability of a. monetary authority that is com-
mitted to paying one dollar to whoever presents it with one dollar. The
balance sheet of this monetary authority shows only liabilities and no
assets, and the value of the authority's liabilities (quantity of money
multiplied by the price of money) is precisely equal to the excess of
assets over liabilities in the private sector. From this point of view,
Samuelson's "contrivance of money" is, in essence, no different from
the financial intermediary of section VII.
We turn now to a brief investigation of a model in which people
live for three periods, and in which efficiency can be achieved (under
some circumstances) without introducing into the economy a sector
with negative net worth.
It is not clear whether Samuelson intends his model to constitute a mathematical proof
of the Locke-Hobbes-Rousseau thesis.
PAGENO="0038"
* 30 OLD AGE INCOME ASSURANCE-PART V
IX. A THREE-PERIOD MODEL.
We have argued that in order to have efficiency there must be some-
one in the economy who is, at least periodically, a net borrower. Along
stationary paths, this must be the case in every period. However, there
was nothing in that argument to suggest that t.he net borrower must be
a net borrower throughout his (or its) lifetime. Indeed, this role may
well change hands over time, which was not the case with the ftnancial
intermediary of section VII.
The prime candidate for this state of temporary net borrowership
is the consumer himself. To check on this possibility, let us consider
the following modification of our model: Assume that people live
for three periods rather than two. In the first period of life a person
grows up and is educated and therefore earns nothing; in the second
period he works and earns one unit of output; in the third he is re-
tired. Every person will now be a net borrower at the end of his first
period, a net lender at the end of his second period, and he will be
neither a borrower nor a lender at the end of his last period. (We are
not using the terms "negative net worth" and "positive net worth"
because it is not clear that they are applicable in the present context.)
We shall restrict our attention to stationary paths. Along a stationary
path, everybody receives the same lifetime consumption proffle, say
(0', 02, 0'). Feasibility and efficiency now means that the following
equation holds:
(l+r)0'+O'+
which is similar to the feasibility and efficiency equation in the two-
period case and is derived in the same way. If we let the (constant)
rate of interest along the stationary path be denoted r, then each per-
son's budget constraint is given by
(1+n)C'+O'+ i~r1~
which leads, once again to r=n. In other words, the only rate of in-
terest which can be established along an efficient stationary path is n.
Each person is now viewed as maximizing the utility* fuiiction
U(01,02,03) subject to the budget constraint (with n replacing r).
The maximization will result in an optimal consumption plan, say
(01,02,03). This optimal triple must be examined to see if it can be
brought about by purely competitive (decentralized) trades.
The optimal consumption plan (01,02,03) will be said to be com-
petitively attainable if the following equation holds:
03= (1+n)20'.
To see how this condition is obtained, consider a person of genera-
tion t. When he is young, he borrows an amount 0' from members
of generation t -1, who are in their middle years. Ne~ period, when
he is middle-aged and generation t-1 is retired, he returns the loan,
plus interest. In other words, he returns the amount (1+n) C'. Now
aggregath borrowing by generation t (when young) is given by the
amount (1 +n) ~0', and, therefore, total payment by generation t
PAGENO="0039"
OLD AGE INCOME ASSURANCE-PART V 31
(when. middle-aged) to generation t -1 (when retired) is given by
(1 +n) t+11J1* This quantity, divided by the size of generatiOn t 1,
yields the per capita consumption of the retired, that is, 0~. This is
precisely what the foregoing equation says.
As an example, consider the utility function
U(01,02,03) =~ log C~.
i=1
Maximizing it, subject to the budget constraint, leads to
(1+n)~'= O2=°=I/3
which is clearly competitively attainable.
Why is the decentralized economy in this example efficient? Surely,
it is possible to attribute this result to "the social contrivance of bind-
ing contracts." It is clearly in the interest of the middle-aged to de-
fault and ignore the debt which they have incurred when young. Even
if the rules are such that a person guilty of default is denied access
to the capital market as a lender (and so must lose interest on his
savings), it is still true in many cases that default will result in
increased consumption in all periods. Here, once again, is an oppor-
tunity to appeal to the social contract and here, once again, it would
seem to be beside the point, and for very much the same reasons as
before: The assumption that contracts are not defaulted upon usually
goes without saying in the theory of the competitive mechanism; it
does not exp~ain our result, it merely permits it.
But our example is a very lucky one. For it is not, in general, to
be expected that the optimal solution of a consumer's lifetime alloca-
tion problem will satisfy as stringent a condition as C~ (1 +n) 201.
In fact, with most utility functions this condition will not hold. If
0~> (1 +n) 201 then we shall, once again, have too little borrowing in
the economy and an additional agent with negative net worth would
he needed in order to attain efficiency. However if 0~ < (1+ n) 201 then
we shall have too much borrowing in the economy, and it will be pos-
sible to introduce a financial intermediary with positive net worth
which will guarantee efficiency. The balance sheet of this intermediary
will show I.O.TJ.'s of young people on the~ asset side and nothing on
the liabilities side. In order to see how this intermediary would operate,
let us concentrate, once again, on period t. In this period, a member
of generation t offers to sell a quantity Cl of I.O.TJ.'s. The total supply
of I.O.U.'s by generation t is therefore given by (1+n) tCl. However,
a member of generation t-1 wishes to buy only C~/(1 +~i) inl.O.TJ.'s,
so that the total quantity of I.O.TJ.'s demanded is (1 +n) t203, which
falls short of the quantity supplied. The intermediary now steps in
and buys the excess supply, using as payment the resources which
* generation t -1 pays in when it redeems its own I.O.U.'s, which it
sold to the intermediary in the previous period. Stationarity insures
that these resources will always be precisely adequate to buy the out-
standing I.O.TJ.'s. The result will be that physical output will not be
carried over from period to period, so that efficiency will be attained.
PAGENO="0040"
32 OLD AGE INCOME ASSURANCE-PART V
The aggregate debt of the consumer sector to the fiancial intermediary
will grow like a geometric progression, but each consumer by himself
will be balancing expenditures and receipts over his lifetime; that is,
at the end of his last period of life he will have zero net worth. Under
these circumstances, the intermediary may very well be thought of
as a privately owned, competitive institution.
X. EFFICIENCY AND INFINITY
The possible inefficiency (or nonoptimality) of the competitive
mechanism, as demonstrated by Samuelson and Diamond, has given
rise to a certain amount of speculation, mostly on an informal basis.
Many people (including Samuelson [for example, 1958, p. 474; 1959,
p. 522] and Diamond [1965, p. 1134]) seem to feel that this phenom-
enon has something to do with infinity. What apparently leads one to
point an accusing finger at "infinity" is the fact that for the standard
general equilibrium model (which is finite) we have theorems which
tell us that the competitive mechanism always leads to an optimum
(and, a fortiori, to efficiency). Nevertheless, the role played by "in-
finity" in leading the competitive mechanism astray has remained, at
best, rather vague. In the present section, we wish to explore this ques-
tion somewhat more systematically by trying to construct a finite
model that resembles the infinite model of the foregoing discussion as
closely as possible. As it turns out, inefficiency may well arise in such a
finite model.
Consider an economy with rn agents and m commodities (where
m>2). Each agent is both a consumer and a producer. Let C~ be the
amount of commodity j consumed by agent i, and let Q~ be the amoirnt
of commodity j produced by agent i. Agent i (for i=1, 2, . . . , ~m- 1)
is assumed to desire, for consumption, only two commodities: com-
modity i and commodity i+ 1. (Agent in is assumed to desire com-
modity `in and commodity 1.) Thus, agent i's utility function is given
by
U~=U(G~, O~')i=i, .. ., rn-i, Um=U(G~, ~
where the function U is common to all. As for production possibilities
we assume that agent i can produce commodities i and i+ 1 (with
agent m producing commodities `in and 1), but that he has a relative
advantage in the production of commodity i. More specifically, we
shall assume that agent i can produce any combination of Q~ and Q~'
satisfying
Qi+1
Q2~O, Q~~1»=~j, ~
where ~ is some real number satisfying O<8<1.'°
Recall for a moment the infinite model of Section VI (with durable
output). That model is easily shown to be mathematically equivalent
to a model in which population is stationary, while the storing of out-
`° we shall, from now on, neglect to write separate expressions for the case i=rn. Let us
agree, therefore, that whenever i=m, i+i is simpiy 1.
PAGENO="0041"
OLD AGE INCOME ASSURANCE-PART v 33
put is subject to depreciation at some constant rate, say 8. But the
model which we have described in the foregoing paragraph is an exact
finite analogue of this infinite model with depreciation. For, by chop-
ping off a finite segment of the infinite model and then tying the two
ends together to form a closed loop, one obtains the model that is
being discussed here.
Let us, therefore, investigate the efficiency of the competitive mech-
anism in our finite model, to which we shall henceforth refer, for
short, as the closed-loop model. This investigation turns on whether or
not intermediation is permitted.
Case a: No intermediation.-Under the assumption that intermedia-
tion is altogether absent, it is evident that the closed-loop model is
inefficient. For, without intermediation, an agent who wishes to trade,
say, x units of commodity j for y units of commodity k, must find, in
order to be able to make the trade, an agent who wishes to trade y
units of comodity k for x units of commodity j. This is, of course, the
well-known "double coincidence of wants" (see, for example, Samuel-
son, 1964, p. 51). Now, in the closed-loop model, things are tailored
in such a way that whatever trade an agent might wish to engage in,
at whatever prices, he can never find another agent wishing to engage
in the same trade on the opposite side. In other words, the double
coincidence of wants never occurs. In the absence of intermediation,
therefore, each agent must satisfy all his wants under complete au-
tarky, and the doctrine of comparative advantage tells us immediately
that this is inefficient. More formally, the decentralized solution of
the closed-loop model in the absence of intermediation is given by
for i=1, . . .,m,
where( Q~, Q~~1) is chosen so as to maximize U(Q, Q~~1), subject to
Qi + 1
~ Q~o, Q~+1~
Under the assumption that 6>0, this solution is inefficient. The econ-
omy can produce more of every commodity.
Now let us go back for a moment to the model of section VI. The
inefficiency which beset the economy of that section is precisely the
inefficiency which now besets our closed-loop model. All the arguments
(indeed, the very words) of section VI are applicable to the closed-
loop model, with only minor changes which have to do with substitut-
ing depreciation for population growth.
Case b: The money economy.-The necessity for a double coinci-
dence of wants disappears, as is well known, as soon as barter is
abandoned in favor of the money economy. Thus, one might expect
the introduction of "the contrivance of money" to cure the afore-
n'ientioned inefficiency, just as it cured the inefficiency of the model
of section VI. Such, indeed, is the case. In the money economy, agent
i will produce commodity i exclusively, sell some of it to agent (i-i),
and then use the money which he receives in return to buy some of
agent (i+ 1) `s output. More formally let ~i be the money price of
PAGENO="0042"
34 OLD AGE INCOME ASSURANCE-PART V
commodity i. Then, the equilibrium of the money economy is de-~
scribed by
p~=p~ for all i a.nd j,
Q~=O if i!=j
=1 if i=j,
and (G~, C~÷1) chosen so as to maximize U(C~, Ca),
subject to G~+O~ =1, U~O, 6~'»=O.
This equilibrium is efficient, and it is precisely analogous to the equi-
librium in the money economy that was discussed in section VIIL
People "buy" money in return for goods which they produce, and then
"sell" it for goods which they want. Basically, money here provides
intermediation. Agent i sells to agent (i-i), but he is being paid
in terms of output of agent (i±1).
There are several ways in which money can be introduced into the
closed-loop model. The ~impJ.est is as follows: Think of agent 1 selling
a promissory note to agent 2, who, in turn, proceeds to sell it to agent
3, and so on. The note travels around the loop until it reaches agent in,
who redeems it from agent 1. All other agents, however, accept agent
l's note not because they are interested in agent l's output, but simply
because the note is negotiable. In other words, agent l's note will
assume the role of money.
Case c: The ordii2ary geiiei~al ejv.ilibrivim model.-The closed-loop
model is, after all, a finite general equilibrium model, and for the latter
we have theorems concerning optimality and efficiency. What role,
then, does intermediation play in the standard general equilibrium
theory? The answer to this question is straightforward: In all of gen-
eral equilibrium theory it is assumed (mostly implicitly) that a central
clearing house exists, through which trades are channeled. Only with
such a. central clearing house is it possible to define competitive equi-
hbrium in terms of aggregate excess demands alone. If a central
clearing house were to be introduced in the closed-loop model, com-
petitive behavior would lead to efficiency, as may be verified directly
(without appealing to general theorems). But, the central clearing:
house is obviously an intermediary. It is, in fact, precisely the ana-
log, in the closed-loop model, of our negative net worth intermedi-
ary of section VII.
XI. CONCLUDING REMARK
Before closing, we would like to add a brief comment concerning the
relation of the topics we have been discussing here to the Modigliani-
Brumberg (1955) "life-cycle" theory of saving. Modigliani and Brum-
berg have postulated that aggregate saving can be explained by the
interaction of individual saving for retirement and changes in th~
population structure. It is interesting to note that under the Modigli-
am-Brumberg assumptions of a zero rate of interest and an expo-
nentially growing population, providing for old age in a way which
results in positive aggregate saving is inefficient.1'
`~ We are Indebted to James Tobin for this observation.
PAGENO="0043"
OLD AGE INCOME ASSURANCE-PART V 35
APPENDIX
Let ~t be the level of inventories carried over by the economy from
period t into period t+1. In section VIII a necessary and sufficient
condition for efficiency was stated. The following is an equivalent
assertion:
THEOREM: The economy is efficient if and only if
lim inf x~=O.
t-~oo
PROOF. (a) Necessity: Suppose urn inf x~>O. Then there exists an
c>O and an integer q' such that Wt>C for t~T. Let aggregate~ con-
sumption in period T be increased by , and let consumption in all
subsequent periods remain unchanged. Let the sequence of inventory
levels under this new scheme be denoted { x }. Clearly, x= x~ for
t~ T and X~=Xt-E for t>T. All that must be verified to show that
the new scheme is feasible is x «= 0 for all t. But this follows from
the hypothesis.
(b) Sufficiency is proved similarly: Assume lim inf xt=O. `We must
show that it is impossible to increase aggregate consumption in any
period while keeping aggregate consumption in subsequent periods
unchanged. Let aggregate consumption in period T be increased by an
amount c>O, where T is an arbitrary integer. Denote the new sequence
of inventory levels by { x}. Now for t> T, we have x= x1 - ~, but by
hypothesis there exists a t> T such that ~ Hence x would have
to be negative; that is, the new scheme is not feasible.
REFERENCES
Diamond, Peter A. "National Debt in a Neoclassical Growth Model," A.E.R., LV
(December, 1065), 1126-50.
Lerner, Abba P. "Consumption-Loan Interest and Money" and "Rejoinder,"
J.P.E., LXVII (October, 1959), 512-517, 523.
Meckling, W. H. "An Exact Consumption-Loan Model of Interest: A Comment"
and "Rejoinder," J.P.E., LXVIII (February, 1960), 72-75, 83-84.
Modigliani, F., and Brumberg, R. "Utility Analysis and the Consumption Func-
tion: An Interpretation of Cross-Section Data," in Post -Keynesiai~ Economics,
ed. K. K. Kurihara. London: George Allen & Unwin, 1955.
Phelps, Edmund S. "Second Essay on the Golden Rule of Accumulation," A.E.R.,
LV (September, 1965), 793-814.
Samuelson, Paul A. "An Exact Consumption-Loan Model of Interest with or
without the Contrivance of Money," J.P.E., LXVI (December, 1958), 467-482.
"Reply," ibid., LXVII (October, 1959), 518-522.
-~ . "Infinity, Unanimity, and Singularity: A Reply," ibid., LXVIII (Febru-
ary, 1960), 76-82.
Economics, an Introductory Ana'ysis. 6th ed. New York: McGraw-Hill
Book Co., 1964.
PAGENO="0044"
ECONOMIC ASPECTS OF PENSIONS: A SUMMARY
R,EPORT*
BY ROGER F. MTIrn~Y
RELATION OF THE DIRECTORS TO THE WORK AND PUBLICATIONS
OF THE NATIONAL BUREAU OF ECONOMIC RESEARCH
1. The object of the National Bureau of Economic Research is to ascertain and
to present to the public important economic facts and their interpretation in
a scientific and impartial manner. The Board of Directors is charged with the
responsibility of ensuring that the work of the National Bureau is carried on
in strict conformity of this object.
2. To this end the Board of Directors shall appoint one or more Directors of
Research.
3. The Director or Directors of Research shall submit to the members of the
Board, or to its Executive Committee, for their formal adoption, all specific
proposals concerning researches to be instituted.
4. No report shall be published until the Director or Directors of Research shall
have submitted to the Board a summary drawing attention to the character of
the data and their utilization in the report, the nature and treatment of the
problems involved, the main conclusions, and such other information as in their
opinion would serve to determine the suitability of the report for publication in
accordance with the principles of the National Bureau.
5. A copy of any manuscript proposed for publication shall also be submitted to
each member of the Board. For each manuscript to be so submitted a special
committee shall be appointed by the President, or at his designation by the
Executive Director, consisting of three Directors selected as nearly as may be
one from each general division of the Board. The names of the special manu-
script committee shall be stated to each Director when the summary and re-
port described in paragraph (4) are sent to him. It shall be the duty of each
member to the committee to read the manuscript. If each member of the special
committee signifies his approval within thirty days, the manuscript may be
published. If each member of the special committee has not signified his ap-
proval within thirty days of the transmittal of the report and manuscript,
the Director of Research shall then notify each member of the Board, request-
ing approval or disapproval of publication, and thirty additional days shall be
granted for this purpose. The manuscript shall then not be published unless at
least a majority of the entire Board and a two-thirds majority of those mem-
bers of the Board who shall have voted on the proposal within the time fixed
for the receipt of votes on the publication proposed shall have approved.
6. No manuscript may be published, though approved by each member of the
special committee, until forty-five days have elapsed from the transmittal of
the summary and report. The interval is allowed for the receipt of any memo-
randum of dissent or reservation, together with a brief statement of his reasons,
that any member may wish to express; and such memorandum of dissent or
reservation shall be published with the manuscript if he so desires. Publication
does not, however, imply that each member of the Board has read the manu-
script, or that either members of the Board in general, or of the special com-
mittee, have passed upon its validity in every detail.
7. A copy of this resolution shall, unless otherwise determined by the Board, be
printed in each copy of every National Bureau book.
(Resolution adopted October 25, 1926. as revised February 6. 1933, and
February 24, 1941)
National Bureau of Economic Research, Number 85, General Series. This
report is one of a series emerging from an investigation of pension plans made
possible by grailts `to the National Bureau from the Maurice and Laura Falk
Foundation and the Life Insurance Association of Ameriëa. These organizations
are not. however, responsible for any of the statements made or views expressed.
36
PAGENO="0045"
ADVISORY COMMITTEE ON PENSION STUDIES
In the planning and review of its pension studies, the National Bureau
has benefited from the advice and guidance of this committee. The
committee's concurrence with the views expressed in this report,
however, is not to be assumed. The members of the committee are:
SOLOMON FABRICANT (Chairn~an), National Bureau of Economic Research
ROBERT M. BALL, Social Security Administration
DORRANCE C. BRONSON, The Wyatt Company, Washington, D.C.
J. DOUGLAS BROWN, Princeton University
GEORGE B. BUCK, Ju., New York, N.Y.
ARTHUR L. COBURN, JR., Old Colony Trust Company
JOHN J. ConsoN, formerly of McKinsey & Co.
FRANK G. DICKINSON, Northern Illinois University
F. F. FAURI, University of Michigan
MILTON FRIEDMAN, University of Chicago
GEORGE GARVY, Federal Reserve Bank of New York
WILLIAM C. GREENOUGH, Teachers Insurance and Annuity Association and Col..
lege retirement Equities Fund
CHALLIS A. HALL, JR., Yale University
RALPH W. HEMMINGEB, Bankers Trust Company
ALBERT J. HETTINGER, JR., Lazard Frčres and Company
REINHARD A. HOHAUS, formerly of Metropolitan Life Insurance Company
CHARLES L. JACOBSON, JR., formerly of State of Wisconsin Investment Board
E. GORDON KEITH, University of Pennsylvania
BENJAMIN B. KENDRICK, formerly of Life Insurance Association of America
LANE KIRKLAND, AFL-CIO, Washington, D.C.
MURRAY W. LATIMER, Washington, D.C.
J. D. LOCKTON, General Electric Company
DAN M. MCGILL, University of Pennsylvania
IDA C. MERRIAM, Social Security Administration
ROBERT J. MYERS, Social Security Administration
VITO NATRELLA, formerly of Securities and Exchange Commission
JAMES J. O'LEARY, formerly of Life Insurance Association of America
JosEPH A. PECHMAN, Brookings Institution
RAY M. PETERSON, formerly of Equitable Life Assurance Society
RENE L. ROTHSCHILD, formerly of California State Employees' Retirement
Systems
LAWRENCE H. SELTZER, formerly of Wayne State University
37
PAGENO="0046"
CONTENTS
Page
ACKNOWLEDGMENTS- 40
I. THE PRINCIPAL QUESTIONS
Retirement Saving in an Industrial Society 41
The Shape of Things To Come 42
The Questions in Perspective 44
A Central Theme 46
II. THE CHARACTERISTICS AND FINANCING OF PENSION PROGRAMS,
1940-80
Public and Private Pension Programs: An Overview 47
Old-Age, Survivors, and Disability Insurance 50
The Railroad Retirement System 52
Federal Employee Retirement Systems 52
Veterans' Pensions and Old-Age Assistance 53
State and Local Government Retirement Systems 54
Private Pension Plans 56
The Pension Structure in Perspective 57
III. PENSION PROGRAMS, THE REDISTRIBUTION OF INCOME,
AND SAVING
Public Programs:
The Primary Role of OASDI 58
Other Contributory Government Programs 61
Noncontributory Government Programs 61
Public Programs as a Whole 62
Private Programs:
The Saving Aspect 62
Transfers: The Contribution Aspect 63
Transfers: The Benefit Aspect 65
Pension Programs as a Whole 66
IV THE EFFECT OF PENSIONS ON AGGREGATE SAVING
The Issues Involved 69
The Cagan Analysis 71
Corroboration of Cagan's Findings 73
Future Trends 74
Possible Effects on Saving by Business and Government 75
The Impact on Total National Saving 77
V. THE IMPACT OF PENSIONS ON THE CAPITAL MARKETS: PRIVATE
FUNDS
Insured Plans:
Growth Trends 78
Portfolio Composition 79
Competitive Influences 80
Trusteed Plans:
Evolution of Investment Management 81
Portfolio Composition 83
Competitive Influences 85
Flexibility 86
Multiemployer and Union Pension Funds 88
Nonprofit Organizations 89
Future Capital Market Flows:
Recent Trends 89
A Look at the Future 92
38
PAGENO="0047"
CONTENTS 39
VI. THE IMPACT OF PENsioNs ON THE CAPITAL MARKETS:
GOVERNMENTAL PR0GRAITs
Federal Government Programs: Page
OASDI 96
Federal Employees
Railroad Retirement
Armed Services and Veterans 97
Federal Government Programs as a Whole 97
State and Local Government Systems:
Evolution of Investment Management 98
Restraints on Investment Management___ 102
Future Capital Market Flows:
Federal Government Programs 104
State and Local Government Systems 104
\TJJ~ CONCLUDING OBSERVATiONS
Implications for Saving and Economic Growth 107
Implications for Economic Stability 108
Implications for the Capital Markets -- 109
Implications for Financial institutions 113
Implications for Public Policy 113
TABLES
11-1. Benefit Payments Under Public and Private Programs, 1940-65~. 48
11-2. Contributions and Benefits Under Selected Public and Private Pro-
grams, 1940-65
11-3. Assets of OASDI Trust Funds, 1940-66 50
11-4. Past and Projected Growth of State and Local Government Retire-
ment Systems, 1940-80 55
11-5. Past and Projected Growth of Private PensionPlans, 1940-80 - - - 56
V-i. Growth in Life Insurance Assets and. Pension Reserves, 1940-65.. - - 79
V-2. Life Insurance Growth Rates, 5-Year Periods, 1940-65 79
V-3. Invested Assets of U.S. Life Insurance Companies, 1965 80
V-4. Bell System Pension Trusts, 1935-50 81
\T.5* Assets of Private Noninsured Pension Funds, 1965 84
V-6. Purchases of 25 Selected Common Stocks and Percentage of Total
Common Stock Purchases by Institutional Groups, January
1953-October 1955 87
V-7. Net Acquisitions of Financial Assets by Insured and Trusteed Pen-
sion Plans Covering Individuals in Private Employment, 1960 and
1965~_ 90
V-S. Possible Future Net Acquisitions of Financial Assets by Pension
Funds Covering Individuals in Private Employment, 1970, 1975,
and 1980
V-9. Possible Future Assets of Pension Funds Covering Individuals in
Private Employment, 1970, 1975, and 1980 94
VT-i. Assets of State and Local Government Retirement Systems, Fiscal
Years 1957 and 1966 100
VI-2. Possible Future Net Acquisitions of Financial Assets by State and
Local Government Retirement Systems, 1970, 1975, and 1980~ 105
CHARTS
V-i. Portfolios of Private Noninsured Pension Funds, 1955-65 84
VT-i. Portfolios of State and Local Government Employee Retirement
Funds, 1957-66 100
PAGENO="0048"
ACKNOWLEDGMENTS
Most of the research for this study was completed while I was S.
Sloan Colt Professor of Banking and Finance at the Graduate School
of Business, Columbia University. I am grateful for the research leaves
and other facilities which that professorial chair afforded me. The
cooperation of my colleagues at Teachers Insurance and Annuity
Association and College Retirement Equities Fund, where I now serve
as vice president and economist, has greatly assisted the writing of
this report.
During the course of our pension studies, I have benefited from the
cooperation and assistance of a large number of individuals. My senior
colleagues in the studies, Daniel M. Holland and Phillip Cagan, have
been an invaluable source of ideas and suggestions beyond the scope
of their specific study papers. Elizabeth T. Simpson has worked with
all of us throughout the project, correcting our mistakes and pushing
us to clarify and to make more precise our emerging conclusions. At
various points in the text, I refer to the major contribution of John
J. Carroll in his study of the income redistribution effects of pension
arrangements.
Working closely with me have been four extraordinarily capable
research assistants: H. Robert Bartell, Jr., Peter 0. Dietz, Richard L.
Menschel, and Thomas E. White. The first two of this group subse-
quently wrote their doctoral dissertations in aspects of the pension
field and have continued to expand the range of their contributions
to knowledge.
All of us engaged in these studies gratefully acknowledge the assist-
ance and suggestions received in bountiful measure from members of
the Advisory Committee on Pension Studies. As a panel of experts and
as individuals, they have aided us on innumerable general and specific
problems.
The guidance and suggestions of Solomon Fabricant and Geoffrey
H. Moore, the two Directors of Research of the National Bureau of
Economic Research under whose supervision these studies were pur-
sued, have *been invaluable. They and such other members of the
research staff as the late Frank G. Dickinson, Raymond W. Goldsmith,
F. Thomas Juster, and Ralph L. Nelson, to mention just a few, have
contributed once again to making a research undertaking at the Na-
tional Bureau a demanding, stimulating experience for the researcher.
I also thank, for their helpful suggestions, Joseph A. Beirne, A. J.
Hettinger, Jr., and Theodore 0. Yntema, who comprised the reading
committee of the Board of Directors.
James F. McRee, Jr., edited the manuscript and H. Irving Forman
prepared the charts.
The support of these research studies by grants from the Maurice
and Laura Falk Foundation and the Life Insurance Association of
America is gratefully acknowledged. Neither these organizations nor
any of the other individuals mentioned above are responsible for the
views expressed in this report.
ROGER F. MURRAY
40
PAGENO="0049"
I. THE PRINCIPAL QUESTIONS
The emergence of a new structure of fiscal operations, financial
institutions, and capital market participants inevitably raises a host
of questions about its implications for the American economy and, in
particular, the saving and investment process. Despite the fact that
the original Federal legislation for the old-age and survivors msur-
ance program dates back to 1935 and pioneer arrangements by private
industry to provide retirement income are 50 years old and more, the
present structure is properly described as new in the sense of only now
becoming of genuine economic significance. Now it is possible to
visualize the impact of public and private pensions on such basic
economic processes as saving and capital formation.
RETIREMENT SAVING IN AN INDUSTRIAL SOCIETY
How can a well-developed industrial society like the United States
save for the purpose of providing retirement income? As fewer in-
dividuals directly own the means of producing those goods and services
which they wish to use in retirement, it is evident that fewer of them
can store tangible assets for future consumption. In the New Testa-
ment parable, the rich farmer could build new barns to store grain
and other goods for years of future ease.' His present-day counterpart,
however, would hire a larger safe deposit box to store financial claims
and equities.
The point is simply that rising living standards are achieved
through the development of an economy which relies upon the ex~
changing of claims to goods and services rather than the satisfaction
of wants from the individual's own current output or a bartering of
consumption goods. The individual members of our 20th-century soci-
ety, therefore, cannot and do not save for their own retirement needs in
real terms. It is true that an individual can provide for some of his fu-
ture needs by acquiring outright ownership of his home. But, he cannot
store up, in physical terms, future property taxes, insurance premiums,
maintenance expenses, and the cost of cutting the grass.
That only a small fraction of saving for retirement can take the
form of accumulations of tangible goods, whether we think of in-
dividual cases or of society as a whole, means that the process of
saving and investment, the exchanging of financial claims, must de-
velop and grow to accommodate an additional type of transaction.
Furthermore, as arrangements are made to handle these transactions
outside the family unit, the claims and financial assets entering into the
money and capital markets are multiplied. The simplest illustration
is the case of the son who used to support his parents from current
income, but now supports them only as a member of the working pop-
ulation while they live on OASI benefits, a supplementa.l private pen-
sion, life insurance benefits, a savings account, and dividends from
common stocks.2
1Luke 12: 16-19.
2 The growth of financial institutions for such transactions has been explored in depth
by Raymond W. Goldsmith In Financial Intermediaries in the American Economy Since
1900, Princeton University Press for National Bureau of Economic Research, 1958, and
The Flow of Capital Funds in the Postwar Economy, New York, NBER. 1965.
41
83-200-67-pt. 5-4
PAGENO="0050"
42 OLD AGE INCOME ASSURANCE-PART V
This shifting to public and private organizations on a group basis
and the adoption of insurance-type arrangements are significant devel-
oprnents in themselves and account for the historical record of a high
growth rate for savings in this form. But, these secular trends were
jolted out of their regular pattern by a shift in the preferences of the
American people. In the long history of employment in manufactur-
ing, workers have accepted a substantial portion of productivity gains
in the form of a shorter workweek to enjoy more leisure time. From
1950 to 1965, on the other hand, there was no reduction in the average
number of hours worked. (Paid holidays a.nd vacations were increased,
to be sure.)
The same period witnessed a spectacular growth in pension coverage
for manufacturing industries. Workers did, in fact, receive a substan-
tial portion of their real gains in the form of future retirement bene-
fits, along with other income maintenance programs,. during the post-
war years. Also, this lasted long enough to be called a trend. It is rea-
sonable to conclude, therefore, that employee preferences have shifted,
to some extent, from the desire for more leisure during their working
years to a greater degree of financial independence during retirement
or to earlier retirement.
The main threads of the story. then, are as follows: (1) The long
shift in our society from agriculture a.nd a degree of self-sufficiency
in the family unit to an industrial, urban society; (2) the increase in
life expectancy, the widespread application of fixed retirement ages,
provisions for early retirement, and the resulting spectacular growth
in man-years of living in retirement; (3) a preference for a degree of
independence in old age as compared with other benefits during active
employment; (4) the growth in institutional arrangements for attain-
ing goals on a group, rather than individual, basis through .the em-
ployrnent of insurance principles.
Basic to all of these developments has been a change in the attitude
of the American people toward their old people, now sometimes known
as "senior citizens." Dependency on public or private charity based on
a means test is more widely recognized as destructive of human dignity
and the full potential for self-realization that lies in every individual.
Concurrently ha.s come the widespread acceptance of social insurance
as a normal institutional arrangement in our industrial society.
The economic aspects of pensions develop, then, from basic forces
at work in our society. We ca.n be confident that the structure of ar-
rangements will not be dismantled, nor will it fall into disuse in an-
other generation. Rather, it will grow and develop, spread and be-
come more inclusive, and become an increasingly dynamic force in
economic processes. For these reasons, too, we must look ahead, not
dwell upon history, to detect the most important questions and issues of
public policy and private action.
THE SHAPE OF THINGS TO COME
Obviously, the pension structure will evolve over the next decade and
more in response to many factors: the economic environment, rates of
growth in different sectors, employee preferences, cmiii public policy
decisions. Nevertheless, it is possible t.o identify the basic forces at work
and to rely upon them in projecting the probable course of flows of
PAGENO="0051"
OLD AGE INCOME ASSuRANCE-PART v 43
income and capital funds. Thus, one of the basic components of our
study has been a projection of the magnitudes of such determinants
of future economic influences as coverage, contributions, beneficiaries,
benefit payments, and fund earnings and accumulations.3
The questions which arise in making these projections are discussed
at length in Holland's study. The extrapolation of past trends has
a certain validity because of the long-time span of pension commit-
ments. Moreover, most of the pensioners in the year 1975 will already
have qualified for a large share of their prospective benefits. The costs
have been incurred and may have actually been firnded in whole or
in part through insurance contracts or trust funds. Despite revisions
in employee benefit plans from time to time and the increasing fulfill-
ment of pension promises, the basic framework of pensions for this
group is not likely to change materially.
Hence, it is possible to make helpful estimates not only of today's
fund flows but also of the transfer payments and capital market flows
of 1970, 1975, and 1980. Such estimates enable us to contrast with pres-
ent and past influences on income and expenditures those transactions
which `will be generated in the future. What differences will they make?
How will recipients of benefits respond to the creation and realization
of their equities in pension programs? Who pays the cost of these bene-
fits? How are the costs distributed overtime?
The answers to these questions may be different at various stages in
the evolution of the pension structure. It becomes essential, therefore,
to look at the system from several vantage points: In its early stages,
in its rapid growth phase, and in approaching maturity. In chapter
II, we have attempted this view of the structure of retirement income
arrangements from 1940 to 1980.
The purpose of this view is, first, to appraise the redistribution of
income brought about by the emerging pension structure. Chapter
III explores this process of income transfers. The incidence of costs
and the distribution of benefits are familiar problems of analysis in
other contexts. Our inability to produce precise answers should not
deter us from exploring the issues. We should even venture to face that
old question: Is there some limit to the pensions `we can afford? Is this
a meaningful question and is there any quantitative answer to it?
What might be the result in the distant future of the maturing of
pension commitments?
In addition to the direct influence of transfer payments on incomes,
spending, and saving, we should trace the impact of fund accumula-
tions under pension arrangements. How do they affect aggregate sav-
ing? Our discussion of this major question in chapter IV is based
largely upon Phillip Cagan's perceptive analysis.4 This thorny ques-
tion also has its time dimension, and the pattern of saving reactions
over future years becomes of great `interest in the formulation of eco-
nomic policy.
~ Daniel M. Holland, Private Pension Funds: Projected Orowtli, New York, NBER, 1966.
deals with these variables in projecting the fund flows through industrial pension plans
and the retirement syatems of State and local governments. This study will be referred to
as Holland's projections throughout this volume.
Phiuip Cagan, The Effect of Pension Plans on Aggregate saving: Evidence frons a
~5ample ~urvey, New York, NI3ER, 1965. This study will be referred to as Cagan's paper
on saving throughout this volume.
PAGENO="0052"
44 OLD AGE INCOME ASSURANCE-PART V
Entirely apart from these questions about the impact of the pension
structure on aggregate saving, however, we face a range of questions
relating to the influence on the capital markets of directing the flow of
funds through financial intermediaries known as pension trusts, life
insurance companies, and governmental retirement systems.5 The as-
sets acquired and held by these institutions are influenced, of course,
by the nature of their liabilities. Our projections of the fund flows
should, therefore, be a guide to the influences which we can expect to
see investment managers exert on the capital markets. For private
pension programs, these possibilities are discussed in chapter V, and
the operations of State and local government retirement systems are
analyzed in chapter VI.
In chapter VII, the findings of the study are pulled together in the
form of some conclusions, an identification of some of the implications
for public policy, and a recognition of the gaps in our knowledge.
about certain aspects of the subject.
THE QUESTIONS IN PERSPECTIVE
When viewed in their totality, t.he questions relating to the economic
effects of the development of public and private pensions are as all-
inclusive as a study of the American economy as a whole. Some nar-
rowing of the field of investigation was inevitable, therefore, to make
the project manageable. For example, we have not undertaken to study
the impact of pensions on labor mobility, a. matter of great impor--
tance in these days of rapid technological change. Nor have we been.
especially concerned with the tax treatment of pensions and the aged..
The role of pensions in contributing to economic stability is treate&
by implication, but this has not been treated as a major topic.
Otherwise, we have attempted to respond to most of the issues out-
lined in the National Bureau's exploratory survey, published in 1957,
as Suggestions for Research in the Economics of Pension-s. That sur-
vey attempted to delimit the economic aspects by excluding from con-.
sideration medical, sociological, and psychological problems of the~
aged. While admittedly of grea.t importance, these topics require spe-
cialized treatment a.nd a different preparation for the investigators.
The conception of this particular study was that it would build upon.
the work of the National Bureau in the fields of capital formation and
financial institutions. The timing seemed good since the research
program on postwar capital markets had been largely completed..
\\Tjthout that wealth of material, this study could not have been under-
taken with confidence. In a very specific sense, we are simply carrying~~
on these previous studies with particular emphasis on a new and
dynamic factor in the perpetually changing roles of financial inter-
inediaries.
In identifying the present and future impact of public and private
pension systems on saving and investment as the most important
economic aspects of pensions, we do not mean to imply that these are
H. Robert Bartell. Jr.'s study of union and jointly administered pension funds and
Elizabeth T. Simpson's study of the pension plans of nonprofit organizations. Penston
Funds of ~1fnltiem.ployer Industrial Groups, Unions and Nonprofit Organizations, fill in
gaps in our knowledge .about these special types of funds.
PAGENO="0053"
OLD AGE INCOME ASSURANCE-PART V 45
the oniy important considerations. Berle, Harbrecht, and Tilove
have discussed some of the issues arising from the institutional owner-
ship of a large and growing share of equity securities. Dan M. McGill
has dealt in a major study with the problem of assuring the fulfillment
of pension expectations.~ Similar concerns about the legal aspects of
pension commitments prompted Merton C. Bernstein to write The
Future of Private Pensions.8 Finally, the Cabinet Committee Report
of January 1965 has raised a wide range of issues regarding the regu-
lation of private pension arrangements.
These various studies, and numerous other writings, deal with prob-
lems of fairness and equity to the employee: matters of proper public
concern. As citizens, our sense of justice is particularly outraged if an
individual has been led to believe that provision has been made for his
old age and this turns out not to be the case. When he has lost his
capacity to support himself, it seems a particularly cruel form of in-
justice to have his expectations destroyed by careless, incompetent,
or venal administration of his retirement income program.
Our concentration on the economics of pensions should not be in-
terpreted as reflecting a lack of concern for this whole range of ques-
tions. We are simply trying to provide a basis for judgment about
the economic implications of the range of steps which might be taken
in the area of the public interest in the pension structure.
Comparable choices had to be made in deciding whether we ought
to be concerned with the adequacy of pension benefits. Here we de-
termined to express no judgments apart from our role of analyzing
the costs and economic impact of steps taken to improve the level of
benefits.'°
The narrowing of the focus of this study to the saving and invest-
ment process may seem to exclude some of the burning issues of our
time. Clearly, it skirts some areas of lively controversy. Yet, we make
no apologies for the design of the study, believing that we have at-
tacked the hard core of the subject. Among the wide range of topics,
we had to make a selection on the basis of our backgrounds and the
objectives of the National Bureau's research program. Furthermore,
we were concerned about losing sight of the relationship of manmade
pension arrangements to the underlying realities of the economic
process. Perhaps, in the final analysis, economic growth, economic
stability, and inflation are such prime determinants of our ability to
provide systematic programs to spread income beyond the years of
productive employment that they deserve the priorities which we have
given them by our concentration on the saving and investment process.
6 Adolph A. Berle, Jr., Power Without Property, New York, 1959; Paul P. Harbreeht,
Pension Funds and Economic Power, New York, 1959; Robert Tilove, Pension Funds and
Economic Freedom, New York 1959.
Summarized in his Fulfilling Pension Enpectations, Homewood, Ill., 1962.
New York, 1964.
Public Policy and Private Pension Programs, A. Report to the President on Private
Employee Retirement Plaits, by the President's Committee on `Corporate Pension Funds
and Other Private Retirement and Welfare Programs, Washington, 1965. Reference will
be made elsewhere to earlier studies by other agencies of Government.
10 The writer, as an individual, has participated in hearings held by the Subcommittee
on Retirement Income of the Special Committee on Aging, U.S. Senate, July 12-13, 1961,
and March 4, 5, 10, 1965.
PAGENO="0054"
46 OLD AGE INCOME ASSURANCE-PART V
A CENTRAL THE?~iE
As previously emphasized, in our urban industrialized society, we
deal primarily in claims rather than in tangible goods. A typical em-
ployee today accumulates the right to receive income after age 65 from
the OASI system of the Federal Government and from a pension plan
established by his employer. He can read bookints and calculate what
his benefits and those of his survivors will be if and when he attains
age 65 in his present employment.
These prospective benefits are translated by him into a certain degree
of financial independence, a level of living rather than a money in-
come. At that future date, he expects to have a reliable command over
the goods and services which others will be engaged in producing. His
very inability to provide for his own retirement by storing up a major
portion of his living standard makes him wholly dependent upon the
purchasing power of those accumulated claims.
If, by reason of stable economic growth during his remaining work-
mg years and during the period of his retirement, productivity gains
are substantial, those claims will give him a command over a good level
of living. Stable prices and quality improvements are essential for this
happy outcome.
Furthermore, he must take into account the attitudes of those who
will be providing him with goods and services during his years of re-
tirement. They must be willing to forgo current consumption of real
output in exchange for their own accumulation of claims to retirement
income. Their willingness to do so depends upon their standard of
living, a function of the productivity gains achieved in the economy
through investment in all forms.
If the claims to real output of the retired population seem burden-
some to the younger employed workers, they will be willing to tolerate
or even to stimulate a lightening of the burden through the traditional
route of inflation. On the other hand, rising real incomes for the work-
ing population make acceptable the diversion of real output to those
no longer working. This is especially so if widespread coverage by pen-
sion programs relieves the worker of personal responsibility for retired
members of his immediate family.
In short, the emerging pension structure has a great potential for
good if confidence in claims to retirement income is sustained by an
economy which is showing stable economic growth without inflation.
Since saving and capitai formation play a key role in this process, it is
crucially important to determine that public and private pension pro-
grams operate in a manner to contribute constructively to the kind
of economic progress which wifi translate their promises into realities.
The central theme of the research project summarized in this volume
is the search for a better understanding of the economic effects of the
mammoth structure of commitments, benefit payments, and fund ac-
cumulations which we are still in the process of creating. Is it well
designed to accomplish its exceedingly worthwhile objectives? Can it
be improved in some respects? Does it suggest the desirability of other
economic policies to reinforce or offset its effects? And finally, what
difference does it make if the shares in the total provision of retire-
ment income are altered as between private plans and the tax-sup-
ported programs of the Federal Government?
PAGENO="0055"
OLD AGE INCOME ASSURANCE-PART V 47
IT. THE CHARACTERISTICS AND FINANCING or PENSION PROGRAMS,
1940-80
A pension, broadly defined, is a regular payment, normally for life,
to provide retirement income to the recipient upon completion of his
working years. The payments may be made to the worker or his
spouse, or for dependent children. If withdrawal from gainful em-
ployment is occasioned by accident or ill health, the payments may be
called disability benefits, but the distinction is not important. It is
really immaterial whether "retirement" is caused by attainment of a
specified age or by inability to continue at work. For purposes of
economic analysis, the significant elements of the pension structure
are not the details of how a person qualifies for the benefits of a public
or private program, but, rather, the nature and financing of commit-
ments to individuals far into the future. Nor is it important whether
the contributions are based upon actuarial calculations or a share of
profits.
Our interest, then, is in any program, whatever the benefit formula,
whatever the source of funds, and whatever the terms and conditions,
which serves to maintain income for those who have partially or
wholly retired from gainful employment. Private charity and per-
sonal gifts, although obviously major voluntary elements in the care
of older people, are not within our purview because they do not repre-
sent systematic arrangements involving the shifting of group commit-
ments to agencies or institutions which are expected to fulfill them.
Individual saving for retirement and old age is less clearly distin-
guishable from pension programs and must be recognized at certain
points, even though it is not a part of the emerging structure for
groups of individuals.
In short, we are concerned with those arrangements which are clearly
distinguishable from traditional forms of personal saving and ex-
penditure. The questions, as we have seen, relate to the implications of
the present and prospective pension structure for the saving and in-
vestment process. What difference does it make that we have been
creating, and continue to create, an extensive network of arrangements
for the systematic provision of retirement incomes to most Americans?
What are the new and different factors introduced by this emerging
pension structure? What are the implications for economic policy
and for the capital markets? The starting point for answers to these
and related questions is an identification of the peculiar character-
istics of the present and prospective structure.
PUBLIC AND PRIVATE PENSION PROGRAMS: AN OVERVIEW
Somewhat arbitrarily, the period from 1940 to 1960 can be con-
sidered the period of the development and rapid growth of institu-
tional arrangements. By 1960, the public and private programs had
become firmly established, powerful influences in our structure of in-
come distribution. These were the decades of most rapid growth in
the creation of promises and expectations for greater financial security
in old age.
In contrast, it is possible to View a second 20-year period from 1960
to 1980, not only as a period of amendment and enlargement, but also as
PAGENO="0056"
48 OLD AGE INCOME ASSURANCE-PART V
years of rapid maturing in the fulfillment of expectations. That is to
say, by 1960 the pension structure had come of age. To a greater ex-
tent in the future, it will be providing benefits instead of expectations.
It is possible to measure the scope of the emerging pension struc-
ture in several ways: The extent of coverage, the number of bene-
ficiaries, the amount of benefits, and the funds being accumulated to
pay benefits in the future. Simply to trace developments since 1940
in the whole range of programs to maintain retirement incomes, how-
ever, it is helpful to look first at the growth in actual benefits paid
(table Il-i).
TABLE ll-1.--BENEFIT PAYMENTS UNDER PUBLIC AND PRIVATE PROGRAMS, 1940-65
[Annual totals in millions of dollarsj
Program
1940
1945
1950
1955
1960
1965
OASDI
Railroad retiremenL
Federal civil service
Other Federal employee
Veteran
Old-age assistance
State and local government
Private employee
35
118
68
54
424
475
142
140
274
147
94
76
952
727
193
220
961
311
192
301
2,224
1,478
320
370
4,968
577
380
489
2,746
1,765
595
850
11,245
962
816
779
3,437
2,216
1, 083
1, 750
18,311
1, 155
1, 384
1, 561
4,196
2,615
1, 780
3, 180
Total
1, 456
2, 683
6, 157
12,370
22, 288
34, 182
Percentage of personal income
1. 9
1. 6
2. 7
4, 0
5. 6
6. 4
Source: "Social Security Bulletin," Stat
istical Supple
ment, 1965; `
`Social Securi
ty Bulletin,"
April 1967.
The table traces the income maintenance effects of the whole range
of programs at various points in time. The six programs of the Federal
Government, it is strikingly apparent, provided 85 percent of all the
benefit payments disbursed in 1965. State and local government retire-
ment benefits and those provided by private programs had compound
annual growth rates of 10.6 percent and 13.3 percent, respectively, for
the 1940-65 period, but they were overshadowed by the Old-Age, Sur-
vivors, and Disability Insurance System. This reflects, of course, the
result of making increases in benefits applicable to present beneficiaries
and the different cha.racteristics of the programs.
The unique characteristic of a national social insurance system such
as OASDI is the matching, over a span of years, of collections from
employers, employees, and self-employed with current benefit pay-
ments. As the program matures and as benefits are increased, the rate
of tax on earnings will have to rise except as the maximum amount of
annual earnings taxable and creditable toward benefits is periodically
increased. By 1955, the OA~SDI system had reached the stage in its
evolution at which benefits and administrative expenses were equal
to 82 percent of total income, and by 1960 the proportion was 95 per-
cent. The initial contingency reserve fund had been accumulated and
receipts and disbursements approached an approximate balance.
Other Federal programs, such as those for veterans, retired members
of the Armed Forces, and individuals who qualify for old-age assist-
ance on the basis of need, are noncontributary pay-as-you-go pro-
grams, in which the cost of benefits is met by current appropriations
from general revenues. No attempt is made to anticipate future costs
or to create any kind of a reserve fund. Total receipts and the total
PAGENO="0057"
OLD AGE INCOME ASSURANCE-PART V 49
of benefits and expenses have been rising at the same pace, therefore,
and will continue to do so.
In contrast, the usual concept of funding aims to keep the contribu-
tion rate unchanged over a long period for a given benefit formula.
This sort of "level premium" per dollar of employee compensation, to
use the analogy from ordinary life insurance, involves prepayments
to be accumulated at interest. The excess of contributions over bene-
fit payments will be relatively large in a new program, in a newly
liberalized program, or in the case of a young employee group. The
accumulation of these excess receipts is expected to provide the earn-
ings necessary to close the gap between leveling contributions and
rising benefit payments as the program matures.
The relation between total contributions and the benefit payments
shown in table 11-2, for selected programs which are wholly or par-
tially funded, illustrates the lag of benefits behind contributions and
the degree of maturity achieved at different points in time.
TABLE 11-2.-CONTRIBUTIONS AND BENEFITS UNDER SELECTED PUBLIC AND PRIVATE PROGRAMS, 1940-65
[In millions of dollars]
Program
1940
1945
1950
1955
1960
1965
Railroad retirement:
Contributions 1
Benefit payments
Benefits as percentage of contributions
Federal civil service:
130
118
91
279
147
53
546
311
57
588
577
98
910
962
106
1,113
1,155
104
Contributions
Benefit payments
Benefits as percentage of contributions
State and local government:
Contributions
Benefit payments
Benefits as percentage of contributions
Private employee:
Contributions
Benefit payments_
Benefits as percentage of contributions
141
68
49
267
142
53
310
140
45
541
94
17
380
193
51
990
220
22
678
192
28
905
320
35
2, 080
370
18
744
380
51
1,740
595
34
3, 840
850
22
1, 610
816
51
2, 895
1, 083
37
5, 480
1,750
32
2, 197
1,384
63
4, 220
1, 780
42
7, 750
3,180
41
Source: Table Il-i; "Social Security Bulletin," Statistical Supplement, 1965; Institute of Life Insurance.
1 After 1954, contributions include transfers from OASDI under the financial interchange arrangement with the Railroad
Retirement System.
While the relation between benefits and contributions is affected by
a number of factors other than the maturing of the programs (such
as changes in employment, extension of coverage, extent of funding,
and integration with OASDI), it can be used as a very crude measure
of progress toward the maturing of a system.1 In these terms, we can
identify the railroad retirement system as having matured in 1955,
the Federal Civil Service Retirement System as now in middle age,
State and local government retirement systems in vigorous young
manhood, and programs covering individuals in private employment
as having rapidly outgrown adolescence.
In short, table 11-2 verifies the statement that by 1960 the pension
structure had come of age. What may lie ahead is discussed in the
brief reviews of the major programs which follow.
A retirement system is said to be mature when contributions plus interest earnings on
fund accumulations equal the sum of benefit payments and administrative expenses. This
precise definition of maturity Is rarely encountered in a growing economy. More frequently,
the transition from extremely rapid growth to a more moderate pace of asset accumulation,
the maturing process, takes the form of an increasing proportion of contributions being
immediately disbursed in the form of benefits.
PAGENO="0058"
50 OLD AGE INCOME ASSURANCE-PART V
OLD-AGED SURVIVORS, AND DISABILITY INSURANCE
Clearly identified as a social insurance system, OASDI emphasizes
the objectives of portability and inclusiveness. The objective of pro-
viding a basic level of income to all disabled o.r retired individuals,
their dependents, and survivors is served in the United States essen-
tially by a wage-related benefit, calculated by formula to replace a
higher portion of low than high earnings from past employment.
The setting of minimum and maximum levels of benefits and small
payments to individuals over 72 not otherwise covered gives the pro-
gram some characteristics of a fiat-rate plan, but the wage-related
features are deeply embedded in the structure.
With the attainment by 1955 of close to 90-percent coverage of eli-
gible individuals in paid employment, attention was `turned to in-
creasing OASDI benefit provisions and periodic revisions were made
to reflect changes in living costs. Because of their applicability to
beneficiaries, these changes have been promptly reflected in the amount
of benefits distributed. Currently, and in the future, the OASDI sys-
`tern is a major factor in Federal fiscal operations. Its impact came
largely through transfers of income during the 1957-67 decade, when
-the trust funds did not grow as holders of Federal securities. Table
11-3 shows the accumulation of the present reserve fund.
TABLE 1-3--ASSETS OF OASDI TRUST FUNDS, 1940-66
un millions of dollarsj
Year end
Old-age and
survivors in-
surancetrust
Disability
insurance
trust fund
Total
fund
1940
`1945
1950
1955
1960
1965
`1966
2 031
7,121
13,721
21,663
20,324
18,235
20,570
-
-
-
-
2,289
1,606
1,739
2,031
7,121
13,721
21,663
22,613
19,841
22,308
Source: "Social Security Bulletin," Statistical Supplement, 1965, and
SSB, July 1967.
Moderate fluctuations are more likely than substantial growth for
the trust fund accumulations in the future. Current projections show
sizable additions to assets in future years, but their realization depends,
at least, in part, upon the assumption of no changes in the law. In
reality, the trust funds serve primarily to take care of shortrun dis-
parities between payroll tax collections and benefit payments. Also,
the interest collected on the Government securities held by the funds
`amply covers the administrative expenses of the system. In any event,
changes in the assets of the trust funds are not a measure of the "sav-
ing" by covered individuals in the sense that we measure pension say-
`ing by adding together the assets of plans funded through life insur-
ance companies, pension trusts, or governmental retirement systems.
A different, but not necessarily appropriate, accounting approach
would show monumental totals on a balance sheet which would also
reflect, on the asset side. the present worth of income to be received in
the future from payroll taxes and interest on the trust fund assets.
The corresponding liability item would be the present worth of future
PAGENO="0059"
OLD AGE INCOME ASSURANCE-PART V 51
benefit payments plus the cost of administering the system. These
calculations are not customarily used, except for actuarial purposes
to determine actuarial balance, but the claims to future benefits are
just as real to the wage earner as an insurance policy or an equity in a
private pension plan. That these equities are very large is suggested
by the estimate that the survivorship protection of OASI alone had
a face value of about $700 billion at the start of 1966.2
The OASDI program, nevertheless, operates as a giant mechanism
for fulfilling retirement income expectations. The design of supple-
mental pension plans integrated with OASDI recognizes the reality
of this growing equity of the employee in a wide range of prospective
benefits. With the attainment of almost universal coverage, social se-
curity has become completely integrated into the flow of income and
expenditure in the economy. Prospective benefits are the major un-
recorded financial asset of a large proportion of American families.
Actual monthly benefits are going to more than 23 million individuals,
almost 16 million aged 65 and over, with the result that the reality of
the program is evident to all.
Periodic revisions to take account of rising living costs and living
standards will probably be made in the future as they have in the past.
Whether, as has been proposed, some adjustments will be made auto-
matically is currently being debated. In any event, even the present
program will generate higher levels of benefits as more individuals
qualify for monthly payments closer to the maximum level. Recent
cost estimates prepared by Robert J. Myers, Chief Actuary of the
Social Security Administration,~ suggest the following pattern for
combined old-age, survivor, and disability benefits:
[In millions of dollarsj
Year
Low-cost
estimate
High-cost
estimate
1965 (actual)
1970
1980
18,310
23,019
30,404
18, 310
23,498
31,725
Number of Beneficiaries
En~Z of Year: Thousands
1940 149
1945 177
1950 398
1955 633
1960 809
1965 911
Source: Social Security Bulletin.
These estimates do not, of course, allow for subsequent revisions in
the benefit structure. They simply illustrate that, under any particular
formula for determining benefits, the growth in payments is moderate
after the initial impact of a change made applicable to present as well
as prospective beneficiaries.
Robert M. Ball, "Policy Issues In Social Security," Social Security Bulletin, June 1966,
p. 4.
Robert J. Myers and Francisco Bayo, Long-Range Cost Estimates for Old-Age, Sur-
vivors, and Disability Insurance System, 1966, Social Security Administration, actuarial
study No. 63, January 1967.
PAGENO="0060"
52 OLD AGE INCOME ASSURANCE-PART V
As the OASDI system is modified and developed in the years to
come, it will continue as the major source of retirement income benefits
for as far ahead as one can see.
THE RAILROAD RETIREMENT SYSTEM
An interesting illustration of the maturing of a pension program is
found in the operations of the Railroad Retirement System. Disregard-
ing the active hiring during World War II, covered employment was
virtually unchanged at 1.2 million between 1939 and 1956. Since then
it has declined by about one-third. As a consequence, the median age of
employees rose from 41.5 to 46.2 years between 1950 and 1960, and is
continuing to rise. Covered payrolls, the base for financing pension
commitments, moreover, showed no net increase between 1950 and 1965.
Meanwhile, the pension rolls, including retirement, disability, and sur-
vivors benefits, skyrocketed as follows:
By the 1960's, therefore, the number of beneficiaries (which includes
survivors) began to exceed the number of covered employees, and this
condition is likely to continue. Given stable payrolls, the system ob-
viously had to increase the contribution rate sharply to meet the dou-
bling of benefits between 1955 and 1965 (see table IT-i) or draw upon
the trust fund at a substantial rate. This fund, after reaching $3,374
million in 1954, has grown slowly to $3,946 million at the end of 1965
and cannot support a large excess of benefit payments over receipts.
The Railroad Retirement Act Amendments of 1966, in addition to in-
creasing benefits, introduced a tax of 2 cents per hour on employers
to cover the current cost of a temporary supplemental program. The
regular combined tax rates on employees and employers exclusive of
the cost of hospital insurance, became 15.2 percent in 1966 and are
scheduled to rise to 19.2 percent by 1973. Even this schedule is esti-
mated by the actuaries to be insufficient to meet promised benefits.
While this industrywide program is not typical, it illustrates the
problems of a social insurance system when the vitality and growth
of the economic activity involved begin to erode. The. consequences of
failing to make provision in previous decades for accruing retire-
ment costs are also evident. It is apparent that railroad employees and
the public as consumers of rail transportation services will continue
to be the source of growing transfers of income to retired workers and
their dependents for many years to come. Since air transportation com-
panies currently provide for full pension costs, their competitive cost
position will be enhanced in the future~ relative to rail transportation.
Thus, the mechanism for handling retirement income payments may
have marginal effects upon the allocation of resources among competing
forms of transportation.
FEDERAL EMPLOYEE RETIREMENT SYSTEMS
The Federal Civil Service Retirement System was intended to be
fully funded, but Government contributions have lagged far behind
the accrual of benefit costs. The appropriation of additional funds by
the Congress to meet full costs apparently ranks low in the order of
priorities, in the absence of more serious concern among employees
about the fulfillment of the Federal Government's pension commit-
PAGENO="0061"
OLD AGE INCOME ASSURANCE-PART v 53
ments. As a consequence of past and present practices, the accumula-
tion of funds in the Civil Service Retirement Fund has been as shown
below.
Assets
Year (millions)
:1940 $634
1945 - 2, 182
1950 4, 202
1955 6, 477
1960 10, 480
1965 15, 981
Source: Social Security Bulletin.
By mid-1966, the unfunded liability of the system was more than
two and one-half times the size of the trust fund and is continuing to
grow even without taking into account the possible increase arising
from the cost-of-living adjustment factor.
The trend toward a rising ratio of benefit payments to contribu-
tions shown in table 11-2 suggests that the accumulation of funds
will slow in the years ahead as benefit payments continue to climb.
Other Federal employee retirement plans are largely supported
from general revenues. One of the major programs of this type is for
career military personnel. Benefits will be rising especially rapidly
during the next decade, reflecting the large number of erJ~rants into
the armed services in the early 1940's, who will be retiring. The prac-
tice of adjusting retired pay to new scales adopted for those on active
duty is an additional factor stimulating the growth of benefits. If we
assume that current compensation to members of the Armed Forces is
lower by reason of the retirement benefits, especially those relating to
early retirement, it is evident that the cost of the military establish-
ment is materially underestimated by the failure to recognize the cost
of future pension commitments as incurred.
The mounting claim on future general revenues is a substantial one.
On certain assumptions as to the levels of active-duty compensation,
the benefits could rise from $1.6 billion a year in 1966 to $6.0 billion
in 1980. The accrued but unrecognized costs to date probably exceed
$60 billion, but the actual amount of the commitment is difficult to
determine. The variables include the number of career officers and
enlisted men, the future pattern of total compensation, and early-re-
tirement provisions. Some estimates suggest that the cost could rise
from the current rate of almost 10 percent of total military pay and
allowances to 25 percent by 1980.
VETERANS' PENSIONS AND OLD-AGE ASSISTANCE
Permanent disability and pension benefits are, in effect, interchange~
able for the 3.2 million veterans and 1.9 million survivors receiving
$4.2 billion of benefits in 1965. The future growth will tend to occur in
the payments to veterans over 65 years of age with non-service-con-
nected disability. Obviously, the size and scope of the program will be
determined in the future. Presumably the extensive coverage of social
security is meeting a part of the problems of older veterans with
limited incomes, and this fact may moderate demands for pensions'
involving a means test.
PAGENO="0062"
54 OLD AGE INCOME ASSLTRANCE-PART V
The economic effects of veterans' disability benefits are not signifi-
cantly different from old-age assistance. Both are supported from
general revenues and are displaced to some extent by OASDI benefits.
By early 1966, over 48 percent of old-age assistance recipients were
receiving OASDI benefits. The proportion of the population over age
65 receiving old-age assistance has continued to fall from 22 percent.,
in 1950~ to about 11 percent as the proportion receiving social security
benefits has increased from 17 percent to ahnost 80 percent.
The displacement of old-age assistance by OASDI has, therefore,
proceeded at an orderly pace, yet special problems and individual case.s
will undoubtedly require the continuation of old-age assistance. In
1964, 17 percent of all workers awarded retirement benefits received
monthly amounts in the range of $32 to $40, and a similar proportion
of all retired-worker benefits being paid were at or below the $40
minimum then in effect. Earlier studies indicated that about one-
fourth of those receiving minimum benefits were recipients of public
assistance.4
Lack of recent covered employment and very low covered earnings
accounts for individuals receiving minimum benefits, with farmwork-
ers representing a disproportionate fraction of the cases. The continued
displacement of old-age assistance by OASDI depends upon providing
employment opportunities at minimum wage levels for more men and
women of working age, especially among the nonwhite population. If
a vigorously growing economy can make equality of productive job
opportunities a reality and can improve the preparation of individuals
to hold such jobs, OASDI is well designed to perform its function.
Because the benefits are wage-related, they improve rapidly as the in-
dividual gains regularity of employment at prevailing wage rates.
A sustained high level of employment and programs to improve the
skills of potentially employable individuals would, therefore, grad-
ually reduce the need for old-age assistance. However, future progress
cannot retroactively raise the low covered earnings records of individ-
uals now becoming eligible for benefits. If all goes well, we should
expect old-age assistance benefits to continue to grow in absolute
amounts but at decreasing rates, as in the 1960-65 period.
STATE AND LOCAL GOVERNMENT RETIREMENT SYSTEMS
Rapidly growing retirement systems of State and local governments
represent one of the most dynamic elements in the entire pension struc-
ture. A measure of the growth, the total wage bill for employees cov-
ered by retirement plans, increased from $8 billion in 1950 to $32.2
billion in 1965. The compo~md annual rate of increase was 9.7 percent,
compared with 6.8 percent for the similar Federal civil service
payrolls.5 Holland's projections assume a 4.7 percent per annum
growth in employment over the next 10 or 15 years, which is close to
the rate of growth during the 1950's and early 1960's.
Not only is employment growing rapidly but pension coverage of
full-time employment is at an extremely high 95 percent. The prom-
ised benefits are relatively liberal, a.nd the widesprea.d (but not uni-
Lenore A. Epstein. "Workers Entitled to Minimum Retirement Benefits Under
OASDHI." Social Security Bulletin, March 1967, pp. 3-13.
5 Social Security Bulletin, Statistical Supplement, 1965.
PAGENO="0063"
OLD AGE INCOME ASSURANCE-PART V 55
versal) practice of systematically funding them through a broad range
of financial assets makes State and local government retirement sys-
tems substantially similar to private plans in their irnpa~t on the
capital markets. The salient features of their past and projected
growth are shown in table 11-4.
TABLE 11-4.-PAST AND PROJECTED GROWTH OF STATE AND LOCAL GOVERNMENT RETIREMENT SYSTEMS,
1940-80
1940 1945 1950 1955 1960 1965 1970 1975 1980
Covered employees (million) 1. 4 1. 8 2.6 3. 5 4. 5 5. 8 7. 6 9. 6 12. 2
Beneficiaries (million) . 2 . 2 . 3 . 4 . 7 . 9 . 9 1. 1 1. 3
Benefit payments (billions of dollars) . 1 . 2 . 3 . 6 1. 1 1. 8 2. 0 2. 6 3. 4
Retirement system assets (billions of dollars) 1. 6 2. 5 5. 2 10. 6 19. 7 33. 6 50. 7 76. 9 114. 9
Source: 1940-65, actual from Social Security Administration and Bureau of the Census; 1970-80, from Holland's pro.
jections adjusted to calendar years.
These retirement systems are significantly different from the other
tax-supported programs. Substantial saving through the acquisition
of financial assets is involved in the relatively high rate of funding in
progress and in prospect. Additions to fund assets were 10.5 percent
of covered payrolls in 1965 compared with about 7.4 percent for the
Federal Civil Service Retirement System.
The most significant feature of the State and local government re-
tirement systems is that they are still relatively young and, according
to Holland's projections, show no signs of reaching by 1980 that stage
in the maturing process when benefit payments are a high proportion
of current contributions. Tha~t is to say, these programs over the future
will make an increasing contribution to personal saving and to capi-
tal market flows. If the projected growth in State and local govern-
ment employment materializes, these retirement systems will emerge
as a major financial institution with vigorous growth and influence
in the resource allocation process.
A comparison with Holland's projections for private employee pen-
sion programs, discussed below, will serve to illustrate the vigor of
this growth.
COMPOUND ANNUAL RATES OF GROWTH
[In percentj
1940-60
1960-80
State and local
Private
State and local
Private
government
employee
government
employee
Covered employees
Beneficiaries
6.0
6. 5
8.6
11. 6
5.1
3. 1
3.4
7. 7
Benefit payments
Fund assets
12. 7
13. 4
15. 5
16. 6
5. 8
9. 2
9. 3
6. 8
Source: Tables 11-4 and 11-5. Growth rates are approximate for the 1940-60 period because they are based on rough
estimates for the early years for private plans.
Relative to programs covering individuals in private employment,
this comparison suggests that a more rapid growth in coverage, com-
bined with a slower rate of growth in beneficiaries and payments to
them, will enable the State and local government retirement systems
to sustain the higher growth rate in asset holdings which first be-
came apparent in the early 1960's.
PAGENO="0064"
56 OLD AGE INCOME ASSURANCE-PART V
PRIVATE PENSION PLANS
The development of private pension programs by business firms,
labor unions, and nonprofit organizations has been especially rapid
since 1940. Holland's projections are shown in table 11-5; they illus-
trate the extent of the possible maturing of private plans by 1980.
The preceding comparison of the principal indicators of growth pro-
jected for the 1960-80 period with the record of the preceding 20
years shows evidence of the maturing process. Possibly this compari-
son overstates the case. Holland took no account of retirement plans
for the self-employed and their employees, which may begin to grow
from a very small base over the next decade. Annual contributions
to such plans in excess of $1 billion might be attained by 1975.6
TABLE 11-5.-PAST AND PROTECTED GROWTH OF PRIVATE PENSION PLANS, 1940-80
1940 1945 1950 1955 1960 1965 1970 1975 1980
Covered employees (million)
Beneficiaries (million)
Benefit payments (billions of dollars)__
Ftjod assets (billions of dollars)
4. 1
.2
. 1
2. 4
6. 4
.3
. 2
5. 4
9. 8
.5
. 4
12. 0
15. 4
1.0
. 8
27. 4
21. 2
1.8
1. 8
52. 0
25. 4
2.8
3. 2
85. 4
32. 6
4. 1
4. 7
116. 9
37. 2
6.0
7. 3
155. 2
41. 4
8.0
10. 6
193. 1
Note: Because of instances of coverage under more than 1 plan, it is believed that coverage data for 1965 and previous
years may overstate the number of covered employees. See Walter W. Kolodrubetz, "Growth in Employee-Benebt Plans,
1950-65," Social Security Bulletin, April 1967, pp. 10-27.
Source: 1940-65, actual from Institute of Life Insurance, Social Security Administration, and author's estimates;
1970-80, from Hollasd's projections.
The projected slowdown in the rate of growth of coverage reflects
the fact that a high proportion of potentially eligible employees has
already been included. Of all employees on nonagricultural payrolls,
excluding government, 47.1 percent had been covered by the end of
1960, compared with only 25.0 percent at the end of 1950. If this group
of employees is further reduced to exclude part-time workers and
young workers (under age 25), the coverage ratio approximated 65
percent in 1960. High coverage already prevails in transportation,
public utilities, finance and insurance, mining, and many sectors of
manufacturing. In short, the large groups most readily covered by
group plans have already been included. Future extensions of coverage
will have to come increasingly from multiemployer and other plans
designed to accommodate small employers and small groups of
employees.7
In addition to this "saturation" factor, trends over the years in
pension arrangements have caused some shift of individuals from cov-
erage in active employment to beneficiary status. Increased disability
coverage, earlier vesting, greater use of the normal retirement age as a
mandatory retirement age, and considerable liberalization of the pro-
visions governing early retirement have been among the major devel-
opments.8 The effect has been to accelerate the maturing of these pro-
grams for individuals in private employment.
6 This estimate is based on studies made by. the author, for testimony before the Com-
mittee on Finance of the U.S. Senate, regarding the Self-Employed Individuals' Retire-
rnent Act. See Hearings on HR. 10. 86th Congress, 1st sess., June iS. 1959, pp. 210-
215 and Hearings on HR. 10, 87th Congress, 1st sess., July 28, 1961, pp. 144-102.
See Holland's projections, ch. 2.
8 See Bankers Trust Co., 1965 S~tiidy 0! IndustrIal Retircrnent Plans, for an analysis of
recent trends.
PAGENO="0065"
OLD AGE INCOME ASSURANCE-PART V 57
Since all employees covered under private plans are covered under
social security, the future role of OASDI will obviously affect these
supplementary arrangements. Flat-benefit plans, which provide a fixed
monthly pension for each year of credited service and which are most
commo.n among multiemployer and union plans, are not influenced by
increases in social security benefits, except as employees may lose inter-
est in bargaining for higher supplements. This has not occurred to date
because t.he supplementary benefits started at low levels. At some point,
of course, priority might be given to other elements of total
compensation.
In the more typical wage-and-service-related benefit formula char-
acteristic of single-employer plans, the social security benefit is taken
into account in determining an attainable objective as to the replace-
ment of some fraction of final compensation. This form of integration
with OASDI implies that a higher proportion of compensation above
the amount creditable for social security will be replaced by the sup-
plementary plan. The individual's salary progression and the trend
of wages and salaries generally will, therefore, have an important in-
fluence on the promised retirement income, especially with the trend
toward basing pensions on final average compensation.
The general assumption underlying Holland's projections is that
social security benefits will be increased in the future at about the same
pace as in the past. A shift from a "floor of protection" to a "level of
adequacy" concept in the determination of OASDI benefits could,
therefore, have a pronounced effect upon the projections used in this
study: benefit payments might still follow the pattern outlined in table
11-4, since they relate to conmiitments already made, while contribu-
tions and fund accumulations would not show the expected rate of
growth during the later years of the period.
There are other factors currently at work, however, which could
sustain a larger excess of contributions over benefits for a longer pe-
riod of years than Holland has projected. The most important single
influence in this direction might be an acceleration in the pace of
increases in money wages combined with the trend toward relating
benefits to final average compensation. In addition, the inclusion of
larger death, disability, survivor, and early retirement benefits in pen-
sion arrangements is causing increases in contribution rates. Even-
tually these trends have the effect of increasing benefit payments, of
course, but in the interim the pace of the maturing proceess is slowed.
The dynamic evolution of pension arrangements during the next
decade cannot be plotted with precision. However, the accuracy of
projections is not crucial to the analysis of economic effects. The direc-
tion and rates of change of income and saving flows are the important
elements, rather than the timing of their incidence.
THE PENSION STRUCTURE IN PERSPECTIVE
The preceding review of salient characteristics of the public and
private pension structure in being and in evolution suggests certain
summary observations:
1. We have come to a pause in the creation of new types of programs
and attention is being devoted primarily to refining, strengthening,
and generally improving their operations.
83-200-67--pt. ~-5
PAGENO="0066"
58 OLD AGE INCOME ASSURANCE-PART V
2. During the next decade, the flow of annual benefit payments
could increase by at least 60 percent, just from existing programs. A
100-percent increase is more likely. In the latter event, benefit pay-
ments would represent perhaps 7.6 percent of persona.l income, com-
pared with 6.4 percent in 1965.
3. The irreversible nature of retirement income commitments run-
ning far into the future means that we are issuing claims on goods
and services to be produced in the future at a rate substantially in
excess of the prospective rate of growth in real output.
4. The full implications of this structure of costs and benefits for
the distribution of income, consumption and saving, and the func-
tioning of the capital markets need further exploration. (These are
the topics of succeeding chapters.)
Systematic efforts, on a group basis, to spread income over the latter
segment of the lifespan, instead of merely over the working years,
obviously call for a complicated machinery if they are to be broadly
inclusive of a mobile, diverse population, like that of the United States.
The coordination and integration of various programs have made
excellent progress at the level of specific benefit formulas. Can they be
related to broad economic objectives as well? Can we establish a more
effective framework for the analysis and appraisal of the economic
effects of the decisions we are about to make regarding the future
structure of pension arrangements?
III. P~xs~ox PROGRAMS. THE REDISTRIBUTION OF INCOME. AND SAVING
The preceding chapter's overview of the characteristics and financ-
ing of pension programs highlighted the past, present, and potential
future aggregate flows of income which the programs generate. We
examined the development of a network of arrailgements designed to
transfer income (goods and services in real terms) to individuals
no longer in the work force. This income is transferred from other in-
dividuals, of course, in a. variety of ways. This process and its possible
economic effects are the subjec.t of this chapter.
PUBLIC PROGRAMS
The Primary Role of OASDI
As we have seen, the major factor in the current and future col-
lection of receipts and distribution of benefits is the Old-Age, Sur-
vivors, and Disability Insurance System. A possible effect of creating
these expectations of old-age and survivor benefits is that individuals
would save less in other forms. That is, they would substitute. the ac-
cumulation of claims against the system for other financial or tangible
assets. Because coverage has become so close to universal, it is not
possible to test this possibility by comparing the saving behavior of
large groups of people not covered by OASDI with those who are
covered.
In cha.pter IV, below, we deal systematically with the question of
the impact of other accumulations of pension claims on aggregate sav-
ing. This is done by examining the saving behavior of households
similar, in most characteristics, except coverage under a plan which
supplements OASDI. In the absence of a similar approach to deter-
PAGENO="0067"
OLD AGE INCOME ASSURANCE-PART V 59
mining the effect of OASDI coverage on saving in other forms, we
turn to the reasoning developed, by Cagan and Katona as discussed in
chapter IV.
In essence, their hypothesis is that the promise of some minimum
level of income in old age stimulates motivation to supplement it in
other forms. These forms have included supplemental pension plans,
life insurance, and liquid-asset holdings. It is common practice, for
example, for life insurance agents to provide information about
OASDI benefits to prospective purchasers of their policies. This ap-
proach succeeds because the agent can present a program for 1mprov-
ing the adequacy of financial security to be built upon a base which
already exists. This is similar to the experience with life insurance
provided to members of the Armed Forces; on balance, it has encour-
aged the sale of policies by agency companies.
The evidence summarized in chapter IV supports the view that the
creation of equities in the OASDI system has not served as a sub-
stitute for saving in other forms. Widely expressed fears that indi-
viduals would lose their incentive to save if this comprehensive pro-
tection were provided have proved to be without foundation. However,
the limitations of the body of evidence on saving behavior should be
kept in mind. To be more precise in describing experience since the
adoption of OASDI, we should state the conclusion in these terms:
tection against the hazards of old age, disability, and death, the rec-
ord shows that it has been a stimulus to other individual and group
programs to serve the same ends.
2. The increases in OASDI benefits have not kept pace with im-
provements in real living standards. To date, the program has not
provided an increasing share of total old-age income objectives.
3. We have no body of assembled evidence, `there~fore, as t.o what
would be the effect upon saving in other forms of achieving a goal of
"adequacy" as it might be defined in relation to some widely accepted
concept of living standards.
Robert M. Bail, the Commissioner of Social Security, urges that we
consider social security rather as a retirement system than as an anti-
poverty program, pointing out that it is the only retirement system for
80 percent of the beneficiaries. He argues that a test of adequacy is
relevant and timely.1 Similarly, the President's 1967 budget message.
discusses proposed changes in benefits in terms of adequacy.
From what we know about the saving function of individuals, it is
impossible to specify the level at which prospective benefits cease to
stimulate saving in other forms and begin to inhibit it. The type of
analysis made by Cagan is a contribution to the search for reliable
evidence, but does not suggest answers which would necessarily be ap-
plicable to a program of the size and scope of OAS1DI. Revisions of a
universally applicable system can be expected to change the attitudes
and expectations of a whole society.
Nevertheless, it is essential that this question be studied so that, if
decisions to move toward a concept of adequacy are made, their con-
sequences for saving in the economy may be taken into account. This is
not to say that economic factors should control. After all, fiscal policy
1 Robert RI. Ball, "Policy Issues in Social Security," ~S'ociaZ ~eouritsj Bulletin, June 1966,.
`)p. 3-9.
PAGENO="0068"
60 OLD AGE INCOME ASSURANCE-PART V
could be modified to offset the effects of changes in OASDI if the p05-
sible effects were known and recognized. Changing the balance of sav-
ing and investment in the economy without analysis of the long-range
consequences will not contribute effectively to the attainment of living
standard goals for citizens of all ages and circumstances.
The possible effects on saving habits of changing retirement income
expectations is, however, only one aspect of the operation of OASIDI.
It ma.y be the most significant aspect, but it relates only to the matter
of saving motivation. The other major economic aspect of the system is
its role as a collector of taxes and as a. simultaneous distributor of bene-
fits. The growing volume of transfer pa.yments can have important
economic effects on a current basis.
Early in our research program, John J. Carroll undertook a thor-
ough exploration of the possible effects of the whole range of tax-
support.ed programs.2 Such analysis depends upon a host of assump-
tions about the shifting and incidence of taxes, the consumption func-
tion, and consequences of changes in the distribution of income.
For example, Carroll assumed that the employee's share of the pay-
roll tax is not shifted at all, but that t.he employer's share is shifted to
the extent of two-thirds forward to the consumer and one-third back-
ward to the employee.3 For the benchmark year 1957, families with in-
comes below $4,000 were estimated to have contributed about 25 percent
of the payroll taxes and to have received about 75 percent of the bene-
fits. The net ga.in to the lower income groups represented about 52 per-
cent of the benefit payments. The t.ransfers were largely from the
middle-income groups.
There figures are illustrative of the power of the OASDI system in
redistributing income, whether or not they measure the magnitude
precisely. If we make the usual assumption that the marginal propen-
sity to consume of families receiving benefits is materially higher than
that of those bea.ring the costs of the program, we can conclude that
the result is in the direction of an increase in consumption and a reduc-
tion in saving. The calculation of transfers between income brackets
is obviously more reliable than estimates of the impact on consumption
and saving. In the latter case, we have to estimate family expenditure
patterns, not just their income levels.
Carroll's analysis uncovered many ga.ps in the availability of reliable
current information. Some of these gaps are being filled by the Social
Security Administration and other agencies. It is not altogethe.r
visionary to expect that in the future we shall be able to consider revi-
sions in the OASDI tax and benefit structure from three points of
view: (1) The level of benefits in t.he light of changing living costs
and standards, (2) t.he actuarial balance between receipts and bene-
fits over the visible future, and (3) the effects of the projected redis-
tribut.ion of income on consumption and saving.
The same type of analysis is also relevant to choices between in-
creases in the tax rate and increases in the maximum earnings base
2 The methodology employed in dealing with OASDI was developed in his monograph
Alternative Methods of Financing 01(1-Age, yarvivorS, and Disability Ifl8UraflCe, Michigan
Governmental Studies No. 38, Institute of Public AdminIstration, 1960. The results of
th~ National Bureau analysis were presented to the Advisory Committee for comment in
the form of "Some Notes on the Redistributional Effects of Mandated Pension Programs"
and are the basis for the relevant analysis in this chapter.
2 The effects of alternate assumptions on the shifting and incidence of the payroll tax
were also calculated.
PAGENO="0069"
OLD AGE INCOME ASSURANCE-PART V 61
to which it is applied. There are a number of considerations in making
such choices other than the extent of income redistribution involved,
yet, surely, this is relevant to the comprehensive design of the overall
tax and expenditure policies of the Federal Government.4
Apart from the redistribution of income between income classes,
social security transfers claims for goods and services from one age
group to another. The productively employed meet the cost of belle-
fits to those no longer at work. When the level of benefits is improved
and taxes increased, the young worker may be paying more than the
cost of his prospective benefits. Some students of the system argue
that this will ultimately create clissatisf action on the part. of those most
important in raising productivity. Of course, the young workers may
rest content with the notion that their turn to enjoy improved benefits
will come eventually. Oi' they may be amenable to an erosion of the
real value of old-age income promises by inflation as a means of
lightening the impact of the transfers, While such questions undoubt-
edly have some substance, it is by no means clear that. these are addi-
tional considerations not already taken into account in the spending
and saving decisions of the individuals concerned.
Other Contributory Government Pro grams
The other major contributory governmental programs are Railroad
Retirement, Federal Civil Service, and State and local government re-
tirement systems. Compared with OASDI, Carroll found that for the
benchmark year 1957 a smaller fraction of contributions to these pro-
grams (9 to 13 percent) was attributable to the lower income groups
and a somewhat smaller proportion of the benefits paid out (GO per-
cent). The actual net transfers to families with incomes less than $4,000
were very modest, however, because of the low level of benefit payments
relative to contributions in the Federa.l, State, and local government
employee systems.
Because of the surplus from their current operations, the Federal
Civil Service and State and local government retirement. systems pro-
vided a. net increase in saving after giving effect to the stimulus to con-
sumption of the transfers to lower income families with higher mar-
ginal propensities to consume. In these tax-supported programs, we
can see the effects upon saving and the capital markets of the funding
process at work simultaneously with the redistributing of income.
N on contributory Government Programs
Of the two major noncontributory programs supported by general
revenue taxat,~on, veterans' benefits and old-age assistance, the latter is
the more effective in the redistribution of income. `With tax collections
distributed across all income groups and benefits allocated to the lowest
income families, it is evident that this program will stimulate con-
sumption, precisely as it is intended to do.
Veterans' benefits are heavily concentrated among lower income
families, but a portion of the service-connected disability payments go
to middle and upper middle income families. Carroll found reliable
data particularly difficult to assemble in this section of his analysis and
his estimates are, therefore, tentative.
The long-range research program of the Social Security Administration includes most
of the topics mentioned here,
PAGENO="0070"
62 OLD AGE INCOME ASSURANCE-PART V
For the benchmark year 1957. these two noncontributory programs
generated net transfers to families with less than $4,000 incomes on a
substantial scale. The gain to these lower income groups was almost as
great as that of OASDI, despite the fact that benefit payments were
only 60 percent as large. Shifting individuals from old-age assistance
to the social security benefit rolls or making the corresponding sub-
stitution of benefits, then, reduces the income redistribution effects and
marginally reduces the shift from saving to consumption. Because
these noncontributory programs are supported from general revenues,
a major share of the cost is borne by upper middle income and high
income groups.
Public Pro gram~s as a Whole
Looking at the gain-loss pattern of the six types of tax-supported
pension programs mentioned above, Carroll found that families with
incomes of less than $4,000 received about $8 billion in 1957 in the form
of transfers from families whose incomes exceeded that level. Of this
figure, S5.3 billion went to families with incomes under $2,000 per year.
The estimated resulting net increase in consumption expenditures was
$1.3 billion.5
These figures are cited merely to provide an indication of the direc-
tion and magnitude of income redistribution effects. The growth in the
economy a.nd in the scale of the. tax-supported programs requires a
complete reexamination of the whole subject. Subsequent research and
the development of bet.ter data about beneficiaries should make possible
greater precision in the definition of current alterations in the pattern
of income distribution, spending, and saving.
Much more important, however, is the need for developing the data
and related models to permit. a simulation of the future economic effects
of alternative decisions regarding the financing and benefit structure
of tax-supported pension programs. Carroll's pilot study of redis-
tributive effects suggests that in the future the growing level of public-
plan operations, apart from State and local government retirement
systems, will generate a substantial offset to that saving which is done
through private plans. This is especially true of the Federal noncon-
tributory programs. Rational decisionmaking in this area requires
that we develop the data and techniques to verify and measure these
possible developments on a prospective basis.
PRIVATE PROGRAMS
The Saving Aspect
The growth of pension programs for individuals in private em-
ployment has been so spectacular since 1940 that attention has been
concentrated primarily on the growth in coverage and fund accumula-
tions, a veritable explosion in pension expectations. The response of
individuals, as will be shown in chapter IV, has been to add their
growing equities in private pension plans to saving in other forms.
Taking net fund accumulations exclusive of capital gains as the
These figures for 1iTh7 have little current sIgnificance. hut they provide some measure
of the Impact on saving of these programs at a specific point In time. The results are
not alarming, but the direction of the effects is clear.
PAGENO="0071"
OLD AGE INCOME ASSURANCE-PART V 63
measure of saving through private pension plans, the past. and pro-
jected trend in this form of saving is shown below.
Addition to Fund Assets
Year Billions
1950 $1. 8
1955 3. 6
1960 5. 3
1965 7. 6
1970 7. 5
1975 7. 6
1980 7. 4
Source: Institute of Life Insurance, Securities and Exchange Commission, and Hol-
land's projections for 1970-SO.
These figures are designed to reflect changes in book values without
taking account of realized and unrealized capital gains on common
stocks and other investments with equity characteristics. They do not
recognize, therefore, the addition to assets represented by the equity
in corporate retained savings. This is another way of saying that the
earnings on fund accumulations are substantially understated. In
chapter V, we attempt to recognize this factor, which became of ma-
jor significance after 1955, in projecting possible future asset accumula-
tions.
If full recognition were given to appreciation in equity investments
at an arbitrary 5-percent rate, the changes in fund accumulations at
market values might be more like $9, $10, $13, and $15 billion for 1965,
1970, 1975, and 1980. However, if these high returns are realized, there
n-iight well be a reduction in contributions because of the earlier
achievement of substantially full funding, an increase of benefits, or a
combination of the two. In that event, these higher projected levels of
saving through private plans based on the recognition of market
values would not be realized.
On balance, the trends suggest that the contribution to aggregate
saving of these programs for individuals in private employment will
be reaching a peak in another decade or so. The implications for capi-
tal formation and the capital markets are discussed in chapter V.
Transfers: The Contribution Aspect
A measure of the dynamic growth in private arrangements is the
increase in contributions. The record of the recent past and Holland's
projections are as follows:
Contributions
Year Billions
1950 $2. 1
1955 3. 8
1960 5 5
1965 ~`. 8
1970 8. 1
1975 9. 3
1980 10. 3
Despite the growth of collectively bargained plans, which are rarely
contributory, and a gradual reduction in the proportion of conven-
tional plans requiring employee contributions, the fraction of total
contributions provided by the employer has apparently remained
PAGENO="0072"
64 OLD AGE INCOME ASSURANCE-PART V
stable at 86 percent for the past decade. One dollar of contributions
in eight, therefore, represents a form of individual saving by the
employee directly analogous to saving through life insurance or any
other medium.
The interest earnings of the funds represent a return for currently
setting aside the discounted cost of benefits payable in the future.
These earnings are presumably shared by the employee and the em-
ployer approximately in proportion to their contributions. The fact
that the earnings are not taxed gives them full value.
It is more difficult to determine the cost incidence of the employer
contribution. It is a part of total employment costs which may be re-
covered in part by the individual firm if (1) the plan serves to attract
more productive employees; (2) it reduces turnover, training ex-
penses, accidents, and spoilage; and (3) it provides for orderly retire-
ment and promotions as a part of more efficient personnel management.
If the employer cost merely substitutes for other forms of compensa-
tion, it has been shifted backward to the employee. The incidence is not
necessarily either equal or proportionate among individual employees.
This shift can ta.ke the form of a slower increase in money wages rela-
tive to productivity gains. Collective bargaining over a total pay pack-
age appears to accept this concept, with pension and other fringe
benefits substituted for or added to pay increases in cents per hour.
The well-paid and highly paid employee with long service may wel-
come this bargaining emphasis. The employer's larger contribution
for a more liberal pension benefit is not taxed to the employee currently
as income; the tax impact is deferred until retirement, when his effec-
tive tax rate will be significantly lower. Within limits, the marginal
dollar of employer contribution to pension benefits is worth more to
the employee than the incremental dollar of money wages added to his
own retirement savings.
A different situation may exist, however, when a pattern plan is
negotiated throughout a major industry without regard to productiv-
ity gains. Depending upon product demand elasticities, a price rise to
recover this added cost may cause a decline in output and employment
or a shift from the utilization of labor to other factors of production.
The net effects nmv be extremely difficult to trace. They obviously
depend upon the economic environment and other influences at work
in the economy at the time.6
It appea.rs that the shifting and incidence of the employer's cost will
vary with the characteristics of the industry and the firm within it.
There is no clear answer to the question but only a limitless array of
specific cases. Perhaps the best general conclusion we can reach is that
the net cost is probably shifted both forward and backward, with a
longrun reduction in corporate profits, and, therefore, corporate sav-
ing, taking place only in marginal instances.
Another "contribution" to private pension programs is the favorable
income tax treatment given to employer contributions and fund earn-
For a searching discussion of the various possIbilities, see Challis A. Hall, Jr., "Retire-
ment Contributions, the Spending Stream, and Growth." Federal Ta~r Policy for Eco-
noinic Growth and Stability, papers submitted by panelists appearing before the Subcom-
mittee on Tax Policy, of the Joint Committee on the Economic Report, November 1955.
pp. 7SG-797. I am indebted to Dr. Thomas E. white for his analysis of the elements of
this problem during his participation in the pension research project.
PAGENO="0073"
OLD AGE INCOME ASSURANCE-PART V 65
ings. The Office of Tax Analysis of the U.S. Treasury Department
made the estimates for the year 1963 shown below.
Millions
Revenue gain from taxation of benefits received in excess of employee
contributions after the retirement income credit +$325
Revenue loss from exemption of interest earnings of pension funds at
individual tax rates -550
Revenue loss from not taxing to employees employer contributions
as made ~ iso
Net revenue loss -1, 375
Source: Private Pension Plans, hearings before the Subcommitee on Fiscal Policy of the
Joint Economic Committee, 89th Cong., 2d sess., p. 2, May 1966, p. 416.
Other factors equal, this treatment of qualified pension plans will
increase the revenues to be collected from taxpayers in general. Pre-
sumably, those not covered by pension arrangements share in the costs
without sharing in the benefits. These tax benefits are, of course,
equally applicable to the retirement systems for Government em-
ployees. Also, in the case of OASDI, all of the benefit payments are
exempt from taxation, with the result that there is no deferred recov-
ery of income taxes from the recipients.7
Transfers; The Benefit Aspect
The most dynamic phase of the growth in private programs is the
current and prospective increase in the flow of benefit payments. The
record of the recent past and Holland's projections show this trend:
Benefits
Year Billions
1950 ~o. 4
1955 . 8
i~60 1. 8
1965 3. 2
1970 4. 7
1975 7. 3
1980 io. o
The bulk of private pension benefits, almost by definition, go to
individuals who have had regular employment for an extended period
at good levels of compensation. Also, tile benefit structure is moderately
progressive (i.e., providing a higher proportion of higher compensa-
tion) to counterbalance the reverse pattern of OASI. A recent large
sample of industrial plans shows tile typical range for the growing
proportion of final pay plans:
Average Annval Compensation During Median Benefit Range Ecoclusive of
Final Years of Service OASI
Dollars Percent
4,500 20-30
8,000 31-35
15,000 31-35
35,000 36-40
Source: Bankers Trust Company, 1965 Study of Industrial Retirement Plans.
Since we are uncertain about tile incidence of contributions, it is not
feasible to estimate a redistribution of income between family income
A possible change In this treatment of benefits Is under discussion. An anomaly in the
present method of tax treatment Is the increase In the value of the benefits to an individual
with substantial other income in retirement.
PAGENO="0074"
66 OLD AGE INCOME ASSTJRAXCE-PART V
classes as a result of transfers through private pension programs. lYe
may suspect that the redistributive effects are less than in the case of
S tate and local government retirement systems, but how much less is
open to question.
Some insight into the distribution of beneficiaries by income class can
be gleaned from the Statistics of Income. Federal income tax returns
for 1964, on which at least one taxpayer was 65 or older, show the
following reported income from pensions and annuities (taxable
pGrtion)
Adjusted gross income classes
Number o
-
Thousands
f returns
Percent of
total
Taxable in
~
Millions of
dollars~
come reported
Percent of
~~total
Under$5,000
$5,000 to $9,999
$10,000 or more
918
260
140
69.7
19.7
10.6
1,369
581
444
57.2
24. 3
18.5
Total
1,318
100.0
2,394
100.0
Source: Internal Revenue Service, `Statistics of Income, 1964, Individual Income Tax Returns," publication 79, jan-
uary 1967, p.90.
Presumably, a large number of individuals had small benefits but
were not required to file returns.
These data do not distinguish the source of the pension or annuity
income, but they suggest that the benefits in a recent year went pre-
dominantly to middle-income family units. Private plans could not
have had a materially different pattern from the public programs also
included.
If employer contributions are largely shifted, and at least in part to
employees, it would appear that the effects of transfers between income
brackets work moderately in the direction of increasing consumption,
but at least for the present the impact c.annot be substantial.
PENSION PROGRAMS AS A WHOLE
Our exploration of the redistributive effects of the whole range of
public and private pension programs has produced only tentative con-
clusions. It has uncovered gaps in information which our resources
could not hope to remedy. The fruitfulness of further research and
information gathering about family incomes and budgets is evident.
The intelligent appraisal .of alternative programs requires that this
type of research be pursued with the aid of improved information
and techniques of analysis.
The present and prospective transfers of income between age groups
are more easily measured. In a sense, they represent the degree of suc-
cess which has been achieved by the pension structure. For these claims
to be honored in full in real terms, individuals must surrender a portion
of current output either in exchange for promises of benefits in the
future or in payment of tax assessments. That is to say, workers during
their active years, on the whole, defer consumption to their years of
retirement. In programs covering large groups, it is almost certain that
the present value of the future benefits to individual participants will
be somewhat greater or less than the consu1nption they forgo in the
PAGENO="0075"
OLD AGE INCOME ASSURANCE-PART V 67
present. Such individual gains and losses will be minimized, however,
if the benefit formula is equitable and the plan has been in effect for
an extended period.
Tinder a. pension plan in the design of which individuals have, had
an opportunity to express their views, that is, a voluntary or negotiated
plan, this kind of arrangement appears to involve no special burden
on the individual participant. While the saving feat.ure may be com-
pulsory and the utility of retirement income may vary between in-
divicluals depending upon their preferences and expectations, the
burden of participation in a plan characterized by equity and uni-
formity is not material.8
Wage- and service-related benefit plans covering individuals in pub-
lic and private employment, therefore, do not involve significant
burdens.
Tax-supported old-age benefits, on the other hand, involve the
burdens characteristic of any element of the revenue gathering and
distributing structure. The transfers between incon'ie brackets which
social insurance is designed to make set up pattern of gains to those
who earned less and losses to those who earned more during their years
of active employment. But, this is no special characteristic of pension
arrangements; it is common to almost all segments of the tax structure
which involve the redistribution of income among income classes and
age groups.
There is evidence to the effect that a program involving greater
income redistribution creates greater resistance to that program on
the part of those bearing the burden of the transfers. In a study of
social security systems around the world, Henry Aaron found that
countries relying more heavily on general revenues tend to spend less
on social security in proportion to national income than countries
relying more heavily on payroll taxes to finance benefits.9
The gain-loss patterns of social security, veterans' benefits, and old-
age assistance, however, cannot be considered apart from the whole
structure of transfers involved in the fiscal operations of governmental
units. The income redistribution effected by pension arrangements
must be examined in the context of the redistribution effected by the
personal income tax, excise taxes, or estate and gift taxation. More-
over, it should be noted that to some extent social security programs
merely replace voluntary and intrafamily transfers in favor of the
aged. The substitution of tax burdens for the individual's own per-
ception of his obligation to an aged or disabled person represents,
therefore, a change oniy in the form of the burden on the active
worker.
We conclude that the structure of public and private pensions
involves no special burdens which are not characteristic of any broad
scale of income transfers over time and across a population enjoying
a wide range of incomes and living standards.'° We also conclude that
8 ThIs conclusion does not Ignore the reality of possible Income transfers from younger
to older workers, from men to women, from transient to career employees, and among
Individuals in different Income classes.
Henry Aaron. "Social Security: International Comparisons," in Otto Ecksteln, ed.,
Studies in the Economics of Income Ma~ntenance, Washington, 1967, pp. 28-29.
1O For closely reasoned discussion of this question of burdens, see Daniel M. Holland,
"The Pension Structure," Federal Ecependiture Policy for Economic Growth and Stability,
Joint Economic Committee, U.S. Congress, Nov. 5, 1957, pp. 1007-1009.
PAGENO="0076"
68 OLD AGE INCOME ASSURANCE-PART V
less burden is involved for individuals in a collective sense when the
full cost of benefits is recognized currently and the excess of future
over current costs is funded according to a systematic program. The
reasoning runs as follows:
1. The excess of current contributions over current benefit pay-
ment; that is, pension saving, finances capital formation. As a result,
productivity gains are greater than would otherwise be the case.
2. As the pension system matures, or comes closer to maturity, the
increasing transfers of output to beneficiaries are made from a larger
total output. The fraction of the current output transferred from
active to retired workers is, therefore, a smaller and less burdensome
one.
3. If a cohort of workers over their active working years saves the
capital accumulation and the earnings on it to enlarge total output
sufficiently to generate the goods and services which will satisfy their
claims in retirement, it has placed minimal burdens on other workers.
However, this line or reasoning is based on a series of assumptions
which are probably not equally valid. In the first place, it is assumed
that investment will expand to absorb the full addition to savings at
full-employment levels of output. A study by Simon Kuznets supports
this assumption."
Second, it is assumed that benefits are in a fixed relationship to past
savings. This is surely not entirely valid. In fact, we know that eco-
iiornic growth and rising living standards will be accompanied by im-
provements in the level of benefits, involving a supplemental trans-
fer of current output from t.he working to the retired group. However,
if inflation erodes the real value of money claims, the price rise will par-
tially offset this supplemental transfer process.
In summary, the maturing process never ends in a growing economy.
There is a rising level of aspirations which results in transfers to re-
tired workers in excess of what they have contributed through pension
saving. Our reasoning, therefore, is that the burden of pension bene-
fits is lessened when they are provided at least in part from additions
to saving which has been employed in incremental capital formation
to expand real output. This should be true on the average for indi-
vidual burdens and for the burden on productivity of the economy as a
whole. One of the conditions promoting this result is a minimal effect
on the incentives of those engaged in production.
Pay-as-you-go arrangements and those which use the taxing power
of Government to make transfers of income are, in this sense, the most
bllrdensome pension program since they entail no contribution to capi-
tal formation.12 We should be cautious about reaching fixed conclusions
on these matters. Within broad limits, the tolerance levels of indi-
viduals to accepting burdens and limitations of their expenditure deC-
cisions are not fixed for all time. We know that people can and do ad-
lust to all kinds of circumstances, including burdens. This process takes
11 Capitni in the American Economy: Its Formation and Financing, Princeton university
Press for NBER. iDGi.
1~ Pay-as-you-go plans of industrial organizations or of Government, of course, need not
be more burdensome in every case. Contributions not made to fund future costs may he
directly invested to produce future productivity gains from which rising benefit payments
can he met. The after-tax return from these direct investments must be persistenly high
relative to the cost of capital from other sources to make pay-as-you-go programs less
burdensome than those which are systematically funded.
PAGENO="0077"
OLD AGE INCOME ASSURANCE-PART V 69
place slowly and is greatly aided by a sense of equity in the division of
the cost incurred. The role of economic analysis is to measure, more
effectively than in the past, the incidence of cost burdens and the distri-
bution of benefits. If this is done well, the choices and decisions on
priorities among programs to enhance the level of living will be made
not by prototypes of the economic man but by an electorate better in-
formed of the economic consequences of efforts to meet desired social
objectives.
In broad perspective, the limitations on what the American economy
can "afford" to provide in the way of income maintenance to older
members of the population are costs (burdens of transfers) incurred
and the conflicting claims of other high-priority objectives for its
dependent members; that is, education of dependent children competes
for resources against better living standards for dependent senior
citizens. The limitations are real resources and the willingness of indi-
viduals to share gains in real income. The Federal Government's tax-
supported programs are essental to deal with the large number of cases
in which a stable employment relationship does not exist. On balance,
they are no more burdensome per dollar of expenthture, and perhaps
less so in the case of OASDI, than a wide range of public service
activities.
Funded employee retirement plans, whether public or private,'3 in-
volve less burden on the incentives of those actively engaged in pro-
duction. Through the saving and investment process, they finance the
economic progress which produces higher retirement living standards.
This is presumably the rationale for deferring the incidence of in-
come taxation on growth in the individual `s pension equity. In essence,
it is also an answer to the question: What difference does it make
whether now and in the future a greater proportion of retirement in-
comes is provided under private auspices or under tax-supported
governmental programs?
IV. THE EFFECT or PENSIONS ON AGGREGATE SAVING
Nonfarm households are the principal group of savers, typically
accounting for about three-fifths of gross national saving in the Amen-
can economy.' Logically, then, we should give primary attention to
the question of how retirement income programs affect the saving of
individual family units, commonly referred to as personal saving to
distinguish it from saving by corporate business, unincorporated busi-
nesses and farms, and units of Government.
THE ISSUES INVOLVED
From the national income and product accounts prepared by the
15.5. Department of Commerce, the picture below emerges of personal
income, personal saving, and retirement saving. However measured,
retirement saving represents a maj or share of personal saving and its
importance has increased during the first two postwar decades. By all
13 In this context, there is no distinction between the retirement systems of State and
local governments and the plans for Individuals in private employment, since both generate
private saving and the financing of activities which enlarge the output of goods and services.
`Based upon the flow-of-funds accounts of the Board of Governors of the Federal Reserve
System. Households provide an even larger proportioa of net saving in the economy.
PAGENO="0078"
70 OLD AGE INCOME ASSURANCE-PART V
past standards, the systematic provision of retirement income by fund-
ing through financial institutions, as distinguished from transfer pay-
ments, has become the most important new element in the saving
process.
Average,
1946-50
Average,
1951-55
Average,
1956-60
Average,
1961-55
Personal income (billions of dollars)'
Personal saving (billions of dollars)'
Retirement saving (bi!iisr.s of dollars
Personal saving as percentage of personal income
Retirement saving as percentage of persona! income
Retirement ss~ing as psrcentoge at personal saving
203.0
11.7
2.7
5. 8
1. 3
23. 1
283.4
17.2
4.6
6. 1
1. 6
26.7
366.0
19.9
7.2
5. 4
2. 0
36.2
472.0
23.2
9.9
4. 9
2. 1
42. 7
I National income accoants of the U.S. Department of Commerce, Survoy of Current Business."
2 Flow-of-funds accounts of the Boord of Governors of the Federal Reserve System. Measures increase in assets held
by governmestsl and private pension prcgrams, excluding OASDI.
Note: If persons! caving is expressed as a percsstoge of disposable personal income, the saving ratios for the four
5-yesr periods are 6.4, 6.9, 6.2, and 5.6 percent, respectively. Using the ratio to personal income is constdared mere rele-
vant since most of the pension saving is, in effect, from pretax income; i.e., employer costributioss to public and private
plans are cot recognized as taxable incsme to the employee.
Whether this form of saving represents a substitution for other
forms or a net addition to the aggregate saving of households becomes
a key question in determining the future volume, composition, and
trend of saving and capital formation in the American economy. The
research of Raymond W. Goldsmith 2 has established the long-term
stability in the ratio of saving to income. Despite the development of
new financial institutions such as life insurance companies, savings and
loan associations, consumer finance companies, investment companies,
savings banks, and credit unions, the record of the past 90 years
shows remarkable stability in the saving ratio.
There have, of course, been great changes in the preferences of in-
dividuals for different types of financial assets. Financial interme-
diaries have enjoyed varying rates of growth as the public has shifted
its favor from one to another form of saving medium. The secular
trend toward the institutionalization of saving has afforded ample
opportunities for growth in the case of most intermediaries despite
the changing emphasis of individuals on different characteristics of
saving media.3
It is quite consistent with the record of the past, therefore, to expect
that households will substitute for other savings the accumulation of
an interest or equity in a pension program. With this form of income
for life after retirement assured by coverage under a public or pri-
vate plan, it would appear to be less necessary or desirable to save
during working years by means of savings accounts, life insurance,
or homeownership. If this is the way people react, we should expect
them to substitute retirement saving for saving in other forms. The
substitution might well be less than complete, however, because of
certain characteristics of saving through a pension plan: (1) The
realization of full pension benefits may depend upon continuity of
employment in the event of delayed or graded vesting. (2) An equity
CA ~Stndy of S'aving `in the United ~tetes, Princeton. N.J., 1955. vol. I, pp. 4-22.
a Exceptions are the postal savings system (discontinued in 1966) and possibly U.S.
savings bonds. For a comprehensive view of these trends, see Raymond W. Goldsmith, Pi-
1IU.I1CIIII Tntcrnlttharies iii the Aineuicait EConomy t~iItC 1900, PrInceton University Presi
for NBER, 1958.
PAGENO="0079"
OLD AGE INCOME ASSURANCE-PART V 71
in a pension plan is illiquid. It cannot be drawn down as can other
forms of saving in the interim period prior to retirement for other
purposes. (3) Contributions to retirement programs are usually com-
pulsory. As a result, the participant may have accumulated for his
benefit sums well in excess of the amounts which he would voluntarily
save in any form. Workers employed at lower pay scales might, in-
deed, have their pension equities as virtually their only form of saving.
A working hypothesis derived from general observation and experi-
ence would be that households covered under pension plans would save
more, other influences on the saving ratio held equal, than those not
covered. The substitution would be substantially less than complete
because of the special character of retirement savings and their lack
of availability to serve other desired objectives.4
THE CAGAN ANALYSIS
How, in fact, do households react in their saving patterns to cov-
erage under a pension program supplementary to social security? This
is the question Phillip Ca.gan sought to answer. Available resources
did not afford an opportunity to survey a representative sample of
American households, but the cooperation of a sizable group of Con-
sumers Union subscribers made possible a detailed analysis of their
saving behavior and holdings of financial assets. This group was
clearly not representative of the population at large because of higher
incomes, greater educational attainments, and the predominance of
salaried workers as heads of households.
The deficiency in representativeness of the sample is partly com-
pensated for by certain of its favorable characteristics. Consumers
Union subscribers may be presumed to be better informed about their
financial affairs, more thoughtful about their financial decisions, and
better equipped to respond to lengthy questionnaires. The 11,513
households actually used in the analysis also represent a very substan-
tial number of cases of middle- and upper-income households in which
saving is important. Over 59 percent had annual incomes between
$5,000 and $10,000, and almost 34 percent enjoyed a household income
of $10,000 and above in 1958-59.
It is plausible, although not demonstrable, that this more knowl-
edgeable group of households reflects the reaction to pension coverage
which will be characteristic of a much more representative sample
of American households in years to come as they become better in-
formed about their pension equities and as they see more of their
friends, relatives, and coworkers actually receiving benefits. If we
assume that employers and the representatives of employees will con-
tinue actively to promote better education of covered workers about
prospective benefits, as they have compelling reasons to do, one could
argue that the reactions of the select group of Consumers Union meni-
bers would be indicative of the saving behavior of the population at
large during a subsequent period.
For a discussion of possible ranges of substitution, see Daniel M. Holland, "What Can
We Expect From Pensions ?" Harvani Business Review, July-August 1959, pp. 125-140;
and Challis A. Hall, Jr., "Retirement Contributions, the Spending Stream and Growth,"
Federal Tace Policy for Economic Growth and Stability, Joint Committee on the Economic
Report, Washington, 1955, pp. 786-797.
PAGENO="0080"
72 OLD AGE INCOME ASSIJRANCE-PART V
However, it is not necessary to repeat here the close reasoning and
analysis which Cagan applied to his survey of households. Only certain
highlights of his study need to be introduced at this point. Excluding
households with significant gains and losses and extreme saving ratios,
Cagan found the average saving ratios (saving as a percentage of
income) to be as shown below.
AVERAGE SAVING RATIOS
[In percent]
Covered
households
Not-covered
househoids
Discretionary saving in the form of increases in cash and securities minus increases in
nonmortgage debt
Contractual saving in the form of increases in equities in life insurance, annuities, and
real estate
2.8
5.9
2.1
5.7
Subtotal
8.7
7.7
Pension saving in the form of increases in equities in pension plans
2.8
Total saving ratio
11.5
7.7
Source: Phillip Cagan, `The Effect of Pension Plans on Aggregate Saving," New York, NBER, 1965, t3ble 4, p. 21. The
number of households included is 8,0J8 for tee covered group and 2,911 for the not covered.
Although respondents to the survey clearly underestimated the in-
crease in their pension equities, presumably because of their lack of
information about employer contributions, the proposition that saving
in forms other than through pensions is not reduced in the average
household, and may even be increased, is clearly supported by Cagan's
findings. Furthermore, he shows that this pattern of behavior is not
restricted to either discretionary or contractual saving forms.5
The conclusion that pension saving represents a net addition to per-
sonal saving, i.e., that it is not substituted for other forms of saving,
is sufficiently startling in the light of the history of saving habits to
call for some explanation.6 The explanation which seems to fit the facts
best is that there is a "recognition effect" associated with pension cover-
age. That is to say, the head of a household recognizes that a reason-
able degree of financial independence in retirement is attainable for
him when a pension program is made applicable to him in addition
to his social security income prospects. At. this point, his motivation
to save on his own to increase the adequacy of his retirement income
is stimulated by the realization that such a goal is within his grasp.
The fact that it was not similarly attainable by his parents, who are
in some degree dependent upon him, may heighten the impact of this
recognition effect.
Conceivably, this stimulus to saving in other forms will weaken
over an employee's lifetime as his pension benefits become fully vested
and he has in fact saved to supplement his prospective benefits. Cagan's
study suggests that while the recognition effect predominates, there
does exist evidence of an offsetting substitution effect in the case of
those who contribute 5 percent or more of their compensation to the
pension plan and enjoy full vesting. The strongest recognition effect,
Tests of the validity of these conclusions and examinations of the data for possible
biases are set forth in Cagan's analysis and are not repeated here.
George Garvy wa~ among the brat economists to recognize this possibility and to suggest
reasons for it. See his "The Effect of Private Pension Plans on Personal Savings," Review
of Economics and Statistic8, August 1950, pp. 223-226.
PAGENO="0081"
OLD AGE INCOME ASSURANCE-PART V 73
on the other hand, is found in those cases where the individual con-
tributes at a modest rate and has only partially vested rights to his
pension. The recognition effect seems to be least powerful when the
household head does not contribute to the plan and has no vesting.
The argument that a contributory pension plan is more meaningful
to the employee than a noncontributory one is apparently supported
by these findings. Further, the transformation of a remote contingent
benefit prospect into an emerging firm promise as vesting is earned
seems to act as a catalyst in the individual's making of decisions to
save for his old age.
The operation of both recognition and substitution effects in pro-
ducing the average behavior of households in the Consumers Union
panel suggests that the impact of pension programs on aggregate per-
sonal saving can change over time if there are changes in the structure
of pension arrangements. This must be kept in mind when we turn
to the question of implications of Cagan's findings for the future.
CORROBORATION OF CAGAN'S FINDINGS
Writing in The Mass Cb'rtsumption Society, George Katona of the
Survey Research Center, Institute for Social Research of the Univer-
sity of Michigan, discusses the recognition effect in these words:
Being assured of some, for most people insufficient, funds after retirement, the
provision of adequate funds during old age no longer appears an insurmountably
difficult problem; being closer to the goal Stimulates people to work harder to
achieve the goal, and, therefore, collective retirement plans promote individual
saving.
This hypothesis is derived from the goal-gradient hypothesis which assumes
that effort is intensified the closer one is to one's goal.
GoaLs that are believed to be attainable stimulate us to a much greater extent
than improbable dreams.7
This, then, is a rationale based upon the study of human behavior
for the results obtained in Cagan's study.
The other major question raised about Cagan's conclusions relates
to the nature of his sample. Admittedly, it is debatable whether the re-
actions of .the particular group of Consumers Union respondents can
be accepted as valid for the American population as a whole. Fortu-
nately, important evidence on this question has been provided in a sub-
sequent study by Katona and his staff at the University of Michigan.
With the assistance of a research grant from the Social Security Ad-
ministration, Katona surveyed a representative sample of households
in the continental United States in 1962 and 1963. His findings coin-
pletely support the existence of the recognition effect of pension
coverage.8
Pension coverage, according to this survey, increases both savii~g
behavior and saving-mindedness. Those expecting an adequate retire-
ment -income, furthermore, save at least as much on average as those
who expect an income below their needs. The effects of rising levels of
aspiration, stimulated by accomplishment, in motivating saving are
seen as the powerful factors; fear of the future seems to provide much
less motivation for saving for the distant prospect of retirement.
`New York, 1964, pp. 184-485.
8 George Katona, Private Pen8ion8 and Individual Saving, Monograph No. 40, Survey
Research Center, Ann Arbor, MIch., 1965.
S3-200-----Gi-pt. 5-6
PAGENO="0082"
74 OLD AGE INCOME ASSIJRANCE~PART V
FUTURE TRENDS
The Ca~an study, like any cross-sectional analysis, has its maximum
significance in relation to the enviromnent in which it was made. The
similarity of results from Katona~s survey, 4 years later, might suggest
that we are dealing with fairly stable phenomena. But, we must bear
in mind the possibility of changes in personal saving patterns in re-
sponse to changes in the pension structure.
The trend to noncontributory plans should erode the substitution
effect which Cagan observed to be stronger when employee contribu-
tions are substantial. Earlier coverage for employees, together with
the prospect of sooner achieving vesting, might increase the recog-
nition effect. Greater knowledge of the meaning of pem~sion coverage
does not seem to have a discernible effect. upon the saving ratio. Thus,
we can identify some factors which might, on balance, modify saving
behavior. It would be hazardous to guess the net result, but it seems
unlikely that in the near future pension coverage will reduce saving in
other forms. This conclusion is supported by the evidence suggesting
that the recognition effect does not "wear off" with the passage of
years of covered employment.
Apart from these changes in the terms of pension a.rrangemnents, of
course, there is the possibility that if benefit formulas are liberalized,
some of the retirement saving motivation might be sapped. The ex-
treme case would be that in which larger social security benefits and
higher supplementary pension benefits would substantially equal the
reduced cost of maintaining the same standard of living as during the
last working years. This situation is not in prospect for the visible
future, but the movement in that direction persists, particularly in
reducing the cost of maintaining living standards by programs such
as Medicare and by the lower income ta.x applicable to retirement in-
come. Rising levels of aspirations, after all, can apply to standards
of living during the period before as well as the period after retire-
ment.
In any event, the value of repeating the Cagan and Katona surveys
within the next decade is apparent. With the further extension of
coverage and the maturing of the pension structure further advanced,
it should be fruitful to test the reactions of households to pension
coverage once again.
Putting to one side this rather speculative question of a possible
shift in the balance between the recognition and substitution effects of
pension coverage, What are some of the implications for personal say-
ing in the American economy of the prospective growth of pension ar-
rangements? Holland's lrojections show that benefit payments are
likely to be rising at a more rapid rate than contributions in the years
ahead. The rate of growth in retirement saving could slacken. The
implications are pictured below by comparing possible 1975 with ac-
tual 1965 relationships.
PAGENO="0083"
OLD AGE INCOME ASSURANCE-PART v 75
1965
1975 1
Personal income (billions of dollars)
Personal saving (billions of dollars)
Retirement saving (billions of dsllars)
Personal saving as percentage of personal income
Retirement saving as percentage of personal income
Retirement saving as percentage of personal saving
537. 8
27. 2
10. 9
5. 1
2. 0
40. 1
894. 9
48. 5
15. 1
5. 4
1. 7
31. 1
1 Personal income and personal saving from a representative projection of national income accounts made by
the Natiosal Industrial Conference Board in `The Economy in the Next Decade," "The Conference Board Record,''
December 1965. This projection is roughly in line with other projoctiens made by governmental and private
agencies. Retirement saving is calculated from Holland's projections of about $13.6 billion in 1975 for private
and State and local governmsnt programs, and it is assumed that, as in 1965, this will be nine-tenths of a
$15.1 billion total including the funding of Federal Government plans.
This exercise suggests that, unless Holland's projections materially
underestimate the rate of retirement saving in another decade, the
impact on aggregate personal saving is now at a peak and will dimin-
ish. That is to say, the fraction of personal saving represented by retire-
ment saving is currently close to its maximum. Unless there is a new
burst of growth in retirement saving during the years ahead, or in
other forms of saving, the projected saving ratio will not be realized.
No doubt, a portion of a possible decline in the saving ratio will dis-
appear if consumer dura'bles are included, since consumer expenditures
are expected to rise more rapidly in this area than in the case of services
or nondurable goods. Also we may be victims of measurement prob-
lems. Saving through expenditures on such intangible capital assets as
education (we could even add cultural and personal capital derived
from "investment" in travel, the arts, and good works) is not treated
as saving in conventional accounting. One of the consequences of gains
in living standards can simply be a shift to saving through the ac-
cumulation of such intangible assets.°
However, for our subsequent analysis of capital market influences,
we must concentrate on the implications of the unfolding pattern of
financial saving in the economy. In this dimension of the economic
effects of pensions, we conclude that the impact of rapid growth is close
to its peak; that in the absence of a resurgence of growth in pension
arrangements, we have already experienced the major part of their
influence on saving and the capital markets during the second postwar
decade.
POSSIBLE EFFECTS ON SAVING BY BUSINESS AND GOVERNMENT
On the basis of our earlier analysis in chapter III, it appears that
the growth of private pensions has probably not had any material
effect on saving `by business.
Federal, `State, and local government saving is substantial in the
absence of major wars and depressions. Changes in the net worth of
these units, taking into account tangible assets as well as financial
claims, are somewhat irregular and `sensitive to economic change.
With so many forces affecting the levels of revenues and expenditures,
it is difficult to isolate the effects of pension arrangements in specific
terms. At times, `of course, the question relates ,to changes not in saving
but in dissaving.
For the significance of these forms of saving, see F. Thomas Juster, Houselaolci Capital
Formatson and Financing, 1897-19 62, New York, NBER, 1966.
PAGENO="0084"
76 OLD AGE INCOME ASSURANCE-PART V
The revenues of the Federal Government, and other units of govern-
ment in certain instances, are lower than they would otherwise be as
a result of not taxing currently as income to the covered employee
those contributions which the employer makes to pension programs.
In 1965, such contributions exceeded $10 billion, of which two-thirds
were supplied by private and one-third by public employers. This
figure does not accurately reflect the. true situation because the Federal
civil service and some State systems are greatly underfunded; that is,
the employer's contribution substantially understates the growth in
the employee's equity in the program. The extreme case, of course,
is the retirement program for members of the armed services under
which benefits are paid as current budgetary expenses. The member
earns a right to retirement benefits, but that equity is not recognized
until he starts to draw them. The same is true of additional billions of
employer contributions to a wide range of health and insurance benefit
programs.
The figure which appears in the flow-of-funds accounts as retire-
ments saving is, we should constantly bear in mind, only the increase
in assets held in trust funds or as offsets to life insurance reserves. This
is not necessarily the same figure as the present value of pension
claims accrued during a year after discounting for mortality and
turnover factors. What the real figure is no one knows, but, particularly
in Public programs, there is a very substantial understatement. It is
clear, however, that we are measuring only financial claims which are
specifically recognized, rather than the growth of individuals' pension
equities.
Whatever the true magnitudes, there is a partial offset in that pen-
sion benefits received as a result of employer contributions are taxed
as income, with the exception of OASI benefits, which are, of course,
not ta.xed. The offset is reduced by the more favorable tax treatment
of persons over 65 and the fact that lower rates are applicable on the
average for the person no longer actively employed.
Thus we can say that the growth of public and private pensions has
eroded the revenue base for taxation based upon income. The simplest
illustration is the recent trend in some State and local government re-
tirement programs to take over a portion of the employee's contribution
instead of increasing salaries. The same result is achieved by increasing
benefits without changing the employee's contribution rate.
Another important tax exemption is that of the income earned by a
pension fund. The amount of this untaxed income has, of course, grown
rapidly as public and private systems have accumulated assets. Exclu-
sive of Federal programs, such untaxed earnings are about one-half
of the employer contributions.
Substantial but unmeasured reductions in Federal, State, and local
expenditures are made possible by the flow of pension benefits. We have
observed the partial displacement of old-age assistance by social se-
curity as an example. Given a target of well-being sought for the aged,
the existence of pension benefits may afford governmental units sub-
stantial economies in their necessary outlays for supporting housing
projects, other community facilities, and agencies.
The effect on Government saving is difficult to determine, as Cagan
has pointed out. It seems doubtful t.hat this one element in the tax
PAGENO="0085"
OLD AGE INCOME ASSURANCE-PART V
structure affects the level of expenditures by Government. If Federal
and State taxes on income are higher as a consequence, the principal
result may be a reduction in the level of other saving, particularly
personal saving by upper-income individuals. However, saving in the
aggregate is larger, we have seen, by the full amount of retirement
saving. Thus, a discernible result may be to make savings by different
income groups less unequal than would otherwise be the case. That is,
the lower-income groups may save more, and the higher-income groups
less.
If the pension structure is designed to shift income earned in pro-
ductive effort from the working years to the years of retirement, it is
not unreasonable to shift the incidence of taxation with it. If there were
no differential in applicable rates of tax, revenues would be unchanged
after the transitional period. However, the double personal exemption
for those over 65, the retirement income credit, and the exclusion of
OASI benefits from tax are all steps designed to lighten the tax burden
on pension benefits. Thus, employees in the aggregate, if we assume
some stable revenue requirement, pay higher income taxes on wages
and salaries during employment and lower taxes after retirement. In
this light, the public and private pension structure is a mammoth trans-
fer operation, probably without any important impact on Government
saving in either direction.
In another sense, also, the argument is admissible that the tax benefits
afforded systematic programs to provide retirement incomes supple-
mentary to OASI having nothing to do with Government saving. A
long-term decisioii is reached that encouragement should be given to
certain financial arrangements. Having made this decisioii, govern-
mental actions are taken to establish whatever pattern of revenue and
expenditure policies seems most likely to afford promise of sustained
economic growth and rising living standards.
TIlE IMPACT ON TOTAL NATIONAL SAVING
Our research has supported the proposition that pension saving is
a net addition to personal saving. Less clearly established, perhaps, is
the extension of this conclusion to state that it is a net addition to total
national saving. The impact on saving by business amid Government is
not~ clear, but it seems doubtful that it is materially affected.
There is also some evidence that this major impact has already been
felt. If it is desirable to sustain the growth of saving in the economy,
some other economic policies may be more fruitful in the future. On
balance, however, the tax and other policies which have fostered pen-
sion fund growth appear to have been appropriate, `in the environ-
ment of the postwar years, to the objective of increasing saving. If
balanced economic growth should require either a higher level of say-
ing or a higher level .of consumption in the future, it will be most
desirable to resurvey `this whole topic again.
It is evident, however, that the long-term nature of pension arrange-
ments makes them unsuitable as a focal point for short-term adjust-
ments in the pattern of spending and saving. Tax and expenditure
policies designed to foster stable economic growth can clearly `be more
PAGENO="0086"
78 OLD AGE INCOME ASSURANCE-PART V
effectively applied to areas which are more sensitive to temporary
influences, in. their application, however, recognition should be given
to the secular changes brought about by the maturing pension structure.
V. Tnr IMPACT OF Prxsioxs Ox THE CAPITAL Mxnhi~T: PRIVATE
FUNDS
The impact of pension fund operations on tile capital market depends
upon the investment policies followed by portfolio managers. Our
concern is ~rimariiy with differential effects. What difference does it
make when funds flow through these financial intermediaries, rather
than through other channels? This chapter provides some answers
to this question in the case of pension plans for employees of business
firms and nonprofit institutions. Governmental plans of all kinds are
dealt with in chapter VI.
INSURED PLANS
Growth Trends
When pension commitments are funded by premium payments to,
or deposits with, a. life insurance company, they are described as being
insured. The form of the contract is not material to our consideration.
Suffice it to say, the insurance company is seeking to provide as good
a return on the fund as possible within the statutes governing invest-
ments after allowance for adequate reserves for possible losses. The
actual process involves a. mingling of the funds with all others avail-
able for employment in tile capital market.'
Thus, we can think of insured pension funds as a. portion of life
insurance loans and investments. In general, it is valid to assume that
receipts are invested across tile whole range of assets, enhancing the
companies' ability to lend and invest but not changing it in any mate-
rial respect. Tile reserve liabilities of a life insurer under a group
annuity contract a.re different from those created by an equivalent
premium volume from, say, ordinary life insurance; but, it is not clear
that this fact affects the investment policies of the company. The
objective of earning a high, stable rate of return is similar, if not
identical, in both cases.
The growth of pension programs has added materially to the assets
of life insurance companies, as indicated in table V-i. At the end of
1965, assets balancing insured pension reserves of $27.3 billion were
equal to 17 percent of total assets of all life insurance companies in the
United States and 19 percent of invested assets (excluding cash,
policy loans, home office buildings, and premiums due). The propor-
tions have increased over the years because assets attributable to pen-
sion plans have grown more rapidly than assets employed in other areas
of the business. In 1950, both proportions were only 9 percent. Over
the period 1940-55. life insurance assets grew at a compound annual
rate of 7.4 percent, but the growth rate slowed down in the mid-1950's
and wa.s only 5.8 percent in 1955-65. This slackening occurred in spite
of a generally rising growth rate for life insurance in force. The latter
rate increased from 7.8 percent in 1940-55 to 9 percent in 1955-65. The
change in growth patterns is shown in table V-2.
1 Certain qualifications to this general statement are discussed below.
PAGENO="0087"
OLD AGE INCOME ASSURANCE-PART V 79
TABLE V-i--GROWTH IN LIFE INSURANCE ASSETS AND PENSION RESERVES, 1940-65
Yearend
Increase in
total assets
(billions)
Increase in
pension reserves
(billions)
(2) divided
by (1) 1
(percent)
(1)
(2)
(3)
1940
$1.6
$0.2
13
1945
3.7
.4
09
1950
4.4
.8
18
1955
5.9
1.3
22
1960
5.9
1.3
22
1961
7.2
1.4
19
1962
6.5
1.4
21
1963
7. 8
1. 7
21
1964
8.3
2.0
23
1965
9.4
2.1
22
Source: Computed from Institute of Life Insurance, "Private and Public Pension Plans in the United States" and
"Life Insurance Fact Book."
1 Computed before rounding.
The reason that assets increased less than insurance in force is
that the saving component in the contract mix fell. For example, the
proportion of endowment policies decreased from about 15 percent of
individual life insurance in force in 1950 to 7 percent in 1962.2 Group
insurance, meanwhile, increased its share of the total (exclusive of
credit life insurance) from 21 percent to 33 percent. Another view of
the same trend is visible from the flow-of-funds accounts of the Fed-
eral Reserve System. If the increase in life insurance reserves is ex-
pressed as a fraction of total financial assets acquired by households,
the proportion falls rather continuously from 21 percent in 1946, the
first year for which data are available, to 10 percent in 1964 and 1965.
TABLE V-2.-LIFE INSURANCE GROWTH RATES, 5-YEAR PERIODS,
1940-65
[In percentj
Insurance
in force 1
Total assets
Total assets
less pension
reserves
1940-45
5. 6
7. 8
7. 2
1945-50
8. 7
7. 4
6. 7
1950-55
8. 2
7. 2
6. 3
1955-60
9. 2
5. 7
5. 0
1960-65
8. 7
5. 8
5. 5
1 Excludes credit life insurance.
Some loss of market share in the saving field suffered by life insur-
ance companies has not been accepted with equanimity. To regain their
former position, managements are emphasizing the sale of ordinary
life contracts in intensified sales efforts~ But, more importantly, for
the particular concerns of this study, a strenuous effort is being made
to regain lost ground in the pension area, as more fully treated below.
Portfolio Composition
Although the life insurance industry has participated actively in
practically all segments of the capital markets, it has become the
dominant institutional lender in two areas: directly placed long-term
loans to business and mortgages on commercial and industrial prop-
2 Institute of Life Insurance.
PAGENO="0088"
80 OLD AGE INCOME ASSURANCE-PART V
erties. The distribution of invested assets at the end of 1965 and
increase in ownership during the preceding decade provide a summary
view of life insurance portfolio activity.
The data in table V-3 provide a picture of how some $16 billion
of insured pension fund accumulations have been invested during the
past decade along with other life insurance assets. The striking feature
of recent years has been the active financing of business firms through
directly placed term loans and mortgages on commercial and indus-
trial properties. Private pension arrangements have, then, made an
important contribution to the extensive financing of the high level of
business investment, and secondarily, to the expansion in the stock of
housing.°
TABLE V-3.-INVESTED ASSETS OF U.S. LIFE INSURANCE COMPANIES, 1965
.
Assets
Amount
(billions)
Percentage
of total
Net change
since 1955
(billioss)
Bands:
U.S. Government
Foreign government
State, provincial, and local
Railroad
1
.9
5.5
3.3
3.5
.6
3.8
2.3
5
.4
2.8
-.6
Public utility
Industrial and miscellaneous
17.0
38.3
11.8
26.4
3.1
20.2
70.2
48.4
22.4
Stacks:
Preferred
2.9
2.0
1.1
Common
6.3
4.3
4.4
9.1
6.3
5.5
Ł~srtgages:
Farm
4.8
3.3
2.6
1-to 4-family
Multifamily asd commercial
29.9
25. 3
20.6
17. 4
12.2
15. 8
60.0
41.4
30.6
Other:
Real estate hnld for investment
Other assets
3. 3
2.4
2. 3
1.7
1. 3
1.4
5.7 3.9
2.7
Total
145.0
100.0
61.2
Note: Excluding cash, policy loans, home office buildings, and premiums due.
Source: Computed from Institute of Life Insurance, "Life Insurance Fact Book," and Federal Home Loan Bank Board
Source Book.
Competitive Influences
The normal competition amon~ life insurance companies for superior
earnings on investments, which has been responsible for many innova-
tions in capital market instruments, has been reinforced by competi-
tion from other managers of pension fund accumulations. In 1950,
when insured pension reserves amounted to $5.6 billion, trusteed plans
had amassed $6.5 billion of assets. By the end of 1965, trusteed assets
of $58 billion were more than double the $27.3 billion of insured
pension reserves.
The flexibility of trusteed plans, especially in equity investment,
accounted for much of the loss in market share susta.ined by life insur-
0 For extensive treatment of the lending and investing activities of life Insiarance corn-
psniee, reference can be made to Life Insurance Association of America. Life Insurance
Companies u.s Financial Institutions. Englewood Cliffs, N.J., 1962. and Andrew F. Brim-
mer, Life Insurance Companies in the Capital Marlet, East Lansing, Mich., 1962.
PAGENO="0089"
OLD AGE INCOME ASSURANCE-PART V 81
ance companies. The industry has obtained relief from taxation which
formerly applied to insured pension plans and has taken a series of
steps to be more competitive. The first was to promote split funding.
In this arrangement, the insurance company managed the fixed-income
portion of the fund, while a bank would handle the remainder coin-
mitted to common stock investments. By offering a return based on the
earnings on new money, instead of the average portfolio earnings rate,
this type of contract has been an important competitive weapon. Since
new money ha~s been invested largely in corporate direct placements and
conventional mortgages, the highest yielding classes of assets, the rate
offered in recent years ha.s been relatively attractive.
The most recent step taken by life insurance companies to improve
their competitive position has been to secure authority to manage
"separate accounts" in which insured pension funds can participate.
These separate accounts enable a company to offer an equity investment
facility outside of the usual range of guarantee and valuation prob-
lems. This development is too recent (only $580 million in separate
accounts at the end of 166) to determine whether it will be of major
competitive assistance, but it is evidence that in the future life insur-
ance companies will be more effectively competitive in the pension field.
It is equally clear that the field of equity investment will become of
greater importance to life insurance company portfolio managers.
TRUSTEED PLANS
Evolution of Inves tnsent Man.agensen t
As previously indicated, trusteed pension funds ha.ve been one of the
most rapidly growing financial intermediaries, especially smce the
end of World War II. Projected growth, as discussed in chapter II,
assures them of a. prominent place among major participants in the
capital markets. Starting from a relatively small base prior to World
~\Tar II, these funds have passed through a complete evolutionary
development of investment management policy.
Initially, the restricted size and scope of noninsured plans influenced
their investment along traditional trust lines. It was customary to
invest in publicly issued bonds of high quality. Indeed the largest
aggregation of funds, those of the Bell System, were largely funded
through the issuance of 4 percent company notes from the start of the
i~lan in 1913, until the late 1930's. The transition of the combined pen-
sion, trusts of the American Telephone & Telegraph Co. and its sub-
sidmaries 1S shown in table V-4.
TABLE V-4.-BELL SYSTEM PENSION TRUSTS, 1935-50
[Book value of assets in millions of dollars at yearend[
Assets
1935
1940
1945
1950
4 percent demand and 1-year notes
Bonds of Bell System companies
U.S. Government securities
125.4
22. 5
2.1
103.8
45. 9
30.7
5. 1
45.7
277.5
200.9
362.0
Other corporate bonds
Other investments
32.9
102.7
646.6
10.1
Cash and accrued interest
2. 0
11. 9
12. 4
11. 7
Total
152.1
225.2
443.3
1,231.3
Source: American Telephone & Telegraph Co. annual reports.
PAGENO="0090"
82 OLD AGE INCOME ASSURANCE~PART V
Prior to World War II, there was, in effect, no investment manage-
ment of the Bell System trusts. They served merely as a medium for
accruing and recording the estimated costs of future pension benefits.
It was a frequent practice of industrial companies to write invest-
ment restrictions into their pension trust agreements which limited
the trustee to securities legal for life insurance companies or fiduciaries
in New York State. As the funds grew, bank trustees urged the inclu-
sion of modest amounts of common stocks, in the range of 10 to 25
percent.. In the late 1940's, with Government bond yields still pegged at
21/2 percent, equities were very attractive for current income. In 1947,
for example, common stocks of good quality offered a current return
of 5 percent while high-grade corporate bonds provided only 2.58
percent.
At. about this time, bank trustees in States such as New York were
greatly concerned about the plight of beneficiaries of personal trusts
restricted to a list of bonds legal for savings ba.nks. In 1950, they
succeeded in obtaining a 35-percent prudent-man rule, enabling them
to buy common stocks up to that proportion in legal trusts. In 1951, life
insurance companies were authorized to buy stocks in limited amounts
in New York State, and in the following year mutual savings banks
received similar authority.
Tile early postwar interest in equities, stimulated by their unusually
favorable yield advantage, was given legislative sanction in an 1111-
portant State by this series of changes in tile laws. It was hardly a
radical step for a company to amend the trust agreement to permit
investment, in secur~tjes legal for life insurance companies or fiduciaries
in New York. This automatically gave tile trustee authority to invest
up to 35 percent of tile fund in common stocks.
Thus, the breakthrough in investment policy came at the inception
of tile period of major growth in trusteed pension fund assets. Hecog-
nit~on was given to tile permanence. built-in growtil features, and lack
of liquidity requirements of the typical corporate pension fund. Also,
as costs mounted, business firms were made more aware of the im-
portance of earning tile better rate of return which equities could pro-
vide. Finally, it became increasingly clear that the company was fund-
ing a distant retirement income benefit which, together with OASI
benefits, would be suitable or appropriate to the standard of living
which the employee would be enjoying at the time of retirement.
Viewed ill these terms, tile commitment being funded is not determined
by some precise formula devised today and effective for decades to
come. but by a level of benefits which would carry out this broad objec-
tive. The better tIle earnings of tile fund, the more fruitful would be
each dollar of contributions.
Tile transition from a concept of meeting a Imown liability, essen-
ta.liv the sinking-fund calculation of the payments needed at an as-
sumed rate of interest to meet a future fixed liability, took place grad-
iiolly. The. dialog between company officials and trustees contmuecl
to broaden the concept of a. pension t.rust fund. The logical result is
the conclusion that a pension trust is a bundle of assets to be employed
as productively as possible on a long-range basis for tIle sole purpose
of meeting pensbin commitments already made (or likely to be made
in the future).
PAGENO="0091"
OLD AGE INCOME ASSURANCE-PART V 83
The actuary's vital role in pension planning is to provide answers to
such critically important questions as the following: (1) What bene-
fits would a fund of a particular size, fed by a certain rate of contri-
butions, be most likely to provide for this group of employees? (2)
What would be the cost of changing the pattern of benefits for this
group of employees? (3) How would alternative rates of contributions
affect the incidence of the costs of providing these benefits? These and
other questions involve assumptions as to mortality trends, turnover
among employees at various ages and after various lengths of service,
and the earnings of the fund.4 The rate of earnings on the accumulated
assets, unlike all of the other elements of cost, is something about
which the trustee can take specific steps.
It is a commonplace to observe that the future cost of a pension pro-
gram, over a long period such as 20, 30, or 40 years, is conjectural.
Except in retrospect, no one knows precisely what the plan costs. But,
the actuary could make a reliable estimate if lie knew the rate of
return that will be earned. This he does not know, nor does anyone
else, except within fairly broad limits. Yet it is a major factor. One
dollar a year accumulated and invested at 4 percent per annum builds
up to $57.20 at the end of 30 years. But, if the earnings are 5 percent,
the accumulation amounts to $68.09 in the same period, a difference of
19 percent.5
The problem of investment management, then, is not to earn any
particular rate but to earn the best possible rate which investment
opportunities permit. Corporate and Government bonds, mortgages,
real estate, equities, or whatever else that provides the best long-term
net yield may be the best medium of investment. (Yield is measured,
of course, in terms of income received plus or minus changes ~n value.)
Current thinking about the management of pension trust assets has
followed this line of reasoning to i:ts logical conclusion. The fund is
not analogous to life insurmlce assets or to a personal trust or to an
investment company or to any other financial institution. It is sui
gener~s. There is simply no other type of trust fund like it.~
This being the case, we might expect a distinctive pattern of pension
fund portfolios to emerge. This expectation cannot be either supported
or denied by experience to date. Investment management in a dynamic
economy like the United States is not a science, nor can one with confi-
dence set forth an estimate of the pattern which will emerge over the
next decade or more. Nevertheless, the trends of recent years are sug-
gestive and some estimates for the future can be entertained as being
plausible.
Port/o7io Composition
The distribution of noninsureci pension fund assets at the cud of
1965, and their growth during the past decade are shown in table V-5
and chart IT-i. This composite picture includes the pension plans of
business firms, nonprofit organizations, unions, and groups of em-
ployers acting jointly with unions.
There are, of course, other factors to be considered these are only the principal ones.
A clear exposition of the whole range of problems. intelligible to a layman, is to be found
In Jnmee A. Hamilton and Dorrsince C. Bronson, Pensions, New York, 1958.
`Computed from Jerome Bracken and Charles Christenson. Tables for the Analysis of
Capital Expenditures, Boston, 1961, "Amount to Which $1 Per Per'iod Will Accumulate,
Received Continuously."
Perhaps some charitable foundations have quite similar investment management
objectives, but relatively few would be identical in all respects.
PAGENO="0092"
84
Source: Securities and Exchange Commission.
NOTE: Book values, end of year.
TABLE V-5.-ASSETS OF PRIVATE NONINSURED PENSION FUNDS, 1965
[Dollar amounts in billions, book value]
Assets
Amount
Percent of
total
Increase
since 1955
Cash and deposits
U.S. Government securities
Corporate bonds
Preferred stocks
$0. 9
3.1
22.7
.8
1.6
5.3
39. 1
1.3
$0. 5
.1
14.8
.1
Common stocks
Mortgages
Other assets2
24.5
3.3
2.8
1 42. 1
5.7
4.9
21. 1
3.0
2.2
Total
58.1
100.0
41.9
Source: Securities and Exchange Commission.
If assets are valued at market, the common stock proportion was just under 55 percent.
2 Includes real estate and other assets not classified eloesihere.
Per cent
OLD AGE INCOME ASSiJEAXCE-PART V
CHART V-i. Portfolios of Private Noninsured Pension Funds, i95~-65
and deposits
S. govt. securities
Corporate bonds
Preferred stocks
Common stocks
1955 `56 157 253 1~ ~ 61
Mortgages
Other assets
`62 `63 `64 `65
PAGENO="0093"
OLD AGE INCOME ASSURANCE-PART V 85
The preceding table and chart illustrate the trends of the past dec-
ade for which good data are available. In contrast to life insurance
companies during the same period, pension fund trustees were not
active in the mortgage market, but, were in a position to expand their
common stock holdings at a rapid pace. Although it is not possible to
identify dollars, it is a fair statement that these stocks were largely
acquired with either incoming cash from contributions and interest
earnings or from temporary investments of new money. In any event,
net purchases of common stocks have not exceeded 60 percent of net
receipts.
One of the problems of obtaining good results from equity invest-
ment is, of course, the matter of timing purchases to avoid a concentra-
tion in periods of high prices. This can be done by a strict dollar-cost-
averaging program or a modification of one employing some sort of a
range concept. In the early stages of programs of accumulation, the
pattern appears to have been one of regularly allocating a substantial
fraction of new money to equities. As funds grew in size and as initial
objectives were achieved, the place of discretionary, as distinguished
from programed, purchasing has increased.
In the trends of recent years, certain possible implications for the fu-
ture are evident. First, it seems clear that common stocks will continue
to be a major outlet for investment even if they do not prove to be
quite as profitable as during the 1947-65 period of major upward re-
valuation of corporate earning power. Whether the proportion in equi-
ties eventually settles down at 40, 50, or 65 percent will depend upon
comparative yields in all areas of the capital markets.
In any event, it appears that there will still be a substantial volume
of business lending in the form of corporate debt, with a major frac-
tion of it placed directly rather than through public offerings. Con-
currently, we may expect a relatively great increase in mortgage lend-
ing and real estate investing. All of these trends are discussed in more
detail below as part of an analysis of the participation of private
pension programs in various sectors of the capital market.
Competitive Influences
Bank and company trustees have responded to the competitive drive
of life insurance companies by seeking higher yields on fixed-income
investments and by seeking to improve performance in managing stock
portfolios.~ The evidence of a more aggressive attitude is found in the
i'ecord of sales of common stocks as a percentage of the average market
value of holdings in corporate pension trust.
Percent Percent
1957 3.8 1962 4.6
1958 4.6 1963 6.3
1959 4.7 1964 7.2
1960 4.5 19651 7. 1
1961 6. 1 19661 7.8
Source: Securities and Exchange Commission.
All private noninsured pension funds, including corporate.
`Evidence of this intensification of interest in investment management performance is
to be found in the number of studies on the subject. See, for example, Peter 0. Dietz,
Pension Funds: Measuring Investment Performance, New York, 1966.
PAGENO="0094"
86 OLD AGE INCOME ASSURANCE-PART V
Other evidence of this more aggressive attitude is to be found in a
widening of the range of companies selected for equity investment,
greater concentration in directly placed corporate bonds, mortgage
lending, and real estate investing. The stimulus of competition be-
tween trustees, always active, has been reinforced by competition from
life insurance companies. As a result, there have been more imagina-
tive approaches to investment management and many innovations in
handling special types of assets. Special funds are established to hold
mortgages and real esta.te, for example, in order to provide economical
administration and diversification for the participating pension trusts.
The use of commingled funds permits smaller accounts to share in the
experience of a large and diversified portfolio.
Another tecimique for stimulating competition among portfolio
managers is the practice of splitting among several bank trustees a.
large ftmd such as that of General Motors, Eastman Kodak, or con-
stituent companies of the Bell System. The right to replace the trustee
on short notice, normally reserved in the trust agreement, a.lso mili-
tates against complacency. Some companies have established their
own investment management organizations (General Electric, United
States Steel, and Bethlehem Steel are examples) to assure the full
attention of qualified personnel.
The consequences of this greater emphasis on efforts to improve
realized yields have included the higher turnover in common stock
portfolios noted above as well as interest in a much wider diversity
of investment opportunities. Somewhat greater attention to the tim-
ing of commitments in both fixed-income securities and equities, a
very difficult undertaking, has become evident with the passing of
time. This trend represents a departure from the kind of dollar-cost-
averaging approach which predominated during earlier years.
FlexibiZity
Data are not available to show the extent to which pension fund
managers shift among different segments of the market for fixed-
income investments, but observations of the writer suggest that. flex-
ibility is becoming increasingly required by competitive forces. Will-
ingness to realize gains and losses has increased with greater recog-
nition of the fact that they are not really relevant to long-term per-
formance. Possibly this diminished "lock-in effect" has contributed to
the narrowing yield differential between discount corporate. bonds
and current coupon issues in the most recent period of high interest
rates.
Comprehensive and detailed information about asset management
policies is not available on a continuous basis, but it is possible to
document the changes which have taken place in the management of
common stock portfolios. From the investigation of the Fuhbright
committee, completed in l956,~ portions of the study made of the
New York State Banking Department in 1955,° and our analysis of
representative portfolios in 1958 and 1959, we can piece together a
general picture of how pension fund trustees managed their equity
U.S. Senate, Committee on Banking and Currency, Institutional Investors and the
~5tock Market 1953-1955, Washington, 195G.
George A. Mooney, Pension and Other Employee Welfare Plans, New York State Bank-
ing Department, 1955.
PAGENO="0095"
OLD AGE INCOME ASSIJRANCE-PART v 87
investments about a decade ago, when this problem had emerged as a
major aspect of total asset management.
These studies indicate that, in these formative years of the invest-
ment management of rapidly growing common stock portfolios, the
practice was to invest in a relatively stable group of well-established
companies. Because the portfolios were newly selected, there were few
occasions for sales. However, there was, even then, a wide range of
opinions as to the most suitable industries and companies.
The Fulbright committee staff selected 25 leading stocks in 17 in-
dustries for its study of activity during the 34 months from January
1953 through October 1955. The concentration of buying in these
"institutional favorites" was highest for corporate pension funds
(table V-6).
TABLE V-6.-PURCHASES OF 25 SELECTED COMMON STOCKS AND PERCENT OF TOTAL COMMON STOCK PURCHASES,
BY INSTITUTIONAL GROUPS, JANUARY 1953-OCTOBER 1955
Group
Purchases
(millions)
Percent
30 corporate pension funds
5 closed-end investment companies
25 life insurance companies
30 bank-administered common trust funds
~201.8
24.0
66.0
20.0
24.0
20.6
20.3
17.5
60 casualty and surety insurance companies
20 open-end investment companies
40 fire insurance companies
30.0
265.2
12.2
15.4
15.1
10.0
Total
619.2
17.9
Source: Fulbright Committee Report.
The indicated concentration of purchases declined lnarkedly during
the next several years. The National Bureau sample of 10 portfolios
for the 18-month period from January 1958 through June 1959 showed
only 16.5 percent of purchases in the same 25 leading companies.b0
Furthermore, 327 companies were represented in these portfolios dur-
ing the period. Only 42 were owned by five or more of the funds, while
228 were owned by only two or one of the funds. Similar data are not
available for later years, but the writer's direct observation supports
the conclusion that fund managers have continued to increase the
flexibility of their operations by broadening the rangeof alternatives.
The other major aspects of flexibility are, of course, in the timing of
equity purchases relative to fixed-income investments and changes in
the proportions of funds allocated to different types of bonds and
stocks. Aggregate data for all trusteed pension funds conceal a wide
range of policies pursued in individual funds. Over the 18-month
period, the banks in the National Bureau sample showed a range from
less than 23 percent to more than 45 percent in the allocation of new
funds to common stocks. In 6-month periods, the range of allocation
was between 7 and 69 percent despite an apparently stable level for
the group as a whole of between 34 and 42 percent. A similar pattern
~° The National Bureau sample was provided on a confidential basis by 9 banks whicla
are leaders in the administration of corporate pension trusts. These banks probably man-
aged more than 50 percent of all corporate pension funds. Each bank selected an vanidenti-
fled fund which was well established, unrestricted as to investments, and under tlae bank's
sole control. In effect, the quarterly statements furnished to us from December 31, 1957,
through June 30, 1959, recorded each bank's unfettered judgment of wlaat investment
policy should be. For the analysis of common stock activity, the College Retirement
Equities Fund was added to provide a 10th portfolio.
PAGENO="0096"
88 OLD AGE INCOME ASSURANCE-PART V
was found in the analysis of purchases of stocks when classified as
defensive, cyclical, cyclical-growth, and growth issues. Divergent re-
sponses to changes in the economic environment and to shifts in relative
values in the marketplace make sweeping generalizations about how
pension funds invest in equities of doubtful validity. The available
evidence supports, rather, the existence of wide variations among the
strategies and selective processes of different portfolio managers.
Little or no support is found for the view sometimes expressed that
the concentration of pension fund investing in a small number of lead-
ing company shares is tending to create scarcity premiums on institu-
tiona.l favorites. Attempts to identify such premiums have been un-
productive and there are no surprises in the results. Pension funds are
only one modest factor in a market which adjusts to underlying factors
of earnings, dividends, growth, and stability. While price adjustments
to changing expectations may take place slowly at times, there appears
to be no compartmentalization of the market for equities which can
create a different valuation for a few or many pension fund favorites
in the marketplace. The increase in competitive pressures on portfolio
managers, combined with the greater flexibility in investment deci-
sions which has developed over the years, has largely eliminated tend-
encies toward concentration in relatively few issues, if, indeed, they
ever existed in sufficient strength to influence selected stock prices.
A[uitiemployer and Union Pen.s ion Funds
Results of the National Bureau study of the asset growth of multi-
employer and union pension funds, 1959-64, are reported more fully in
the following report." The portfolio composition of these funds, as
Bartell has shown, varies in several material respects from the com-
posite picture of trusteed corporate pension funds. A comparison for
the year 1964, for example, shows the corporate funds holding higher
proportions of assets in common stocks (41.6 percent versus 23.5 per-
cent at book values). Mortgages and Government bonds were more
favored by the multiemployer and union funds. These differences are
not diminishing materially over time because of the active interest of
many unions in contributing to the supply of funds for housing. Some
of the newer and more rapidly' growing funds have, however, given
increasing emphasis to common stock investment.
The future pattern of investment policy will undoubtedly he in-
fluenced by certain union positions, notably instances of preferences
for mortgages and real estate over common stocks. Despite the rela-
tively rapid growth rate anticipated for the years immediately ahead,
these differences will have relatively little influence since their pro-
portion of the total assets of trusteed pension funds will increase only
from 6.3 percent to perhaps 7 or 8 percent.
If the unions should succeed in negotiating a voice in the manage-
ment of single-employer funds, more pressure might be exerted on
fund managers to finance housing or other "socially desirable" proj-
ects. On the basis of developments to date, this possibility seems re-
mote and there is some basis for questioning whether there would
11 See Part I of H. Robert Bartell, Jr., and Elizabeth T. Simpson, "Pension Funds of
Multiemployer Industrial Groups, Unions, and Nonprofit Organizations."
PAGENO="0097"
OLD AGE INCOME ASSURANCE-PART V 89
be any significant change in fund flows even if unions did obtain a
voice in investment management decisions.12
Nonprofit Orgar1ti~ation8
The smaller, but still vigorously growing, noninsured pension fuDds
of nonprofit organizations represent less than 3 percent of all private
noninsured funds. As shown in the following study by Simpson
completed as part of the National Bureau's pension research project,'3
the asset composition of these funds is similar to the `typical multi-
employer portfolio.
Again, possible changes in investment policies will not have an
important influence on aggregate capital market flows. Somewhat
greater emphasis on common stocks and mortgage lending is indicated
by the trends of recent years, but these asset shifts are in the same
direction as those evident in single-employer funds.
FUTURE CAPITAL MARKET FLOWS
Recent Trends
A comparison of net flows through `the capital markets during two
recent years, 1960 and 1965, will serve to illustrate the importance, as
suppliers of funds, of insured and trusteed pension programs cover-
ing individuals in `private employment. For three major capital market
sectors, the combined total of net acquisitions of financial assets by
these two investor groups is expressed as a fraction of the net change
in selected categories of assets. In the case of life insurance companies,
the arbitrary assumption is made that assets acquired to fund pension
plan reserves are distributed in the same proportions as all asset
acquisitions.
The important trends to be observed from table V-'? are the shifts
away from corporate bonds to corporate stocks and mortgages. With-
in the mortgage category, moreover, there has been a shift to multi-
family and commercial property lending as contrasted with the prior
concentration `of lending in the one- to four-family market. Direct
investment in real estate has also been growing in importance, although
not of sufficient size to be separately classified in the table.
As both life insurance companies and pension fund portfolio man-
agers have concentrated their corporate bond investments in directly
placed issues, they have relinquished leadership in the market for pub-
licly offered bonds to `State and local government retirement systems.
This shift has now been largely completed to the extent that offerings
and yields make directly placed issues competitively attractive.
The net purchases of corporate stocks (largely common) have run
in excess of net new issues by business firms, excluding investment
companies (see note 2, table V-7). This fact has given rise to wide-
spread discussion of the growth in institutionalization of stock owner-
ship, particularly when the growth of mutual funds is also taken into
account.'4 The net flows in the market for equity securities show a per-
is See the writer's "Management Interests In the Investment of Pension Funds," Pro-
ceeclings of the Eighteenth Annual Winter Meeting of the Industrial Relations Research
Association, December 1965, pp. 312-316.
~ Bartefl and Simpson, "Pension Funds," pt. II.
~ See, for example, Daniel Seligman and T. A. Wise, "New Forces in the Stock Market."
Fortune, February 1964, pp. 92-95, 194-206.
83-200-67-pt. 5-7
PAGENO="0098"
90 OLD AGE INCOME ASSURANCE-PART V
TABLE V-7.-NET ACQUISITIONS OF FINANCIAL ASSETS BY INSURED AND TRUSTEED PENSION PLANS COVERING
INDIVIDUALS IN PRIVATE EMPLOYMENT, 1960 AND 1965
Dollar amounts in billinnsj
Assets
1930
1965
Insured
pIano 1
Trusteed
plans
Total
Percent of
tota! funds
ruised 2
Insured
plans 1
Trusteed
plans
Total
Percent of
total funds
raised 2
U.S. Government securities_
State and local nbligations___
Corporate bonds
Corporate stocks
Mortgages
Other
-$0. 1
. 1
. 4
. 1
.6
. 2
-$0. 1
1. 6
1. 9
.3
. 1
-$0. 1
. 1
2. 1
2. 0
.9
. 2
(3)
(3)
35. 7
60. 6
6.2
Q)
-30. 1
-. 1
. 7
. 2
1.0
. 3
-30. 3
1. 5
3. 1
.6
-33. 4
-. 1
2. 2
3. 3
1.6
* 3
(3)
(3)
25. 6
97. 1
6.3
(3)
Total
1.3
3.8
5.0
(3)
2.1
4.9
7.0
(3~
Note: Individual entries may not add to totals because of rounding.
Source: Board of Governors of the Federal Reserve System, flow-of-funds accounts; figures differ from those compiled
by Securities and Exchange Commission because of differences in concepts.
1 That fraction of changes in life insurance asset holdings which is determined by the year's increase in insured pension
reserves divided by total net acquisitions of financial assets.
Total net acquisitions as percent of net new issues. In the case of corporate stocks, new issues include those of open-
end investment companies. If these are eliminated, the percenfage for 1960 in 111 and for 1935 about 1,100 because
net new issues from other sources amounted to only $300,000,000.
1 Not applicable.
sistent transfer of stock ownership from individuals and "all others"
to professional managers. The data exaggerate, however, the impact of
this trend on the market for equities.
In the first place, net new issues understate the supply of shares suit-
able for institutional ownership. Retirements tend to be dispropor-
tionately large for the smaller companies not widely available for pub-
lic investment. That is to say, retirements are heavily concentrated in
small privately owned businesse.s which sell their assets to large pub-
licly owned corporations. Similarly, when a closely held company
merges with one having a wide distribution of share ownership, the
supply of broadly marketable shares is increased without a new issue
being recorded. At least during recent. years of great onerger activity,
this imparting of marketability to inactively traded common stocks
has substantially exceeded the volume of recorded retirements. Gross
new issues, therefore, better measure the quantity of new shares broad-
ly available for pension fund investment. On average, this might add
upward of $1.5 or $2 billion a year to what can be considered as the
supply of new issues in the marketplace. Spot. checks have shown that
a disproportionately large fraction of the gross financing (including
senior securities convertible into common stock) has been done by coin-
panics which are prominent in pension fund portfolios.
Second, we know that even gross new issues are not a true measure
of the volume of equity financing. In a real sense, retained earnings
represent a privileged subscription, a rights offering if von will, which
has been fully taken up by existing stockholders. From 1961 through
1965, retained corporate profits averaged $18.5 billion a year. compared
with new offerings of common stocks averaging only $2 billion a year.
Pension funeis, of course, "purchased" their proportionate share of
these retained earnings.
PAGENO="0099"
OLD AGE INCOME ASSURANCE-PART V 91
Third, the supply of equity securities is constantly being enlarged
in the marketplace from sales by wealthy individuals or their executors
to meet Federal and State taxes on the transfer of property by gift
or bequest. To the extent that the proceeds of these sales are used to
meet such tax payments, instead of being reinvested in other common
stocks, they measure a demand for equity capital in the marketplace in
the same way as new stock issues. In a typical year. it appears that
sales for these purposes are about as large as gross new issues.
In short, the net acquisitions of common stocks for private em-
ployee pension programs appear very large only when some major
components of the aggregate supply are ignored. In perspective, it is
evident that these investors are still only one important factor in the
market and that their operations can have only a limited iiifluence on
the behavior of share prices.
Transaction studies made by the New York Stock Exchange over the
last 15 years show that, prior to 1965, institutions and intermediaries
typically accounted for less than 25 percent of the volume, on that ex-
change. Studies in March 1965 and October 1966, however, showed 31.4
percent and 32.5 percent, respectively. In the year 1965, when pur-
chases and sales on the New York Stock Exchange (twice reported
sales volume) came to $146.4 billion, the total of trusteed pension
funds' purchases and sales (obviously not all on the New York Stock
Exchange) came to $8.1 billion. Possibly these funds accounted for
5 percent of the total. This would compare with about 4 percent in
1960 and less than 2 percent in 1955.
Pension funds can hardly be credited with a major role in the short-
run behavior of share prices despite the rapid growth in their partici-
pation in the market.. They did, however, perform a stab~l~ziiig func-
tion in the 1962 market break 15 and sustained their net purchases in
good volume during the 1966 market decline.
Most recently, concern has been expressed over the rise in turnover
rates for pension t.rusts and life insurance companies, presumably
under the strong competitive pressure for performance. As such t.urn~
over rates are -frequently computed (the average of purchases and sales
divided by average holdings), a rapidly growing portfolio tends to
show a relatively high figure. Nevertheless, trusteed private pension
funds for the decade, through 1965, typically showed a turnover rate
of between 11 and 12 percent, well below the range of 12 to 16 percent
for all stocks on the New York Stock Exchange, and a 17- to 21-
percent range for open-end investment companies.
As the emphasis on performance mounted in 1966. however, turn-
over rates exploded to 33 percent- -for open-end investment, companies,
the group on which the impact of this emphasis was strongest. In con-
trast, a rise of about 2 percentage points in the turnover rate for
trusteed pension funds and life insurance companies was no more than
the rise in the New York Stock Exchange turnover rate.
WThile developments to date have had only a modest impact on the
structure of the market for common stocks, it must be recognized that
further increases in the rate of turnover in institutional portfolios
1~ See New York Stock Exchange, The stock Market Under stress, New York, March 1963,
and Securities and Exchange Commission, Report of Special Study of Securities Markets,
pt. 4, ch. XIII, Washington, 1963.
PAGENO="0100"
92 OLD AGE INCOME ASSURANCE-PART V
could bring about major changes in the. structure of markets. A rise in
the volume of block transactions of 10,000 shares or more from 3 to 4.5
percent of total New York Stock Exchange volume is the principal
consequence to date of greater institutional activity. An inventory of
institutional ownership of stocks (SEC estimates) at the end of 1965
is shown below.
Market Percent
value of total
(billions)
Personal trust funds
Investment companies
Trusteed private pension funds
Foundations
Property and casualty insurance companies
Life insurance companies
College endowments
Bank common trust funds
State and local government trust funds
$71. 9
41. 1
39. 7
14. 1
12. 4
9. 1
6.4
3. 5
2. 2
10. 7
6. 1
5. 9
2. 1
1. 8
1.3
.9
. 5
. 3
Total for selected institutions
All others, including foreigners
200. 4
474. 3
29. 7
70. 3
Total outstanding
674. 7
100. 0
Source: Securities and Exchange Commission, "Public Policy Implications of Investment Company Growth," Washing-
ton, 1966, pp. 276-277.
No. doubt, much more than 30 percent of all stocks are under some
form of institutional supervision or professional management. Data on
the whole range of investment advisory services are not available.
However, if the turnover of outstanding shares is to increase mate-
rially in the future, many holders other than pension funds will have
to participate on a broad scale.
Neither the proposition that pension funds bid up the share prices
for a limited group of "institutional favorites" nor the proposition
that they are contributing to. price volatility in selected shares can be
supported by available evidence. On the contrary, it appears that port-
folio managers are broadening their range of investment alternatives
and taking advantage of the long time-horizon implicit in their
decisions.
A Look at the Future
The longrun future of pension fund activities in the capital market
will depend upon the economic environment and the demands for
capital which are created. Rather than speculate on these matters, how-
ever, it may be more useful to set forth some educated guesses about
the future flow of funds to indicate the range of possibilities over the
next decade and more.
For this purpose, it is not too important whether insurance coin-
panics, bank trustees, or others administer the funds. An equity invest-
1r1ent., a directly placed corporate obligation, a real estate mortgage
loan, or a real estate purchase does not carry the label of the decision-
maker into the capital market. Furthermore, life insurance companies
are improving their capabilities to handle equity investments just as
bank trustees are building up their facilities fo.r mortgage and real
estate investing. In the future, it will clearly make less difference to
the flow of funds in the capital market whether a dollar of pension
contribution goes to any particular class of institution or type of in-
vestment manager.
PAGENO="0101"
OLD AGE INCOME ASSURANCE-PART V 93
It is vitally important, however, whether Holland's projections of
pension fund flows prove to be close to the actual events. As he cor-
rectly points out, the margin of error is likely to be modest for 1970 or
1975, but can increase materially for 1980. Essentially, he assumed that
the role of social security in the years ahead would not substantially
changed from what it h.as been. A few of the other crucial assumptions
relate to the pace o.f increases in coverage, changes in standards for
vesting and funding, and the rate of return earned by the investment
portfolios.
Because changes in a number of the variables can have the effect of
producing offsetting influences on the size of asset accumulations, we
can have confidence in using Holland's average annual fund accumula-
tion despite some recent surge of growth in years of high corporate
earning power. His projections for the pace of annual net addition to
assets follow:
Year Billions
1965 $6. 7
1970 7. 5
1975 7. 6
1980 7. 4
Source.: Holland's projections, table 28, P. 69.
This projected leveling off in fund accumulations, whether it occurs
in another decade or is delayed by current developments, is the most
significant aspect of future influences on capital market flows. Given
an orderly pa.ce of economic growth, it means that pension funds for
individuals in private employment will acc.ount for a diminishing
fract.ion of total ne.t acquisitions o,f financial assets.
As trusteed portfolios become more seasoned-that is to say, as the
average holding period of fixed-income securities lengthens-maturi-
ties and repayments will become much more important. In the 1970's,
we may anticipate that gro.ss acquisitions will continue to rise long
after net acquisitions have leveled off. A projection for a.ny year after
1970 will, therefore, be subject to a wide margin of error when ex-
pressed in terms of net acquisitions. With all of these qualifications, it
may be useful to look at some possible patterns for the future
(table V-8).
TABLE `/-8.-POSSIBLE FUTURE NET ACQUISITIONS OF FINANCIAL ASSETS BY PENSION FUNDS COVERING
INDIVIDUALS IN PRIVATE EMPLOYMENT, 1970, 1975, AND 1980
Assets
1970
1975
1980
Government securities
Corporate and other bonds
Corporate stocks
Mortgages
Other assets, including real estate
(1)
1. 3
4. 3
1. 6
. 3
0. 2
1. 9
3. 0
2. 0
. 5
0. 2
2. 4
2. 0
2. 2
. 6
Total
7.5
7.6
7.4
1 Negligible.
The division of fund flows between bond, mortgage, and real estate
investments is clearly a function of relative yields and terms. Since
it is impossible to determine precisely what these will be, the distri-
PAGENO="0102"
94 OLD AGE INCOME ASSURANCE-PART V
bution of funds among assets other than common stocks shown in
table V-S has been arbitrarily made. Some weight has been given,
however, to the anticipated strength in housing demand around the
turn of this decade.
To project a meaningful flow of funds into corporate stocks (largely
common), however, it was necessary to take some account of the effects
of market appreciation, reflecting the value of retained earnings. If
this appreciation should average 5 percent per annum in the years
ahead-not an unreasonable assumption on the basis of past history
and one consi~te.nt with the pursuit by fund managers of an aggressive
program to increase and maintain a substantial position in equities-
the composition of total portfolios would be approximately as shown
in table \T_9.
TABLE V-9.-POSSIBLE FUTURE ASSETS OF PENSION FUNDS CONCERNING INDIVIDUALS IN PRIVATE EMPLOY-
MENT, 1970, 1975, AND 1980
[Dollar
amounts in billions[
Assets
196
Amount
5'
Percent
1970
~-
Amount Percent
1975
~-
Amount Percent
1980
-
Amount Percent
Book values:
Common stocks
All other assets
$25. 5
59. 5
30. 0
70. 0
$47. 0 38. 7
74. 5 61. 3
$62. 0
98. 0
38. 8
61. 2
$72. 0
126. 0
36. 4
63. 6
Total
85.0
100.0
121.5 100.0
160.0
100.0
198.0
100.0
Market values:
Common stocks
All other assets
40. 0
58. 0
40. 8
59. 2
75. 0 50. 0
75. 0 50. 0
113. 0
99. 0
-
53. 3
46. 7
155. 0
130. 0
54. 4
45. 6
Total
98. 0
- 100. 0
150. 0 100. 0
212. 0
100. 0
285. 0
100. 0
Note: For purposes of this calculation, the artificial assumption is made that no gains or losses are realized, so that
changes in book values reflect only net inflows. The allocations of funds to common stock investment shown in table
V-8 are assumed to apply to the preceding 4 years as well.
I Actual. Based on arbitrary allocation of life insurance company assets to reserves for insured pensions plus Securities
and Exchange Commission data for trusteed funds.
The pattern of net financial asset acquisitions presented here is con-
jectural, to be sure, but it attempts to recognize several important
changes which will or may occur. The striking feature is, of course, the
projected dec;line in net acquisitions of common stocks after 1970. This
possibility deserves to be recognized because of the following three
factors or assumptions:
1. It is assumed that life insurance companies will greatly expand
their equity investments in the next few years through the use of sep-
arate accounts. Similarly, trusteed funds are expected to be heavy in-
vestors in equities in order to increase the proportion in existing port-
folios. The increase in the common stock component of the Bell Sys-
tem pe.nsion trusts is a substantial case in point.. As a consequence of
these factors, common stocks are assumed to increase from 30 percent
in 1965 to 38.7 percent by the end of 1970 at book values, and from 40.8
percent to 50 percent at market values.
2. As the number of beneficiaries climbs in the years immediately
ahead, funds accumulated under insured deposit administration con-
tracts will be use.d to purchase guaranteed annuities for the pensioner.
These funds, a1on~ with assets sunporting exi~t~ng deferred group and
individual annuity contracts, will be invested largely in fixed-income
securities. This is t.he dramatic maturing of pension promises, involv-
PAGENO="0103"
OLD AGE INCOME ASSURANCE-PART V 95
ing a doubling of the number of individuals drawing benefits between
1965 and 1973. It would be reasonable to suppose that at least $30 bil-
lion of the `$130 billion in asset.s other than common `stocks at market
values projected for 1980 will be supporting insured annuity con-
tracts on a guaranteed `basis. If these are eliminated from the total, the
proportion in common stocks would be close to 61 percent instead of
54 percent.
3. With increasing recognition of the fact that market, rather than
book, values are the significant measure of assets and yields, it is as-
sumed that portfolio decisions will be. based primarily on the distribu-
tion of assets valued at `average market prices. The fund, in reality, is
recognizing the purchase of retained earnings with the receipts from
average market appreciation. For example, for the year 1976, in terms
of these projections, the pension funds will receive $7.7 billion in net
additions to `assets plus $5.65 billion in appreciation on the $113 billion
common stock portfolio at the assumed average rate of 5 percent. The
investment in equities is the projected $2 billion in cash plus $5.65
billion in appreciation, or a total of $7.65 billion, which represents 57
percent of total "net receipts" of $13.35 billion for the year.
Obviously, this possible projection is not susceptible of proof, and
the declining allocation of funds to common stocks is not essential to
our conclusions. This exercise in quantifying a possible change in
trend serves its purpose if it only raises a question in the minds of
those who make freehand extrapolations of the rising volume of net
purchases of common stocks by these groups of investors. Even a stable
level of net common stock acquisitions `would mean a diminution of
capital market influence. Such a stable allocation at, say, 60 percent of
net additions to assets would produce the following results in 1980,
using the same assumptions as in table V-9.
[Dollar amounts in billionsj
Book value
Market
value
Amount Percent
Amount
Percent
Common stocks
All other assets
$93 47
105 53
$182
108
63
37
Total
198 100
290
100
The relatively modest change in assets at market value is attributable
to the delay in the projected cutbacks in stock purchases until after
1970 and 1975.
The reader may well conclude that this possible future allocation of
funds is more probable than that shown in table 17-8. However, even
this higher level of acquisitions implies that, in a growing capital
market, the influence of these investors in the market for corporate
stocks has already reached its peak and will be declining in the future.
On the other hand, the possible pattern in table 17-8 would suggest
a resurgence of influence in the markets for directly placed corporate
obligations, mortgages, and real estate.
But these pension fund managers will continue to be a major factor
in the market for common stocks outstanding in the hands of the pub-
lic. The recent growth of block transactions on the New York Stock
Exchange illustrates one of the adaptations of the market mechanism
PAGENO="0104"
96 OLD AGE INCOME ASSURANCE-PART V
which is necessary to handle the growing volume of turnover in insti-
tutional portfolios. Given the projected volume of equity holdings
by 1980, it is evident that an adjustment in the balance between fixed-
and variable-income securities can create substantial swings in the
amount of net acquisitions. There is little reason to expect that pur-
chases will be made at the even pace implied by table V_S. Nor, indeed,
can it be accepted as a certainty that fund managers will inevitably
continue their steady accumulation of common stocks through sus-
tained or even brief periods of adverse experience. In short, niuch
greater variability in net acquisitions of equity securities must be ex-
pected in the years ahead.
One final qualification to all these estimates and conjectures: If
variable annuity pension plans spread widely among industrial em-
ployers, it will be necessary to modify siThstantially the entire range
of possible asset acquisition patterns. Only the future can tell whether
this. is a real possibility. The increasing familiarity of the public with
common stocks and renewed concern over inflationary pressures in
the economy are factors conducive to the adoption of variable benefit
plans. However, stock purchase plans, savings plans which include
equity investments, and profit-sharing plans provide outlets for
this interest in owning common stocks without disturbing the stability
of fixed-dollar pension programs.
VI. THE IMPACT OF PENSIONS ON THE CAPITAL MARKETS:
GOVERNMENTAL PROGRAMS
Because of the widely differing characteristics of the various govern-
mental retirement income programs, no general observations are ap-
plicable to the whole range of Federal, State, and local systems. It is
necessary, therefore, to consider them separately as to present and
future influences on flows through the capital ma.rkets.
FEDERAL GOVERNMENT PROGRAMS
OASDI
While by far the most important of all Federal programs to pro-
vide retirement incomes, the Old-Age, Survivors, and Disability In-
surance program by its very nature is not. an important factor in the
capital markets. Asset holdings of the Federal Old-Age and Survivors
Insurance Trust Fund and the Disability Insurance Trust Fund ap-
pear below:
[Book values in bull
ons of dollarsj
Yearend OASI
Disability Total
1956 22.5 22.5
1957 22.4 0.6 23.0
1958 21. 9 1. 4 23. 3
1959 20.1 1.8 21.9
1960 20. 3 2. 3 22. 6
1961 19.7 2.4 22.1
1962 18.3 2.4 20.7
1963 18.5 2.2 20.7
1964 19. 1 2. 0 21. 1
1965 18.2 1.6 19.8
1966 20.6 1.7 22.3
Source: Social Security Bulletin, Statistical Supplement, 1965, and SSB,
July 1967.
PAGENO="0105"
OLD AGE INCOME ASSURANCE-PART v 97
The practice has been to maintain a contingency reserve of varying
amounts in the trust funds. That is to say, an approximate balance is
sought between payroll taxes and benefit payments so that the assets
held by the trust fund increase or decrease by only small amounts from
year to year. This means that the OASDI program plays only a minor
part in Treasury debt management operations. The trust funds can, of
course, participate in advance refundings and acquire public issues,
but there are not likely to be important net sales or purchases from
this source.
Federal Employees
In the 5 fiscal years ended June 30, 1965, the Federal Employees Re-
tirement Funds acquired $5.5 billion of Federal Government securi-
ties as the trust funds grew at a steady pace to more than $15 billion.
These funds, principally the Civil Service Retirement System, are far
from fully funded by conventional actuarial standards. There has been
no disposition on the part of the Congress to vote the necessary appro-
priations to fund the deficiency, and it appears likely that accumula-
tions will remain modest.
Unless this pattern is changed, these retirement programs for Fed-
eral employees will not become important in management of the public
debt. While receipts rose from a $2.0 billion rate in fiscal 1961 to $2.7
billion in fiscal 1965, expenditures showed an increase from $1.0 billion
to $1.4 billion. The cost-of-living adjustment feature could cause pen-
sion payments to rise even more rapidly under certain circumstances.
Railroad Retirement
The assets of the Railroad Retirement account have shown no sig-
nificant change for a number of years. A mature system, with an aging
membership and faced with the need for periodic increases in contribu-
tion rates, the Railroad Retirement System is unlikely to show any
growth in the years ahead. The present assets of $3.9 billion, further-
more, no longer represent an important accumulation of funds.
Armed Services and Veterans
While pensions to retired members of the armed services and vet-
erans' benefits are significant components of aggregate transfer pay-
ments, they involve no participation in the capital markets. Benefit
payments are current expenditures from general revenues as much a
part of the Federal budget as regular salaries. It seems quite unlikely
that any portion of accrued liabilities will be recognized through the
creation of trust funds and the accumulation of security holdings.
Thus, these large programs will not have any impact on the capital
markets except, of course, as they may indirectly affect the budgetary
position of the Federal Government, and, hence, its demand for funds.
Federal Government Programs as a TVhole
Taken as a whole, Federal social security and pension programs
produce a modest annual cash surplus on the average. If the adminis-
trative budget is in balance, the cash surplus is enlarged and publicly
held debt is retired to this extent. (Or publicly held debt is increased
less if the administrative budget is in deficit.) To this extent, Federal
programs are a factor in determining the new money needs of the
Treasury. On the basis of the past decisions made by the Congress in
PAGENO="0106"
98 OLD AGE INCOME ASSURANCE-PART V
repeated reviews and revisions of the applicable laws and appropria-
tions, however, it seems most unlikely that these programs will ever be
in a position to reduce substantially the publicly available supply of
Federal securities. The capital markets impact of their operations
will probably remain minimal, despite their extremely important role
in redistributing income.
The growth of Federal investment accounts does, of course, raise
some important questions about the rates of interest to be paid on
special obligations because it determines the amount of contribution
from general revenues in the form of interest payments. If only public
issues are held in these accounts, other questions about the maturity
distribution are bound to arise, especially when the Treasury is in-
terested in some rearrangement of the structure of publicly held
issues. As the economy has grown so much more rapidly than the
Federal debt during the postwar years, however, these questions have
become less pressing.
Periodically, proposals are made to employ funds of the trust ac-
counts in public or even private investments. The financing of urban
renewal, low-cost housing, or public projects for various purposes are
most frequently suggested. If this course were pursued, the impact on
selected sectors of the capital markets could be appreciable. No doubt
the relevant markets and participating financial institutions would
adjust to the new direction of fund flows, but in the process, we could
expect a significant impact on yield relationships. Other factors re-
maining equal, it would seem logical to expect a narrowing of the
differential in yields between direct Treasury obligations and the
favored investment media.
A different question regarding Federal programs is whether any
particular degree of funding (current recognition of the accruing costs
of commitments) is either desirable or necessary. Federal employees
and members of the armed services are not concerned about the
goodness of their pension promis~ just because there is an inadequate
sum, or nothing at all, deposited in a trust account for their collective
benefit; they are quite content to rely upon the Federal taxing power.
The basis for funding or lack of it seems to be largely traditional.
In the absence of other adjustments in the pattern of revenues and
expenditures, full funding on an actuarial basis through the issuance
of Federal securities would not change the position of the cash budget
but would enlarge the administrative budget by recognizing the full
cost of Federal Government services and national defense. There
would be a higher debt service charge and a larger public debt. Recog-
nition of the amounts owed to career public servants for deferred in-
come payable in retirement would not really change their position.
This is, of course, only one example of the vagaries of accounting in
the public sector, where the matching of costs and revenues is seldom
precise. It is simply a reminder to economic analysts that statistics
on public and private debt should not be accepted as raw material for
decisionmaking without careful study of the coverage of the data.
STATE AND LOCAL GOVERNMENT SYSTEMS
Evolution of Investment Management
As discussed in chapter II, the retirement systems of State and
local governments are emerging as major participants in the capital
PAGENO="0107"
OLD AGE INCOME ASSURANCE-PART V 99
markets now, and will continue as such in the years ahead. If our pro-
jections are not wide of the mark, these systems represent one of the
most dynamic groups of financial intermediaries. Attention has been
devoted primarily to corporate pension programs during the last 1~
years. It is high time that the changing role of these governmental
systems be recognized as the most important single development affect-
ing the capital markets of the next decade and more.1
Whether or not we have correctly appraised future growth in State
and local government retirement systems, the dramatic changes in their
investment management policies and practices are bound to have a
major impact on the capital markets, a portion of which has been wit-
nessed in recent years. These developments have been especially im-
portant because of the extent of the changes which have taken place.
Historically, the typical State or municipal pension fund was
handled by the State or city controller as a part of the debt manage-
ment function of his office. It was considered analogous to the manage-
ment of the sinking fund for term bonds, and the choice of invest-
ments was similarly restricted. In 1942, for example, the assets of
State and local retirement systems were distributed as is shown below:
Amount Percent of
(millions) total
Cash and deposits
U.S. Government securities
$72
318
3.9
17.1
State and local government securities
Other securities
1,342
131
72.0
7. 0
Total
1,865
100.0
Source: Bureau of the Census.
The increase in Federal income taxes, enhancing the value of the
tax exemption privilege, and the low volume of State and local bond
offerings caused a shift to U.S. Government securities during World
War II. By 1947, over 70 percent of assets were in Federal securities
and only 20.4 percent remained in State and local obligations. Other
securities and mortgages, however, still represented less than 7 percent
of the total.
By 1957, total asset holdings had grown to more than $11 billion.
The low rate of return, contributing to the high cost of retirement
benefits,2 and the example of noninsured corporate funds fostered
a continuing trend toward relaxation of statutory restrictions on eli-
gible investments. The resulting change in the distribution of assets
is shown in chart VT-i.
These expectations may not be realized, of course, if full funding is abandoned by some
of the more important governmental units. It is idle to speculate on whether or when this
will occur, but the possibility must be recognized.
a The extreme case was that of the New York City Retirement Systems, which as late
as 1059 were 72.6 percent invested in New York City bonds and 16.9 percent in ITS. Gov-
ernment securities. The deficiency of interest earned on their contributions below rates guar-
anteed to members of the systems between 1938 and 1964 aggregated $205 million. Maturi-
ties and sales of New York City bonds reduced the proportion to 32 percent of the $4.5
billion of assets on June 30, 1066. Nevertheless, the remaining $1.4 billion of city bonds
represents 57 percent of all tax-exempt bonds held by State and local government retire-
ment systems. Almost one-half of the bonds will mature during the next decade.
PAGENO="0108"
100
OLD AGE INCOME ASSURANCE-PART V
CHART VI-L-Portfolios of State and Local Government Employee Retirement
Funds, 1957-66
Per cent
The net acquisitions of financial assets between 1957 and 1966 fiscal
yearencls are shown in table VT-i.
TABLE VI-1.-ASSETS OF STATE AND LOCAL GOVERNMENT RETIREMENT SYSTEMS, FISCAL YEARS 1957 AND 1966
[In billions of dollarsi
Assets
1957
1966
Net change
Cash and deposits
Federal Government securities
0. 2
5. 1
3. 3
3.4
.
2
. 2
0.3
7. 0
2. 5
17.7
4. 1
1.8
1. 8
0. 1
1. 9
-. 8
14.3
3. 7
1. 6
1. 6
State and local government securities
Corporate bonds
Mortgages
Corporate stocks
Other assets
Total
12.8
35.2
22.4
S~'ir~e. Bureau of he Lensun
0
1957 `58 `59 `60 `~1 `62
Source: Bureau of the Census.
Noric: Book values, end of fiscal year.
PAGENO="0109"
OLD AGE INCOME ASSURANCE-PART V 101
Net acquisitions of corporate bonds during this period, it will be ob-
served, equaled the similar total for the much larger private trusteed
funds during a comparable period. These were predominantly high-
quality publicly offered bonds. Mortgage lending has expanded sub-
stantially in a few large funds and the pace of mortgage acquisitions
has been accelerating. rf1~e relaxation of statutory restrictions on com-
mon stock investments has been proceeding at a slow pace. In general,
the retirement systems have not fully utilized the permitted leeway
because of accounting conventions and other problems.
A change in the distribution of assets tends to reflect the decisions
of a relatively small numbe.r of very large funds. Assets are highly
concentrated in relatively few systems. State-administered funds in
six States, together with the New York City Retirement Systems,
accounted for 53 percent of the $35.2 billion book value of assets at
the end of the 1966 fiscal year. Another nine large State systems ac-
counted for an additional 15 percent of the total.3 In contrast, the
50 companies and union groups with the largest industrial plans ac-
counted for only 37 percent of total assets in 1964, and the top 15
corporate employers accounted for less than 24 percent of the total.
Put another way, there are now 10 individual State and city retire-
ment systems each with more than a billion dollars of assets. The
pension funds of corporate giants like Du Pont and Western Electric
have just crossed that figure. It is no exaggeration to characterize the
leading State retirement systems as the giants of the pension field.
Some years ago, this writer suggested that one might usefully com-
pare the diversification of State and local retirement systems with that
of corporate trusteed pension funds a decade eariier.~ With State and
local government securities excluded from the 1958 totals, the com-
parison was as shown below:
PERCENT OF TOTAL ASSETS AT BOOK VALUE
State and local Corporate
government trusteed pension
retirement funds 1948
systems, 1958
Cash
2
3
U.S. Government securities
46
39
Corporate bonds
Corporate stocks
42
2
43
11
Other assets, including mortgages...._
7
4
Total
100
100
By updating this kind, of comparison, we can measure the lag in the
investment policies permitted and pursued by managers of the State
and local funds:
`The six largest State aggregations of capital are in California, New Jersey, New York,
Ohio, Pennsylvania, and Texas. The above figures somewhat overstate the. case because
State teacher and public employee retirement systems, as in New York, may be admin-
istered by different trustees. 1-Jowever, the statutory authority is usually, the same for all
systems and there is a tendency for generally similar policies to be followed. (Ohio's three
systems are perhaps an exception.)
"Retirement System Investments," Report of the 44th Annual Convention of the
National Association of State Auditors, Comptrollers, and Treasurers, 1959, pp. 114-119
(reprinted as "New Investment Policies Loom for Public Retirement Systems," Commer-
cial ci Financial Chronicle, Oct. 1. 1959).
PAGENO="0110"
102 OLD AGE INCOME ASSURANCE-PART V
PERCENT OF TOTAL ASSETS AT BOOK VALUE
State and local
government
retirement
systems, 1966
Corporate
trusteed pension
funds, 1956
Cash and deposits
U.S. Government securities
1
22
2
14
51
Corporate bonds
Corporate stocks
Mortgages
Other assets
54
6
13
~
100
26
2
~
100
Total
The striking differences are, of course, the persistent lag in common
stock investment and the surge in mortgage lending during the past
8 years.5
In summary, then, we can observe that State and local government
retirement systems in the last 15 years have ceased to be captive markets
for governmental securities. They have entered the private capital
markets on a large scale-notably the corporate bond market,
more recently the mortgage market-and the market for corporate
common stocks on a modest scale. They have increasingly sought pro-
fessional investment advisory services and become conscious of rate of
return and investment management performance.6 They are somewhat
better prepared than in the past to follow the lead of private funds
in improving the earning power of their assets. Will they follow?
How far?
Restraints on Investment Management
Because they have been in existence longer, State and local retire-
ment systems have a somewhat higher level of benefit payments rela-
tive to the total of contributions and interest earnings. The contrast
with corporate trusteed plans is shown below:
Wollar amounts in millionsi
State and local Govern- Corporate trustend pension
ment retirement systems, funds, 1965
1965-66
Amount Percent Amount Percent
Employer contributions
Employee contributions
Earnings on investments
Other income and capital gains
$2, 630
1,771
1,370
45. 6
30. 7
23.7
$4, 368
581
2, 111
545
57. 5
7. 6
27.7
7. 2
Total receipts
5,771
100.0
7,606 -
100. 0
Benefits and expenses
Withdrawals
11, 859
359
32. 3
6.2
2, 040
26. 8
Total disbursements
2,218 -
38. 5
2,040
26. 8
Net additions to assets
3, 553
61. 5
5, 556
73. ~
Source: Bureau of the Census and Securities and Exchange Commission.
1 Benefits only; expenses were not reported to Census Bureau.
° Mv own projection for 1O6~ will apparently he close to the mark, except that I under-
estimated the substitution of federally insured and guaranteed mortgage in~estment~ for
direct T~.5. Treasury obligations (ibid.).
6 This trend is illustrated by the fact that in recent years consultants have been retained
to study tine investment management policies and procedures of a number of public funds.
One of these studies, made by this writer, was published in January 1904 by the New
York State Teachers Retirement System as part of the report of the Review Committee
of the Retirement Board.
PAGENO="0111"
OLD AGE INCOME ASSTJRANCE-PART V 103
Relatively large employee contributions, usually credited to an "an-
nuity" fund as distinguished from the "pension" fund, become almost
like a savings account, which the employee usually can withdraw
upon termination of employment and often can borrow against. In
conventional accounting terms, this fund from employee contributions
prior to retirement has some characteristics of a deposit-type liability.
A stable rate of interest is paid upon it and some withdrawals are
inevitable.
Furthermore, a number of systems are not being funded on an
actuarial basis by adequate contributions from the units of Govern-
ment. The claims or equity of employees may in fact be a large propor-
tion of the total assets on hand. In the two Massachusetts State funds,
for example, benefits and withdrawals represented 58 percent of con-
tributions. The locally administered systems in Illinois show a ratio
of 60 percent. In such cases, it is not surprising to find bond invest-
ments constituting practically all of the systems' portfolios.
Typically, there is no provision for dealing with capital gains and
losses except to include them in the current year's "earnings." The
"lock-in effect" is powerful after a long period of rising interest rates.
Similarly, there are no systematic provisions in general use for recog-
nizing more than current dividends from a portfolio of common
stocks. Their purchase may tend to depress the important rate of
return on employee contributions. This writer's argument for shifting
to a market-value basis for valuation, rate of return, and performance
measurement (or in the alternative to the use of capital gain and loss
reserve accounts) has received little or no acceptance among retirement
system administrators.'
More imaginative investment policies are inhibited by a number of
other factors. Retirement system trustees are not chosen just for their
knowledge of and experience with investments. They have other im-
portant duties to perform as well. The practice of seeking to secure
competent investment advice by competitive bidding, the inability to
pay adequate salaries for expert staff, and the apparent unwillingness
to lay out even very modest sums for investment management are all
factors conspiring to produce uninspired and mediocre portfolio man-
agement. Despite the great progress of recent years, few systems have
adequate staffs, strong investment advisory arrangements, effective
finance committees, and the capability of providing first-rate manage-
ment. These former sleeping giants of the pension-fund field some-
times appear to be oniy partially awake.
Under the circumstances, it is doubtful that State and local retire-
ment systems will soon break out of the statutory, accounting, and in-
stitutional restraints on their effective management of huge aggrega-
tions of capital. While the high cost of pension benefits will create
increasing pressure to improve rates of return, it is not likely that
the public systems will greatly accelerate the pace at which they follow
private funds. Nor is it likely that they will be as flexible in approach-
ing investment opportunities as they occur in the future of a dynamic
capital market structure.
See "Rates of Return: Standardizing Measurement," a paper presented at the Second
Annual Conference for Public Pension Fund Administrators at New York University on
Apr. 29, 1966. Reprinted in Commercial ~ Financial Chronicle, May 26, 1966.
PAGENO="0112"
104 OLD AGE INCOME ASSURANCE-PART 1/
FLTURE CAPITAL MARKET FLOWS
Federal Government Programs
As discussed at the beginning of this chapter, the Federal Govern-
ment's employee retirement programs have oniy a minimal impact
on the capital markets. Principally, their operations are relevant to
management of the public debt.
There is no evidence to suggest that future trust fund accumula-
tions will average appreciably higher than in recent years. Although
the Treasury is afforded some flexibility in debt management as a
result of its captive market for some public issues, it is difficult to see
how this provides much leverage on the recurrent task of funding and
refunding operations. Accordingly, we conclude that the operations
of Federal pension programs are unlikely to have any important ef-
fects on the capital markets in the near-term future.
State and Local Government Systems
The retirement systems of State and local governments, in contrast,
seem destined to be increasingly important factors in major sectors of
the markets for investment funds. The most startling of Holland's
projections was the potential growth in these funds. Although to
date actual additions to asset holdings have modestly exceeded his pro-
jections, it is well to recall his warning about the reliability of projec-
tions beyond 1975. Nevertheless, the approximate doubling of em-
ployees potentially eligible for coverage between 1965 and 1980 is not
unreasonable. High coverage and liberal benefits appear inevitable..
On a calendar-year basis, Holland's projections 8 were as follow-s
for fund levels and annual increase in assets in selected years:
[In billions of dollars[
Year
Fund accumulations
Net addition
1965
32.6
2.9
1970
1975
50. 7
76.9
4. 2
6.0
1980
114.9
8.9
The net demand for financial assets generated by these State and
municipal funds could surpass that of pension programs for individ-
uals in private employment during the year 1979. Whether or when
this actually occurs is not as important as the suggestion that., as an
emerging capital market factor, it is this sector, rather than the private
programs, which should engage our attention when we look to the
future instea.d of recounting the past.
Table VI-2 suggests a possible future pattern of acquisitions of
financial assets. To repeat, it must be obvious that capital market
demands, relative rates of return, the pace of relaxation of statutory
restraints on investment selections, and the institutional arrangements
made to manage these portfolios will determine what actually happens.
The projection is presented simply in order to illustrate what current
expectations suggest.
a Table 50, p. 131, adjusted from fiscal to calendar years.
PAGENO="0113"
OLD AGE INCOME ASSURANCE-PART v 105
TABLE vl-2.-POSSIBLE FUTURE NET ACQUISITIONS OF FINANCIAL ASSETS BY STATE AND LOCAL GOVERNMENT
RETIREMENT SYSTEMS, 1970, 1975, AND 1980
In billions of dollars]
1965 1
1970
1975
1980
Cash and deposits
U.S. Government secarities
State and local government securities
Corporate bonds
Mortgages
Corporate stocks
Other assets, including real estate
(2)
-0. 1
-. 3
2.3
. 6
. 4
. 4
(2)
(2)
-0. 2
2.0
1. 0
.9
. 5
(2)
0. 2
(2)
2.5
1. 3
1. 4
. 6
(2)
0. 4
(2)
3. 1
2. 0
2. 5
. 9
Total
3.3
4.2
6.0
8.9
1 Estimate for calendar year based on estimates and interpolation between fiscal year data.
2 Negligible.
The general assumptions implicit in table V]I-2 include: (1) Changes
will come gradually as in the past. (2) The reduction in holdings of
State and local government securities will continue for several years,
but about two-fifths of present holdings will be retained indefinitely
for special reasons. (3) The trend toward mortgage and real estate
activity will continue strong, probably stimulated by a strong housing
demand around 1970 and the subsequent years. (4) Common stock pur-
chasing will continue to lag, but eventually it will exceed one-quarter
of net flows, moderately stimulated by the adoption of variable annuity
options in scattered instances. (5) Mortgages and corporate bonds may
be considered interchangeable asset holdings with mortgages obtaining
a somewhat larger share of the systems' investment flow. (6) At some
point in the future, the Federal Government will offer securities to this
type of investor which are relatively more attractive than at present.
There is no presumption that State and local government retirement
systems, taken as a whole, will succeed in eliminating all of the archaic
statutory, accounting, and institutional constraints on effective asset
management. Some relaxation is expected, however, as the very size of
outlays for retirement benefits creates a demand for the more produc-
tive employment of funds.
The pattern presented here is consistent with the conclusion ex-
pressed earlier that these retirement systems will lag behind develop-
ments in the pension programs for individuals in private employment
and will not follow their lead in a number of respects. This can be illus-
trated by trying the 10-year test-that is, comparing the 1975 pattern
of asset distribution, exclusive of holdings of State and local govern-
ment securities, with the actual distribution of corporate trusteed pen-
sion funds in 1965.
PERCENT OF TOTAL ASSETS AT BOOK VALUE
Projected State
and local gay-
ernment retire-
meet systems,
1975
Corporate
trusteed pension
funds, 1965
Cash and deposits
U.S. Government securities
1
10
1
5
Corporate bonds
Mortgages
Corporate stocks
Other assets, including real estate....
49
18
14
8
5
45
5
Total
100
100
83-200-07--pt. 5~8
PAGENO="0114"
106 OLD AGE INCOME ASSURANCE-PART V
Our projections, in short, presume that in the future State and local
government retirement systems will not pursue the policies of cor-
porate trusteed funds because they will also be following, in part,
the traditional practices of legal reserve life insurance companies.
Even by 1980, our projections would produce an asset distribution
only moderately closer to that of the corporate funds of 1965.
Percent of Total Assets at Book Value, 1980
Cash and deposits 1
U.S. Government securities 8
Corporate bonds 45
Mortgages 19
Corporate stocks 18
Other assets, Including real estate 9
Total 100
Nevertheless, if these projections as to both the volume and distri-
bution of fund flows through State and local government retirement
systems are at all reasonable, a major financial institution will have
emerged in another decade. Now managing assets equal to 57 percent
of those of all mutual savings banks, these systems could be handling
assets 75 percent as great by 1975. Just as growth to date has brought a
material improvement in investment management, we can anticipate
continued progress in the quality of personnel and sophistication in
dealing in the securities market.
As investors in common stocks, however, it is not likely that the
system portfolio managers will depart from the current pattern of
dollar-cost-averaging a broad list of high quality issues. Our projec-
tions suggest the accumulation of more than $10 billion at book value
by 1975 and over $20 billion by 1980. On the assumption of 5 percent
per annum appreciation, the 1980 portfolio would be worth over $28
billion, but might still repre.sent only 23 percent of total assets at
market value. If annual purchases reach the $2.5 billion level by that
date, however, these systems will be a factor in the stock market
even if their turnover in existing holdings remains at a relatively low
level.
Some variable annuity plans have been adopted by State and local
governments, but other methods such as a fixed rate of increase in bene-
fits to meet rising living costs have also been employed. At this time,
it is difficult to tell whether equity funding, especially for teachers,
will spread widely. If it should, of course, the effect upon the volume
of projected net acquisitions of common stocks would be substantial.
The picture which emerges is of an eminently respectable, slow
moving, conservatively oriented structure of financial institutions per-
forming an important function with care and prudence. As competi-
tive factors have encouraged risk-taking in other institutions, the ques-
tion has frequently been asked: "But who is going to hold the really
high quality investment securities?" Here is the answer: State and
local government retirement systems. Their capacity to do so should
be reassuring.
VII. CONCLUDING OBSERvATIONS
1~\T0 have seen that the economic aspects of pensions are as broa.d as
tile flows of income, consumption and saving, and the network of
PAGENO="0115"
OLD AGE INCOME ASSURANCE-PART V 107
public and private arrangements which characterize an urban indus-
trial society. We have reviewed the creation and development of a
mammoth pension structure and have attempted to preview its future
growth and progress in attaining maturity. In the process of looking
at these several segments of the whole, we have been at pains to remind
the reader that pensions are not a separate structure but a part of the
warp and woof of the fabric of our economy. The purpose of this
chapter is to reflect briefly on some of the implications of this study of
economic aspects of public and private pensions.
IMPLICATIONS FOR SAVING AND ECONOMIC GROWTH
The social security and other pension programs of the Federal Gov-
ernment, we have seen, act to sustain consumption and to depress the
ability of individuals to save. The present system of payroll taxes to
finance the OASDI system is less powerful in this direction than the
financing of noncontributory programs through general revenues. It is
simply not possible to lift living standards of the aged without the
redistribution of income which these and other fiscal activities involve.
The size and especially the prospective growth of these redistributive
arrangements should, however, be taken into account in any appraisal
of the influence of the total Federal tax structure on economic growth.
The pension plans covering employees of State and local govern-
ments and individuals in private employment, to the extent that they
are systematically funded, generate saving which is substantially a net
addition to total saving in the economy. This is especially significant
because the saving permits investment in business capital and housing.
The flow of funds to finance business plant and equipment, inventories,
research and development, and trade credit plays an increasing role in
enlarging and improving the efficiency of productive capacity.
On the basis of our projections, the net flows into the markets for
corporate securities and mortgages from both State and local govern-
ment and private pension plans will increase from $10.4 billion a year
in 1965 to more than $12 billion by 1975 and over $14 billion by 1980.
These figures for net flows ignore the funds also provided in the form
of corporate retained earnings of portfolio equity securities. To that
extent, the projections represent an understatement of the volume of
business capital financed. In any event, unless there is a major change
in the trends presently indicated, it is clear that while these funds will
finance a growing amount, they will contribute a diminishing share to
the growth of capital assets in the business sector. This may or may not
be disturbing to our expectations for economic growth, depending
upon how we anticipate developments in other influences on the saving
and investment process.
The projections, of course, are only an expression of the probable
net effects of many influences. The realization of substantially higher
returns can depress the level of contributions and of pension savings.
An acceleration in the pace of the extension of coverage, in the trend
toward more liberal vesting provisions, and in the rate of funding can
continue to increase the rate of pension saving for another span of
years. Our analysis of the working of the pension structure is, there-
fore, more illustrative than predictive.
PAGENO="0116"
108 OLD AGE INCOME ASSURANCE-PART V
Our exploration of the question of the burdensomeness of pension
arrangements suggests that there is no precisely determinable level of
what the economy can afford without sacrificing some of its vitality
and potential for growth. The size of the net burden attributable to
the structure of benefit programs is apparently not great, especially if
the plans for employees of State and local governments and private
organizations continue to carry an important share of the provision of
benefits.
The need for improved data and techniques for the measurement of
the gain-loss patterns involved in huge transfers of income has been
demonstrated by our analysis. The fruitfulness of further investiga-
tion and research in this area is evident for the informed evaluation
of the economic consequences of alternative courses of action.
To set tax-supported pension programs apart from all of the other
fiscal operations of government and to attempt to assess their influ-
ence on incentives and productivity gains is to create an artificial and
unreal framework of analysis. Rather, the issue of what we can afford
in the way of old-age income provision must be considered together
with the whole range of public policies which affect the returns to dif-
ferent factors of production. Indeed, the interrelations between these
and other public welfare objectives must be examined and constantly
reexamined in a changing economic environment.' It is idle to appraise
the influence of pension commitments running far into the future
apart from the whole range of comrnitment~s being made in other
areas. What we can afford, in some meaningful sense, is the total
share of real output that can be diverted from the factors of produc-
tion which provide it without impairing the incentives and motiva-
tions for continued expansion and growth.
IMPLICATIONS FOR ECONOMIC STABILITY
Pension and disability benefits clearly operate as contracyclical
influences in the direction of economic stability. Retirements tend to
increase when employment opportunities wane. The level of con-
tributions, especially in private plans for the funding of past service
liabilities and in profitsha.ring plans, is sensitive to changes in cor-
porate profits. Income maintenance is aided by rising benefits in
periods of slack employment, and pension saving declines slightly in
periods of less active demand for business capital investment.
In any case, the regularity of benefit payments is another of the
built-in stabilizers in the economy.2 Public and private pension pro-
grams will continue to provide a growing share of the income pay-
ments not susceptible to cyclical variations in aggregate economic
activity.
The question has been raised as to whether pension saving is not too
stable in times of deficient aggregate demand for consumer goods and
services. As a consequence, it is argued, the cyclical changes in saving
rates which contribute to stability are muted by the regularity of pen-
sion saving.3 One answer is that the limits on the variability of pen-
1 See this writer's "The Future of Private Pensions: Some Economic Aspects," Journal
of Ri.s'k and Insurance~ March 1967. pp. 27-32.
2 Cf. Arthur F. Burns, "Progress Towards Economic Stability," American Economic
Review. March i960.
[Nelson D. McClungJ Old Age Income Assurance: An Outline of Issues and Alternat7ves,
U.S. Congress, Joint Economic Committee, November 1966, p. 18.
PAGENO="0117"
OLD AGE INCOME ASSURANCE-PART V 109
sion saving are not entirely indigenous to the pension structure. In the
case of private plans, indeed, corporate managements have preferred
to use variable contributions as a method of averaging income. Public
regulation to assure the fulfillment of pension promises seeks to reg-
ularize contributions and so does the Internal Revenue Code. Finan-
cial analysts and the public accounting profession seek regular recog-
nition of pension costs as they are incurred and not when flexible con-
tributions are actually made to the fund.4 Also, variability in employer
contributions is not feasible under the budgeting practices of State
and local governments.
Essentially, however, the reality is that pension systems are not well
designed to provide variability in saving flows. The long-term nature
of their contracts calls for regularity in provisions to meet them. Final-
ly, equity investment is a major outlet for the saving flow, and there is
some presumption that rates of return will prove higher on investments
made during periods of slack economic activity. The stabilizing in-
fluence of pension fund investing on the secondary market for equity
securities would diminish if pension saving flows were permitted to be
highly variable.
One of the most important aspects of economic stability is the ques-
tion of inflation. We have repeatedly observed that inflation can erode
the value of pension promises and warned that, unless the burden of
income transfers is willingly borne by the working members of society,
they will acquiesce in policies which lighten the burden by inflation.
This unfortunate outcome of the pension movement is not now in sight.
But, neither has the full burden been felt. The volume of claims to be
presented has only well begun its long rise. Capital formation at a
high level has spurred real output.. Pension saving in the future may
contribute less to this progress if, in fact, saving is the limiting factor
on economic growth.
As we move into the period of substantial rise in benefit payments
and witness the diminishing pace of pension saving, it will be necessary
to adjust fiscal policy to the changing situation. Again, the availability
of more precise measures of possible future effects of income redistribu-
tion through pension programs will be required to judge the adjust-
ments most appropriate to the emerging situation.
IMPLICATIONS FOR THE CAPITAL MARKETS
The stability of net fund flows makes pension systems almost unique
among the major suppliers of funds to the capital markets. Apart
from the mild cyclical fluctuations mentioned above, no unpredictable
changes in inflows need be anticipated by the portfolio manager. An-
other unique characteristic is the absence of valuation problems. There
is no requirement, as with a deposit-type financial institution or a life
insurance company, to demonstrate on a certain day an excess of assets
over liabilities on the basis of some prescribed or conventional valua-
tion of assets.
These two salient characteristics, the absence of both liquidity and
published statement requirements, impart different dimensions to
Ernest L. Hicks, Accounting for the Cost of Pension Plans, Accounting Research Study
No. 8, American Institute of Certified Public Accountants, i065.
PAGENO="0118"
110 OLD AGE INCOME ASSURANCE-PART V
the portfolio management decision to be made in relation to the long-
time horizon of pension commitments. Despite competitive pressures
to show good performance, the outcome of decisions is still to be judged
over an extended period of time. If the illiquidity of a financial asset
carries a premium in yield, pension funds are in about the best position
to capture it. Hence the evolution of portfolio management has been
steadily in the direction of holding less liquid assets, even in State
and local government retirement systems.
The stability of lending and investing in the capital markets ob-
scures some variability in the pace of forward commitments for
directly placed corporate obligations and mortgage loans. The
tendency to enlarge forward commitments in periods of strong demand
for funds may have the effect of contributing to a situation in which
idle balances are being activated and velocity is rising. However.
there can be no important shift into claims on pension funds and
pension fund managers have little capacity or inclination to supply
liquid assets, such as Government securities, to holders of idle
balances.5 The important role of these funds, therefore, is as a rather
neutral financial intermediary between savers and investors, with little
impact on income velocity or the money markets.
We have observed that both State and local government and private
plans, as suppliers of loanable funds, have shown a strong preference
for corporate securities. Our projections for the future show a con-
tinuation of the growth of participation in the area of real estate
finance. Slow but steady progress has been made in solving the ad-
ministrative and expense problems of handling mortgages. It is now
possible for a pension fund to secure most of the services normally
encompassed in home-office administration from organizations which
economically perform these functions. A mortgage portfolio can be
handled with almost the ease and economy of a bond portfolio. If net
yields after allocated expenses are competitive~ risk factors considered,
there will undoubtedly be a substantial growth in pension fund mort-
gage lending across the Nation in conventional as well as FHA-insured
and VA-guaranteed loans. By the late 1970's, the volume could easily
be comparable to the average net acquisitions of mutual savings
banks in recent years.
The ownership of real estate equities would appear to be a natural
avenue of investment. Sale and leaseback financing has been, in fact,
a growing outlet. The tax-exempt status of a private pension plan,
however, can be impaired by engaging in an unrelated business. The
operation of income-producing property, especially if the purchase
is financed with borrowed funds and only an equity position is re-
tained, is susceptible to being considered such an unrelated business.
The tax benefits from accelerated depreciation to a real estate
operator, and his ability to introduce, substantial financial leverage,
usually justify his paying a higher price for property than a pension
fund is prepared to pay without. these possibilities. Hence, pension
fund ownership of true real estate equities is not likely to grow rapidly
in the years ahead.
For a full discussion of these points. cee Victor L. Andrews. `Xoain~nred Cor-
porate and State and Local Government Retirement Funds in the Financial Structure," in
Private Capital Markets, Engiewood Cliffs. N.J.. 1964, especially pp. 465-4S1.
PAGENO="0119"
OLD AGE INCOME ASSURANCE-PART V 111
Corporate equity securities, we have seen, are likely to continue to
occupy a major position in privately organized pension programs and
to become increasingly important in State and local government re-
tirement systems. Common stocks have historically produced a higher
tota.l yield than bonds or mortgages for the holder in a position to ac-
cept price volatility and irregularity in the realization of long-term
rates of return. Pension funds are particularly well situated to accept
these disadvantages of corporate equity securities. The principal
limitation on their role is the need to support guaranteed annuity con-
tracts and the problems with stocks as an investment medium for em-
ployee contributions subject to withdrawal and borrowing privileges.
In the accumulation of pension fund assets for the provision of fu-
ture benefits of indeterminate amount, common stocks have especial-
ly desirable charanteristics. The recent changes in life insurance opera-
tions to provide for separate accounts for corporate stocks should
stimulate the rate of accumulation over the immediate future. Wrhether
the benefits of equity investment will be more widely shared with
present and future pensioners by use of the variable annuity contract
is less certain. The 15-year record of the College Retirement Equities
Fund in providing variable benefits for educators is persuasive of the
merits of this approach, but many employee groups are iiot anxious to
trade off the certainty and stability of retirement income for the pos-
si'bilit.y of a materially higher but fluctuating average level of bene-
fits. There are also communication problems involved, and CREF's
experience may not be readily transferable to other situations. Never-
theless, a trend toward the greater use Qf variable annuity arrange-
ments is the most important single factor which might affect our pro-
jected capital market flows.
Our projections show modest purchases of TJ.S. Government securi-
ties at some point in the future and an early cessation of net liquida-
tion. This reflects the assumption that in the course of an orderly
approach to public debt management the 41/4 percent interest rate ceil-
ing on long-term bonds will be removed and that the Treasury will find
occasions to offer securities which are attractive in comparison with
alternative investments in terms of yield, freedom from an early call
provision, and marketability. There is always room for marketable se-
curities, especially as the concentration in direct placements, mort-
gages, and common stocks limit flexibility in portfolio management at
times.
Fund management has been criticized both for being too cautious
and for accepting too great risks.6 Equating the proportion in common
stocks with high risk ignores what we know of the role of diversifica-
tion-the right combination of high-risk equities. with a minimum of
covarianc~e between them, can comprise a. portfolio with very limited
risk. The investment of pension funds in unseasoned, marginal enter-
prises, on the other hand, clearly raises questions a.s to whether the
trustees are observing the long-established standards of the prudent-
man rule.
There has not been any visible pattern of relating the aversion t.o
risk in the pension fund portfolio to the risk characteristics associated
`~ See, for example, Old Age Income Assurance, pp. 1-20.
PAGENO="0120"
112 OLD AGE INCOME ASSURANCE-PART V
with the enterprises which contribute to it. If there were, we should
expect to find the highest acceptance of risk among pension funds for
State and local government employees or public utilities. Conversely,
the greatest aversion to risk should be characteristic of plans for em-
ployees in highly cyclical or chronically unstable activities. Actually,
almost the reverse is true. This may be a result of transferring to fund
investment management decisionmaking the attitudes and outlook
most frequently a.pplied to the organization's internal investment
decisionmaking.
Generalizations about the cautiousness of fund managements are,
therefore, difficult to make. All kinds of portfolio policies are emerging
and being followed. The spectrum of policies is nearly as broad as the
range of investment opportunities. Over time, the funds will flow to
the areas of the best returns, the allowance for risk being taken mto
account not perfectly but a.t least rationally.
`Within the framework of recognized standards of fiduciary re-
sponsibility, then, the flow of pension saving will continue to spread
through most segments of the capital markets. It is not material if
these funds acquire predominantly sea.soned equity securities, for ex-
ample, because those who sell to pension funds may be in a position
to reinvest in companies with greater risk exposure. In an aggregate
flow-of-funds view of the capital markets, the significant factors are
aggregate sources and uses, the a.bsence of compartmentalization of
markets, and a market structure which permits prompt responses to
changing demands.
The growth of insured and.noninsured pension plans for individuals
in private employment, because of their flexibility in the allocation of
saving flows, has probably done more than any `other single develop-
ment to improve t.he brea.dth and responsiveness of the capital markets
to changing patterns of demand. They provided a major remedy to
the shortage of equity capital in the period immediately following
`World War II, as evidenced by the subsequent recovery in the value of
corporate earning power to previous levels. They have contributed to a
closing of the yield differential between directly placed and publicly
offered corporate bonds in the upper-quality ranges. They may have
aided the responsiveness of certain classes of mortgage yields to bond
yields.7
The picture which emerges is of an additional source of funds for
the capital markets with a. minimum of permanent commitments to
any particular sector of those markets. In the years ahead, we can an-
ticipate somewhat greater flexibility in the allocation of fund flows
and greater responsiveness to yield differentials which express the com-
parative intensity of demands for funds. Our projections are illustra-
tive of possible patterns of response, but what actually takes place will
be a function of the changing pressures in the ~narket.place. On balance,
these public and private pension accumulations have made a major
contribution to the efficiency of the. capital markets in channeling funds
to the most productive areas of investments.
More definitive conclusions on these points will emerge from the National Bureau's
interest rate study. See Joseph W~. Conard, Tue Behavior of Interest Rates: A Progress
Report, New York, NBER, 1966.
PAGENO="0121"
OLD AGE INCOME ASSURANCE-PART V 113
IMPLICATIONS FOR FINANCIAL INSTITUTIONS
The increasing readiness of noninsured private pension trusts and
State and local government retirement systems to participate in the
mortgage market on a much larger scale suggests that life insurance
companies, savings and loan associations, mutual savings banks, and
commercial banks will find new competition for loans. Some of these
institutions have already been working to establish relationships and
to provide essential services to fund administrators. Mortgage hankers
are also active in tailoring their facilities to these new markets. Such
arrangements and correspondent services will undoubtedly develop
further.
Efforts to develop a secondary market for mortgages seem unlikely
to engage the interest of pension fund portfolio managers. The pos-
sibility of resale is well down on the list of desired objectives. Greater
uniformity of mortgage terms and characteristics would, however, be
an attractive feature of mortgage market developments.
Unless there is a major change in the saving habits and motivation
observed in our study, other financial institutions will have good mar-
kets for financial services which either supplement or complement
pension saving. Mutual fund plan accounts, efficient financing of
household capital, additional life insurance protection, and variable
savings accounts are a few of the possible areas of growth which may
be stimulated by the extension of pension coverage.
IMPLICATIONS FOR PUBLIC POLICY
The Cabinet Committee appointed by President Kennedy chose as
the title of its report "Public Policy and Private Pension Programs."
Major sections of that report dealt with economic aspects of pension
growth and the public interest in private pensions. Although it is not
within our province to make or endorse policy recommendations, cer-
tain of our findings are relevant to the Committee's analysis.
The Committee recognized that it is essential to the operation of
public and private pension programs that the highest standards of
fiducial responsibility be maintained. More complete disclosures of
portfolios and changes in them were recommended. Also, the Commit-
tee advocated strengthening statutory provisions for enforcing recog-
nized standards of fiducial responsibility in preference to the applica-
tion of regulations or formulas which would reduce the flexibility of
asset management.
Our analysis of portfolio management indicates that diversity and
flexibility of investment decisions account for much of the contribu-
tions of these funds to the more efficient functioning of the capital
markets. At the same time, we have found no reason to conclude that
more complete disclosure to participants would hamper portfolio
management.
Another relevant issue of public policy is the question of concentra-
tion of economic and financial power. Limitations on the purchase of
employer securities in private funds have long been recognized on the
grounds that collateralizing a promise with the proinisor's evidence of
debt or a share in its equity is no security at all. At the same time, this
type of limitation seeks to prevent use of the pension funds for pur-
PAGENO="0122"
114 OLD AGE INCOME ASSURANCE-PART V
poses of control, support of the market for employer securities, facil-
itating acquisitions or control of other companies, and in other ways
transforming the fund into an agency for purposes other than its
intended one, the funding of pension commitments. In this report we
have accepted the view that the governing considerations in portfolio
management decisions will be comparative yield expectations and not
the search for control or opportunities to exert influences on portfolio
companies.
Certain safeguards are already operative. The most important is the
established requirement that a trustee show undivided loyalty to his
trust. Conflicts of interest must be avoided. The record of life insurance
companies and bank trustees is excellent in this respect. Economic
pressures are equally powerful in the same direction. Increasing empha-
sis on the quality of investment management to reduce the cost of
pension benefits has strengthened the competitive forces at work. As
large employers have divided their funds among a number of bank
trustees or placed different funds under the management of different
trustees, the measurement and appraisal of results have emerged as a
practical prohibition against any course other than strict attention to
the business of investing.
It is true, of course, that some very large concentrations of assets
are emerging, particularly in the case of State-administered public
employee retirement systems. Despite a few lapses from undivided
loyalty to their participants, public fund administrators also have an
excellent record of probity. Examinations of many cases by State in-
surance departments reinforce the system of internal controls. The
trustees of these systems have their good names, and often public
office, at stake in the administration of the retirement systems.
Admittedly, it is not feasible for a State retirement system to split
its assets among several trustees as industrial corporations have fre-
quently done. But concern is more properly with lethargy and lack
of flexibility in asset management which may affect these large pools
of capital, especially when governmental units show reluctance to
employ qualified staffs to deal with financial management responsi-
bilities of these proportions.
Our study suggests, therefore, that competitive factors and greater
disclosure are exercising strong pressures against the abuse of eco-
nomic and financial power. It is clearly appropriate, however, that
these issues in the realm of public policy should be examined and de-
bated. In a. relatively short span of years, public and private efforts
have brought major new financial institutions into being. Mass cover-
age of the contingency of lOSS of income because of age or disability
has been extraordinarily successful and shows promise of even further
development. It is only prudent that we should take stock of both the
economic and public policy issues which emerge, in this case, from the
realization of accomplishments beyond our expectations. We can only
hope that this kind of review will take place again and again, each time
with better grounds for reaching judgments. We can also expect that
it will become common knowledge that the validity of pension prom-
ises ultimately rests on the capacity of our economy to grow in pro-
ductivity and to achieve higher standards of living for citizens of all
~tges.
PAGENO="0123"
PENSION FUNDS OF MULTIEMPLOYER INDUSTRIAL
GROUPS, UNIONS, AND NONPROFIT ORGANIZATIONS*
BY H. ROBERT BARTELL, Jr.,f and ELIZABETH T. SIMPSONX
RELATION OF THE DIRECTORS TO THE WORK AND PUBLICATIONS
OF THE NATIONAL BUREAU OF ECONOMIC RESEARCH
1. The object of the National Bureau of Economic Research is to ascertain
and to present to the public important economic facts and their interpretation
in a scientific and impartial manner. The Board of Directors is charged with
the responsibility of ensuring that the work of the National Bureau is carried
on in strict conformity with this object.
2. To this end the Board of Directors shall appoint one or more Directors of
Research.
3. The Director or Directors of Research shall submit to the members of the
Board, or to its Executive Committee, for their formal adoption, all specific
proposals concerning researches to be instituted.
4. No report shall be published until the Director or Directors of Research
shall have submitted to the Board a summary drawing attention to the character
of the data and their utilization in the report, the nature and treatment of the
problems involved, the main conclusions, and such other information as in their
opinion would serve to determine the suitability of the report for publication in
accordance with the principles of the National Bureau.
5. A copy of any manuscript proposed for publication shall also be submitted
to each member of the Board. For each manuscript to be so submitted a special
committee shall be appointed by the President, or at his designation by the
Executive Director, consisting of three Directors selected as nearly as may be
one from each general division of the Board. The names of the special manu-
script committee shall be stated to each Director when the summary and report
described in paragraph (4) are sent to him. It shall be the duty of each mem-
ber of the committee to read the manuscript. If each member of the special
committee signifies his approval within thirty days, the manuscript may be pub-
lished. If each member of the special committee has not signified his approval
within thirty days of the transmittal of the report and manuscript, the Director
of Research shall then notify each member of the Board, requesting approval or
disapproval of publication, and thirty additional days shall be granted for this
purpose. The manuscript shall then not be published unless at least a majority
of the entire Board and a two-thirds majority of those members of the Board
who shall have voted on the proposal within the time fixed for the receipt of votes
on the publication proposed shall have approved.
0. No manuscript may be published, though approved by each member of the
special committee, until forty-five days have elapsed from the transmittal of the
summary and report. The interval is allowed for the receipt of any memorandum
of dissent or reservation, together with a brief statement of his reasons, that
any member may wish to express; and such memorandum of dissent or reserva-
*Occasional Paper 105, National Bureau of Economic Research, New York,
1007. This report is one of a series emerging from an investigation of pension
plans made possible by grants to the National Bureau from the Maurice and
Laura Falk Foundation and the Life Insurance Association of America. These
organizations are not, however, responsible for any of the statements made or
views expressed.
f Northwestern University.
~ National Btireau of Economic Research.
115
PAGENO="0124"
116 OLD AGE INCOME ASSURANCE-PART V
tion shall be published with the manuscript if he so desires. Publication does not,
however, imply that each member of the Board has read the manuscript, or that
either members of the Board in general, or of the special committee, have passed
upon its validity in every detail.
7. A copy of this resolution shall, unless otherwise determined by the Board,
be printed in each copy of every National Bureau book.
(Resolution. adopted October 25, 1926, as revised February 6, 1933, and
February 24, 1941.)
PAGENO="0125"
ADVISORY COMMITTEE ON PENSION STUDIES
In the planning and review of its pension studies, the National Bureau
has benefited from the advice and guidance of this committee. The com-
mittee's concurrence with the views expressed in this report, however, is
not to be assumed. The members of the committee are:
SOLOMON FABRICANT (chairman), National Bureau of Economic Research
ROBERT M. BALL, Social Security Administration
DORRANCE C. BRONSON, The Wyatt Company, Washington, D.C.
J. DOUGLAS BROWN, Princeton University
GEORGE B. BUCK, JR., New York, N.Y.
ARTHUR L. COBURN, JR., Old Colony Trust Company
JOHN J. ConsoN, formerly of McKinsey & Co.
FRANK G. DICKINSON, North Illinois University
F. F. FAURI, University of Michigan
MILTON FRIEDMAN, University of Chicago
GEORGE GARVY, Federal Reserve Bank of New York
WILLIAM C. GREENOUGH, Teachers Insurance and Annuity Association and College
Retirement Equities Fund
CHALLIS A. HALL, Jn., Yale University
RALPH W. HEMMINGER, Bankers Trust Company
ALBERT J. HETTINGER, JR., Lazard Freres and Company
REINHARD A. HOHAUS, formerly of Metropolitan Life Insurance Company
CHARLES L. JACOBSON, JR., formerly of State of Wisconsin Investment Board
E. GORDON KEITH, University of Pennsylvania
BENJAMIN B. KENDRICK, formerly of Life Insurance Association of America
LANE KIRKLAND, AFL-CIO, Washington, D.C.
I\IURRAY W. LATIMER, Washington, D.C.
J. D. LOCKTON, General Electric Company
DAN M. MCGILL, University of Pennsylvania
IDA C. MERRIAM, Social Security Administration
ROBERT J. MYERS, Social Security Administration
VITO NATRELLA, formerly of Securities and Exchange Commission
JAMES J. O'LEARY, formerly of Life Insurance Association of America
JOSEPH A. PECHMAN, Brookings Institution
RAY M. PETERSON, formerly of Equitable Life Assurance Society
IRENE L. ROTHSCHILD, formerly of California State Employees' Retirement
Systems
LAWRENCE H. SELTZER, formerly of Wayne State University
117
PAGENO="0126"
CONTENTS
Page
ACKNOWLEDGEMENTS 120
FOREWORD, by Roger F. Murray 121
PART I. Growth in Multiemployer and~TJnion Pension Funds, 1959-64, by
H. Robert Bartell, Jr.
1. Introduction 123
2. Description and Size of Funds 123
3. Portfolio Composition 131
4. Comparison With Portfolios of Corporate Pension Funds 134
5. Summary
APPENDIX I. Estimating Procedures 140
PART II. Pension Funds of Nonprofit Organizations, by Elizabeth T.
Simpson.
6. Introduction 142
7. Size of Funds: Variations by Type of Plan and Type of Nonprofit
Organization 143
8. Growth Rates 151
9. Portfolio Composition 154
10. Conclusions 161
APPENDIX II. Sources and Limitations of Data 163
TABLES
I-i. National and International Unions Participating in Multiemployer
Pension Plans, by Number of Workers Covered, Comparison of
BLS and NBER Surveys 127
1-2. Multiemployer Pension Plans, by Number of Workers Covered,
Comparison of BLS and NBER Surveys 127
1-3. Multiemployer and Union Pension Funds, Assets in 1959, Per
Worker Covered, by Year of Organization 129
1-4. Concentration of Assets in Corporate and in Multiempioyer and
Union Pension Funds 129
I-S. Ten Largest Self-Insured Multiemployer and Union Pension
Funds, 1959 130
1-6. Assets of Multiemployer and Union Pension Funds, by Industry
Group, 1959 130
1-7. Portfolio Composition of Multiemployer and Union Pension Funds,
1959-64 131
1-8. Multiemployer and Union Pension Funds, Sources and Uses of
Funds, 1960-64 132
1-9. Portfolio Composition of Atypical, Selected, and All Multiemployer
and Union Pension Funds, 1964 132
1-10. Portfolio Composition of Multiemployer and Union Pension Funds,
by Size of Fund, 1959 133
I-il. Investment Portfolios of Selected Multiemployer and Union Pen-
sion Funds, 1964, and Corporate Pension Funds, 1959 and 1964. 135
1-12. Investment Portfolio of Selected and Bank-Administered Multi-
employer and Union Funds and Corporate Funds, 1959 136
1-13. Portfolio Composition of ~`Iu1tiemployer and Union Pension Funds,
by Union; All Unions With Funds Totaling $15 million or More,
1959 137
1-14. Number of Multiemployer and Union Funds in Sample Holding
Selected Types of Assets, 1959-64 138
118
PAGENO="0127"
OLD AGE INCOME ASSURANCE-PART V 119
TI-i. Percentage Distribution of Pension Funds by Type of Plan, Non- Page
profit Organizations and All Other Private Groups, 1958-64 1-13
11-2. Percentage Distribution of Pension Funds of Nonprofit Organiza-
tions, by Purpose of Organization, and Type of Plan, End of
1960 144
11-3. Pension Coverage Rates of Nonprofit Organizations, by Purpose of
Organization, End of 1960 145
11-4. Annual Growth Rates of Noninsured Pension Funds of Nonprofit
Organizations and Corporations, 1946-64 152
11-5. Annual Growth Rates of TIAA, CREF, and Agency Life Insurance
Company Pension Reserves, 1946-64 153
11-6. Percentage Distribution of Portfolios of Noninsured Pension Funds
of Nonprofit Organizations, Selected Years, 1951-64 155
11-7. Percentage Distribution of Portfolios of Noninsured Corporate
Pension Funds, Selected Years, 1951-64 155
11-8. Ratios of Market to Book Values for Security Investments of Non-
insured Pension Funds of Nonprofit Organizations and Corpora-
tions, Selected Years, 1951-64
11-9. Percentage Distribution of Portfolios of Insured Pension Funds,
Agency Insurance Companies and TIAA-CREF, Selected Years,
1951-64 156
11-10. Percentage Distribution of Portfolios of Total Pension Funds for
Nonprofit Organizations, Year End, 1958-64 157
TI-Il. Uses of Total Pension Funds by Nonprofit Organizations in the
Capital Markets, 1959-64 157
11-12. Ratio of Gross Income From Interest and Dividends to Book Value
of Total Assets for Noninsured Pension Funds of Corporations
and Nonprofit Organizations and for Agency Insurance Com-
panies and TIAA, 1958-64 159
11-13. Ratio of Gross Income From Interest and Dividends Plus Realized
and Unrealized Gains and Losses to Market Value of Total
Assets for Noninsured Pension Funds of Corporations and Non-
profit Organizations and CREF, 1958-64 160
CHART
TI-i. Ratio of Gross Income from Interest and Dividends Plus Realized
and Unrealized Gains and Losses to Market Value of Total
Assets for Noninsured Pension Funds of Corporations and Non-
profit Organizations and CREF, 1958-64 159
PAGENO="0128"
ACKNOWLEDGMENTS
Roger F. Murray, director of the National Bureau's pension studies,
has been of inestimable help to both authors at all stages of their efforts
from initial idea to final fruition. Also helpful were the suggestions
of the two other members of the National Bureau's staff reading com-
mittee: Daniel M. Holland and Ralph L. Nelson.
We are also indebted to members of the pension studies' advisory
committee, especially to Dorrance C. Bronson, Arthur L. Coburn, Jr.,
George Garvy, William C. Greenough, Ralph W. Hemminger, Ida
C. Merriam, Robert J. Myers, and Ray M. Peterson. We also thank
Robert E. Graham, Jr., Office of Business Economics; Donald L. Hib-
bard, Board of Pensions, United Presbyterian Church; Francis P.
King, TIAA; H. S. Linfleld, Jewish Statistical Bureau; L. R. Linsen-
mayer, Bureau of Labor Statistics; Joseph J. Melone, WTharton
School; members of the Office of Welfare and Pension Plans of the
Department of Labor; and others too numerous to mention who pro-
vided information of various sorts. The studies could never have been
completed without their aid, and we are most grateful to them. We
also wish to thank Joseph A. Beirne, Albert J. Hettinger, Jr., and
Theodore 0. Yntema of the National Bureau's Board of Directors for
their review of the manuscript.
The studies were financed by grants from the Maurice & Laura
Falk Foundation and the Life Insurance Association of America.
Some of Dr. Bartell's research was also financed and encouraged in
other ways by the Division of Research, Graduate School of Business
Administration, University of California, Los Angeles, and by the
United States Savings and Loan League.
We both wish to thank James F. McRee, Jr., for his excellent edito-
rial suggestions. Miss Simpson appreciates the skill of H. Irving For-
man in drawing the chart and the painstaking statistical assistance
rendered by Nadeschda Bohsack and Antonette Burgar.
The authors acknowledge responsibility for errors that still remain
in these papers.
H. ROBERT BARa~LL, Jr.
ELIZABETH T. SIMPsoN.
120
PAGENO="0129"
FOREWORD
Gaps in available data are not an unusual discovery in the explora-
tion of a field as broad as public and private pensions. Early in the
National Bureau's research project on the economic aspects of pen-
sions, it became evident that data on the relatively new multiemployer
and union pension programs were not available in enough detail to
indicate their size, rate of growth, and portfolio composition. Com-
parable data were also incomplete even for the long-established plans
of nonprofit organizations. Lack of convenient sources of informa-
tion, rather than any lack of interest, seemed responsible for these
particular gaps in the otherwise comprehensive measurement of
saving and investment in the field of private pensions.
This third report from the pension research project seeks to remedy
these deficiencies in basic information.1 The results have been espe-
cially useful to Government agencies in revising their time series
data on the private pension structure. The findings have also materi-
~ily aided in the displacement of widely circulated but erroneous im-
pressions by factually supported insights into significant trends.
In recent years, the rapid growth of multiemployer and union pen-
sion funds was expected by some observers to create a new economic
power in the hands of union officials. It was suggested, for example,
that a degree of control over employers might be exercised; that the
character of union organization might be changed by newly created
financial management responsibilities; and that great stimulus to the
financing of middle-income housing might be provided by the in-
vestment of vast sums from these sources. Bartell's survey shows that
aggregate assets amounted to less than $1.3 billion in 1959 and grew
to $3.0 billion by 1964. Although this is a rapid rate of growth, the
aggregate amount in 1964 only represented approximately 4 percent
of the book value of all industrial pension fund accumulations. Fur-
thermore, he shows that `the funds were generally invested in accord-
ance with traditional standards of trusteeship.2
Despite vigorous growth in coverage and assets, therefore, it is easy
to exaggerate the role of multiemployer and union pension funds in
the total structure of public and private pension programs. In contrast,
there is a tend'ency to regard similar programs of nonprofit organiza-
tions as of diminishing importance because of the relatively mature
stage of their development. To appraise their role in accurate perspec-
tive, a different approach had to be taken by Elizabeth Simpson in her
study. The principal questions related to nonprofit institutions outside
of the well-known group of major Protestant churches with long-estab-
I The two preceding reports are Phillip Cagan's The Effect of Pension Plans on &ggre-
gate $aving: Evidence From a ~5ampZe $urvey, Occasional Paper 9,5, New York, NBER,
1965. and Daniel M. Holland's Private Pension Funds: Projected arowtls, Occasional Paper
97, New York, NBER, 1966.
I The impact of new pension plan responsibilities on union organization and leadership
was further explored by Bartell in "Unions and Pension Funds," unpublished Ph. D. dis-
sertation, Columbia University, 1962.
121
83-200-67-pt. 5-9
PAGENO="0130"
122 OLD AGE INCOME ASSURANCE-PART V
lished retirement benefit programs. What are the trends in the cover-
age of employees of other religious bodies? Medical and health, cul-
tural, and social service organizations have been slow to extend pension
coverage. Will there be well-sustained future growth in programs for
these nonprofit activities?
Aided by the rapid growth of educational institutions and the exten-
sion of pension coverage, it appears that the nonprofit sector, contrary
to what might be expected, faces an extended period of strong growth
in asset accumulation. To support; this conclusion Miss Simpson has
surveyed trends in a wide range of activities. Complete coverage was
beyond the scope of this study, but an attempt was made to identify
most of the areas of potential importance.
These studies serve a twofold purpose: to assemble the facts in areas
where data were incomplete and to identify trends which will affect the
future structure of pension arrangements. In the process of serving
these two objectives, it wa.s possible to correct some exaggerated ex-
pectations regarding the adolescence of multiemployer and union
pension funds. Similarly, it was found that a vigorous middle age
rather than approaching senility characterized the programs of non-
profit institutions. For these reasons, the Bartell and Simj~son studies
fit together in this volume of research studies dealing with the eco-
nomic aspects of pensions. The facts and interpretation which they
supply serve to illuminate not just the dark corners but the main
stream of developments in the dynamic evolution of the public and
private pension structure.
RoG1~n F. Mtmi~AY.
PAGENO="0131"
PART I: GROWTH IN MTJLTJEMPLOYER AND UNION
PENSION FUNDS, 1959-64
BY H. ROBERT BARTELL, JR.
1. INTRODUC~rON
Corporate pension funds, which cover the employees of a single
company or a group of financially related companies, are the most
carefully reported sector of the private pension structure. The aggre-
gate estimates of assets and portfolio composition, which are provided
regularly for this sector by the Securities and Exchange Commission,'
are derived from a questionnaire submitted to a large sample of U.S.
corporations. In using the corporation (and primarily larger corpo-
rations) as the unit of inquiry, the Securities and Exchange Commis-
sion survey does not include funds which cover the employees of two
or more financially unrelated corporations or unincorporated business-
es,. usually known as multiemployer funds. Only in recent years has
the SEC attempted to consolidate these and nonprofit organization
funds with corporate noninsured pension funds to estimate a total of
assets of private noninsured pension funds.
The Social Security Administration's yearly report on employee
benefit plans regularly includes an estimate of the reserves of all
private pension plans (`both insured and noninsured), including "pay-
as-you-go, multiemployer, and union administered plans, those of
nonprofit organizations and railroad plans supplementing the Federal
railroad retirement program." 2 Thus, the estimate for multiemployer
and union funds, as in SEC tabulations,. is combined with figures for
the assets of other private pension funds. However, the rapid growth
of these funds in recent years justifies a more detailed description of
their attributes. Furthermore, recent attention being paid to the effect
of union policies on pension fund investments prompts a closer look
at the portfolio composition of those funds-rnultiemployer and
union-where labor unions have some measure of control over invest-
ment policy.
2. DEscRIPTIoN AND SIZE OF FUNDS
The history of labor unions in the United States reveals that one
of the major reasons for the organization of unions in `the late 1800's
was the provision of accident, death, and retirement benefits for certain
classes of workers who were unable to obtain them from ordinary in-
surance companies at rates which the unions considered equitable.
The early unions in the railroad, printing, and construction industries
were good examples of the organizational importance of benefit pro-
Securities arid Exchange Commission, Survey of Corporate Pension Funds, 1951-54,
Washington, 1956, and annual surveys published in its Statistical Bulletin.
2 A. M. Skolnik, "Ten Years of Employee Benefit Plans," Social Security Bulletin, April
1966, pp. 3-19.
123
PAGENO="0132"
124 OLD AGE INCOME ASSURANCE-PART V
grams. Until the first part of the 1900's, the beneficial activities of
these unions were limited to providing death and disability benefits
and homes for aged members.
Although there were several earlier attempts at providing regular
superannuation benefits, the system of the International Typograph-
ical Union, established in 1908, was the first which covered any large
number of workers and which is still in existence. This was followed
by plans established by the Brotherhood of Locomotive Engineers in
1912, the Bricklayers, Masons, and Plasterers in 1915, the Railway
Conductors in 1917, the Bridge and Structural Iron Workers in 1918,
the Locomotive Firemen and Enginemen in 1920, the Railroad Train-
men in 1923, the Printing Pressmen in 1924, and the Brotherhood of
Electrical Workers in 1928. These systems were financed by assess-
ments on the membership and were, except in the case of the railway
brotherhoods, compulsory. Eligibility for pension benefits was based
upon age, length of membership in the union, and inability to get em-
ployment in the trade. Benefits were looked upon as a form of charity
rather than as a right earned for long service in the trade.
Union pension funds had assets of $11.5 million in 1929, accord-
ing to Latimer,3 and only 10 unions had such funds. The American
Federation of Labor reported 14 unions paying "old-age benefits" in
1929, and 20 in 1931, paying just over $6 million in benefits. During the
1930's a number of union pension programs were dropped or their
funds exhausted, so that by 1940 only 12 unions were paying benefits
and the PaYme1~tS caine to $1.7 million.4
Beginning in the 1940's unions shifted their efforts from financing
retirement benefits by member assessments to obtaining employer sup-
port in providing pension payments to retired workers. In the larger,
financially secure firms with relatively stable labor forces, a company-
wide pension plan provided the best compromise between union de-
mands for adequate retirement payments and management's desire to
limit the costs of such benefits. Thus most negotiated pension plans
were restricted to the employees of a single company.
However, workers in industries with small firms and frequent trans-
fers of employees between firms also had need for retirement security
which could not be economically provided by single-company pension
plans. The first areawide pension system supported by employer con-
tributions was initiated in 1929 by Local 3 of the Brotherhood of
Electrical Workers and the Electrical Contractors Association of
~4ew York City. A similar plan for Electricians and the Electrical
Employers Association of St. Louis was agreed upon in late 1929, but
the agreement was voided by the Missouri Supreme Court in 1931 as
a conspiracy in restraint of trade. During World War II, wage con-
trols encouraged unions to seek contract improvements in the form
of welfare benefits. The Amalgamated Clothing Workers and the In-
ternational Ladies' Garment Workers' Union negotiated multiemployer
pension systems in the garment trades, beginning with Local 324 of
the ACWA in 1943. Benefits are financed entirely by employer con-
tributions, but the funds are administered and the level of benefits
determined solely by the union.
Murray. W. Lathner, Trade Union Pension ,S'ystem.s, New York, 1932.
`H. A. Millis and R. H. Montgomery, Organized Labor, New York, 1945, Table 6, p. 334.
See also Bureau of Labor Statistics, Beneficial Activities or American Trade TJniOfl8,
Bulletin 465, 1928.
PAGENO="0133"
OLD AGE INCOME ASSURANCE-PART V
125
The Taft-Hartley Act, passed in 1947, required that all plans estab-
lished subsequent to January 1, 1946, which involved negotiated em-
ployer contributions to a -fund be administered jointly by employers
and employees. Thus the gTeat majority of multiemployer funds are
jointly administered.
The bulk of negotiated single-company funds are established under
collective-bargaining agreements which specify the benefits to be re-
ceived upon retirement. Since the entire cost of the benefits (above
employee contributions, if any) is borne by the corporation, invest-
ment of accumulated funds is left to the discretion of management.
Virtually all multiernpioyer plans are established under collective-
bargaining agreements which specify the amount of contributions to
be made by employers and employees, and the benefits received upon
retirement are dependent upon the size of the fund. Since the size
of the fund is a function not only of contributions but also of invest-
ment yield, both union and employers are concerned with the man-
agement of the accumulated fund, especially as it may affect the level
of benefits.
In summary, union funds are relatively old, administered solely
by the union, and few in number. By contrast, rnultiemployer funds
are newer, more numerous, and jointly administered by union and
management.
SOURCES OF DATA
Until the passage, in 1958, of the Federal Welfare and Pension Plans
Disclosure Act, consistent and comparable data on multiemployer and
union pension funds were not available for all such funds.5 New York
State, since 1956, has had a reporting and disclosure act which requires
all jointly administered funds covering more than 25 workers in the
State to file comprehensive annual reports on the financial condition
of the fund, including a list of all securities held. Reports submitted
are subjected to a desk audit to check their internal consistency, and
State agencies are authorized to make field audits of the funds when
necessary and at least once every 5 years. Therefore, information ob-
tained from these records is ample for investigations of asset size and.
portfolio composition and sufficiently accurate for the generation of
aggregate figures. However, while the State disclosure act has wide
coverage of multiemployer funds, these funds are either national in
scope or cover employees in New York State and adjoining areas.
Thus funds of a local or areawide character in other parts of the coun-
try are excluded. Funds administered solely by unions are completely
exempt from reporting requirements in New York.
Before the passage of Federal legislation, several other States had
reporting acts which covered various types of plans-mostly those
jointly administered.~ However, the reports filed with the Office of
Welfare and Pension Plans of the Department of Labor include vir-
tually all multiemployer and union pension funds in operation. The
annual financial data on fund assets are in summary form and, unlike
For estimates back to 1945, based on estimates for 1957, from New York State reports,
see Raymond W. Goldsmith, Robert E. Lipsey, and Morris Mendelson, Studies in the
National 8alance Sheet of the United States, Princeton University Press for NBER, 1963,
vol. II, table III-5j---2, pp. 182-183.
See National Industrial Conference Board, Management Record, July-August 1958,
pp. 246 if., for a summary of State disclosure laws.
PAGENO="0134"
126 OLD AGE INCOME ASSURANCE-PART V
the New York State reports, do not include a list of specific securities
held. In addition, reports filed under the Federal act are not checked
for accuracy by the Labor Department. Nevertheless, for the purpose
of providing aggregate estimates of assets and broad classes of invest-
ments, the Federal reports are satisfactory since most are audited by
a public accountant before being submitted.
The Bureau of Labor Statistics reported that 798 multiernployer
pension plans, covering approximately 3.3 million active and retired
workers, had been filed by the spring of 196O.~ There is no public list
of these funds, but tables I-i and 1-2 indicate that practically all the
funds of major significance are contained in the present survey, and
that in addition a broad sample of the smaller funds has been included.
TABLE -1.-NATIONAL AND INTERNATIONAL UNIONS PARTICIPATING iN MULTIEMPLOYER PENSION PLANS BY
NUMBER OF WORKERS COVERED-COMPARISON OF BLS AND NBER SURVEYS
Number of plans
Union
BLS NBER
100,000 workers and over:
Carpenters 18 18
Clothing (ACWA) 14 12
Electrical (IBEW) 32 27
Garment, Ladies 48 53
Mine (UMW) (excluding District 50) 2 2
Teamsters 121 1t4
50,000 and under 100,000 workers:
Bakery (BCW) 8 5
Hod Carriers 23 25
Hotel 13 14
Longshoremen (ILA) 21 17
Meat Cutters 31 31
Plumbing 63 56
Retail Clerks 11
25,000 and under 50,000 workers:
Bakery, Amercan (ABCW) 5 3
Building Service 8 8
Engineers, Operating 18 17
Iron 20 15
Maritime (NMU)
Painters 23 24
Retail, Wholesale (RWDSU) 12 13
Sheet Metal 24 22
Upholsterers 1 1
5,000 and under 25,000 workers:
Actors (AFTRA) 1 1
Automobile 6 6
Bookbinders 6 5
Brewery 6
Bricklayers 30 29
Electrical (IUE)
Furniture 6 6
Hatters 23 17
Lathers
Leather Goods 5 2
Longshoremen and Warehousemen (ILWU) 5 5
Machinists 10 10
Marine Engineers
Masters, Mates 2 2
Musicians 1 1
Plasterers 8 9
Printing Pressmen 14 10
Pulp 6 4
Shoe Workers, United 1 1
Stage (IATSE) 13 10
Textile Workers (TWUA) 1 1
Toy Workers 1 1
Typographical 13 10
° Bureau of Labor Statistics, .lfulticnsployer Pension Plans Under Collective Bargaining,
~jaring 1960, Bulletin 1326, June 1962.
PAGENO="0135"
OLD AGE INCOME ASSIJRANCE-PART v 127
TABLE 1-1.-NATIONAL AND INTERNATIONAL UNIONS PARTICIPATING IN MULTIEMPLOYER PENSION PLANS BY
NUMBER OF WORKERS COVERED-COMPARISON OF BLS AND NBER SURVEYS-Continued
Number of plans
Union
BLS NBER
1,000 and under 5,000 workers:
Asbestos 17 11
Boilermakers 2 2
Distillery 5 6
Jewelry 1 3
Leather Workers 1 1
Lithographers 1 1
Mine, District 50 3 3
Newspaper Guild 2 2
Newspaper and Mail Deliverers 1 1
Patternmakers 5 3
Photo Engravers 4 5
Roofers 4
Stereotyphrs 3 3
Watchmen's Association 3 1
Fewer than 1,000 workers:
Engineers, Technical 1 1
Firemen and Oilers 1 1
Garment, United 2
Hosiery 1 1
Machine Printers 1 1
Mailers 2 1
Marble 1 2
Mefal Polishers 1 1
Office 2 3
Radio 1 2
Shoe and Boot Workers 1
Telegraphers 1
Textile Workers (UTWA) 1 1
Not included in BLS survey:
Seafarers 2
Writers Guild 1
Industrial Workers, Allied 1
2 or more national unions 26 10
Unclassified (AFL-CIO directly affiliated and indopendent locals, and unknown) 12 9
All plans 798 715
Source: Bureau of Labor Statistics, "Multiomployer Pension Plans Under Collective Bargaining, Spring 1960, Bulletin"
No. 1326, June 1962, table 4, and National Bureau of Economic Research survey (1959) conducted by the author.
TABLE 1-2.-MULTI EMPLOYER PENSION PLANS BY NUMBER OF WORKERS COVERED-COMPARISON OF BLS AND
NBER SURVEYS
Number of workers covered I
Number
BLS
of plans
NBER
All plans 798
715
tinder 100
100 and under 500
500 and under 1,000
1,000 and under 5,000
5,000 and under 10,000
10,000 and under 25,000
25,000 and under 50,000
50,000 and under 100,000
100,000 and over
Coverage not determined 2
54
230
154
257
51
29
15
2
6
29
175
132
211
51
33
14
3
5
62
Source: Bureau of Labor Statistics, "Multiemployer Pension Plans Under Collective Bargaining, Spring 1960," Bulletin
No. 1326, June 1962, ttable 2, and National Bureau of Economic Research survey (1959) conducted by the author.
1 BLS survey includes annuitants; NBER survey does not.
2 For plans where coverage was not reported, BLS estimated coverage at 1,200. This would cause them to fall into the
1,000 and under 5,000 category.
PAGENO="0136"
128 OLD AGE INCOME ASSURANCE-PART V
Since it was important to secure estimates of trends in assets and
portfolio composition of self-insured multie.mployer and uni'on funds,
a stratified sample of funds covered in the initial survey of 1959 was
drawn and used to provide estimates for subsequent years. A descrip-
tion of the sampling method used is provided in appendix I.
TOTAL ASSETS AND GROWTH
The total assets of rnultiernployer and union pension funds were
approximately $1,270 million at book value in 1959. Estimates through
1964 are presented below in comparison with corresponding data for
single-employer industrial noninsured pension funds:
Multiemployer an
d union funds
Single-emplo
yer funds
Book value (millions)
Percent increase
Bsok value (millions)
Percent increase
1959
1960
1961
1962
1963
1964
$1,270
1,548
1,856
2,209
2,601
3,040
(1)
21.9
19.9
19.0
17.8
16.8
$26,667
30,343
34,302
38,233
42,427
47,285
14.8
13.8
13.0
11.5
11.0
11.5
I Not available.
Growth in a pension fund's assets is particularly rapid, of course, in
the early years following its establishment. A constant amount of con-
tributions will cause a fund's assets to double during its second year
of operations and increase by 50 percent during the third year. Bene-
fit payments are usually low in the early years, and provision for the
gradual funding of past and current service liabilities means that re-
ceipts will exceed expenditures from the fund for a number of years.
The history of a number of multiernployer funds shows that extension
of coverage to new groups of employees and increases in contemplated
benefits proceed rapidly after the fund is established and participants
are persuaded that pension coverage is feasible. This "demonstration
effect" also encourages other employer and employee groups to estab-
lish their own funds. In the aggregate, growth in fund assets is derived
from funding of current and past service liabilities, establishment of
new funds, and extensions of coverage and contribution increases in old
ones. The effect of these factors on overall growth rates is most dra-
matic where newer funds predominate.
Most multiemployer pension plans have been established in recent
years. According to the Bureau of Labor Statistics study, over 60 par-
cent of the plans covering 43 percent of the workers under multi-
employer pension plans were started in the 5-year period from 1955
to 1959. Less than 8 percent of the plans were over 10 years old in
1960.°
The effect. of funding on per capita. asset growth is demonstrated by
table 1-3, which shows assets per active employee by age of the. fund
for a sample of 360 self-insured multiemployer and union funds for
In this paper the terms "self-insured" and "noninsured" are used interchangeably.
Insured plans are included In tables I-i and 1-2, but have been omitted from the remainder
of the study. See app. I and BLS Bulletin No. 1326 for number of Insured and self-insured
plans.
e Ibid., table 1, p. 98.
PAGENO="0137"
OLD AGE INCOME ASSURANCE-PART V 129
which coverage, total assets, and date of establishment were available.
For funds established after 1947, assets per employee rise continu-
ously as the age of the fund increases. The sharp drop in assets per
employee in funds started before 1948 reflects the influence of the large
Clothing Workers', Electrical Workers', and Mine Workers' funds,
which do not provide for funding of past service liabilities on an
actuarial basis. If the funds of the ACWA, IBEW, and UMW are
omitted from the funds established in 1946-47, the assets per em-
ployee of the remaining funds exceed $1,000.
TABLE I-3.---MULTIEMPLOYER AND UNION PENSION FUNDS-ASSETS IN 1959 PER WORKER COVERED, BY YEAR
OF ORGANIZATION
Year organized
Number of
funds in
sample
Assets per
worker csvered
1958 and 1959
89
$130
1956 and 1957
89
265
1954 and 1955
59
443
1952 and 1953
46
523
1950 and 1951
1948 and 1949
47
13
1,023
1,201
1946 and 1947
10
427
1945 and earlier
7
313
Source: NBER survey.
CONCENTRATION OF ASSETS
The assets of multiemployer and union pension funds, like those of
corporate funds, are heavily concentrated in a few large funds. Of
the funds included in the National Bureau survey, those having more
than $10 million in total assets accounted for less than 4 percent of
the funds but for more than 53 percent of the assets of all funds. On
the other end of the scale, 53 percent of the funds had under $`/2 mil-
lion each in total assets, and they held less than 5 percent of the assets
of all multiemployer and union funds. Table 1-4 shows the data on
asset concentration in greater detail.
TABLE 1-4.-CONCENTRATION OF ASSETS IN CORPORATE AND IN MULTIEMPLOYER AND UNION PENSION FUNDS
[In percent of totall
Size of fund (thousands)
Corporate
funds,' 1957
Multiemploye
r and
1959
union funds,
Number of
Total assets 2
Number of
Total assets 2
funds
funds
$100,000 and over
1. 5
55~ 3
0
0
$10,000 to $99,999
12. 3
28. 3
3. 8
53. 6
$5,000 to $9,999
$1,000 to $4,999
$500 to $999
9. 1
25. 5
13.0
6. 0
6. 9
1.2
2. 6
26. 9
13.3
8. 1
29. 4
4.3
Under $500
38. 6
2. 3
53. 4
4. 6
Source: Corporate funds, SEC data, calculated from P. P. I-Iarbrecht, "Pension Funds and Economic Power," New
York, 1959, table 30, p. 224; multiemployer and union funds, in BER survey.
1 Noninsured single-employer pension funds of business corporations.
.2 Assets are recorded at book value.
The 10 largest funds in 1959 had $478 million in assets, or about 38
percent of the total. The funds and their assets are shown in table I-S.
PAGENO="0138"
130 OLD AGE INCOME ASSTJRANCE-PART V
TABLE 1-5-10 LARGEST SELF-INSURED MULTIEMPLOYER AND UNION PENSION FUNDS, 1959
[In thousands of dollarsj
No.
Name of fund
Total
assets
1
2
3
4
5
6
7
8
9
10
IBEW pension benefitfund (union)
United Mine Workers welfare and retirement fund
Teamsters' Central States, Southeast, and Southwest areas pension fund
Amalgamated insurance fund (retirement)
National electrical benefit fund of the National Employees Benefit Board for the Electrical Contracting
Industry
Retirement fund of the New York Dress Joint Board of the ILGWU
Retirement fund of the New York Cloak Joint Board of the ILGWU
The 65 security plan pension fund (District 65, Retail, Wholesale & Department Store Union)
National Maritime Union Pension trust
Amalgamated Cotton Garment & Allied Industries fund (retirement)
$85,010
1 82, 945
61, 333
55, 765
46,000
40, 396
29, 612
26, 449
25,801
24,682
Note
: Assets are shown at book value.
Source: NBER survey.
1 Estimated.
TOTAL ASSETS BY MAJoR INDUSTRY GROUPS
Classification of individual funds by industry is difficult because
some funds include employees working in several different industries.
Therefore, the data presented in table 1-6 on assets by industry of
employment can represent, at. best, only approximations. However,
they show rough similarities to the BLS data on coverage by industry.
About 60 percent of coverage and assets are in nonmanufacturing in-
dustries. High concentrations of both coverage and assets are found
in the apparel, construction, and motor transportation industries.'0
TABLE -6.-ASSETS OF MULT1EMPLOYER AND UNION PENSION FUNDS, BY INDUSTRY GROUP, 1959
Assets 1 (millions)
Percent of total
All industries
Manufacturing
Food and kindred products
Apparel and other finished textile products
Printing, publishing, and allied industries
Leather and leather products
Metalworking
Other manufacturing
Nonmanufacturmg
Mining
Contract construction
$1, 270. 3
512.5
124. 7
246. 7
61. 9
6. 7
36. 2
36. 3
752. 2
85. 1
278. 0
100. 0
40.3
9. 8
19. 4
4.9
. 5
2. 8
2. 9
59. 2
6. 7
21. 9
Motor transportation
Watertranspsrtation
Wholesale and retail trade
Services
Motion pictures and recreation
Other nonmanufacturing
Notciassified
133. 2
110.5
74. 1
42. 4
28. 5
. 3
~
10. 5
8.7
5. 8
3. 3
2. 2
(5)
.4
Note: Figures are rounded and may not add to totals.
Source: NBER survey.
1 Assets at book value.
2 Less than 0.05 percent.
In addition to concentration in certain industries, there is also a
virtual atsence of multiemployer and union pension funds in durable
goods manufacturing, railroads, and public utilities other than trans-
portation. These observations underline the proposition that multi-
10Ibid., table 3, p. 99.
PAGENO="0139"
OLD AGE INCOME ASSTJRANCE-PART V 131
employer pension systems are particularly suited to industries where
small firms predominate and where employment mobility within an
industry is a prevalent ctiara-cteristic.
3. Po~oi~io CoMPoSITIoN
The assets of multiemployer and union pension funds are invested
in a variety of financial instruments. For the most part, the instruments
are the same as those found in corporate pension funds: Federal Gov-
ernment obligations, corporate stock and bonds, and mortgages.
However, the distribution of assets among the various investment
categories and the proportion held uninvested; that is, in cash, show
significant differences between corporate and multiemployer and
union funds when viewed in the aggregate. A closer look, though,
indicates that these differences are not great, that they can be partly
explained by differences in size and operating characteristics, and
that they are diminishing.
Table 1-7 presents the aggregate portfolios of multiemployer and
union pension funds for t'he years 1959 through 1964. Multiemployer
and union funds, as -a whole, have a greater proportion of their assets
in cash, Government securities, and mortgages (40.2 percent versus
10.9 percent for `corporate funds in 1964) and a smaller proportion
in corporate stock `and bonds (56.3 percent) as compared to corporate
pension funds (84 percent) *1~ However, over the 6 years, the multi-
employer and union funds have sharply reduced the propOrtion of
their iortfolio in Government bonds and increased investments in
corporate securities. A continuing difference between the two types of
funds appears to be the greater emphasis of multiemployer and union
funds on mortgage investment, although corporate funds also increased
their investment in this field in recent years.
TABLE -7.-PORTFOLIO COMPOSITION OF MULTIEMPLOYER AND UNION PENSION FUNDS, 1959-64
1959
1960
1961
1962
1963
1964
Total assets (millions)
Percent of book value:
$1,270
$1,548
$1,856
$2,209
$2,601
$3,040
Cash
6.5
5.2
5.5
6.1
6.5
6.2
U.S. Government securities
34.0
26. 1
21.0
18.6
16.0
14.8
Corporate and other bonds
28.0
31.5
32. 1
31.7
31.9
31.8
Preferred stock
1.7
1.8
1.6
1.4
1.2
1.0
Common stock
14.0
17.5
20.4
22.2
22.4
23.5
Mortgag
Other investments
12.3
1.4
15.1
1.2
16.4
1.2
17.0
1.4
17.9
2.5
19.2
1.9
Other assets
2.1
1.7
1.7
1.6
1.6
1.6
Source: NBER survey for 1959, with subsequent y
ears estima
ted as desc
ribed in a
pp. I.
FLOWS OF NEW FUNDS
The changes in multiemployer and union pension fund portfolios are
more sharply highlighted by the figures in table 1-8. "New" funds have
been channeled into mortgages and common stock at a faster rate in the
years since 1959 than the average for the years prior to 1959. Corporate
and other bonds gathered a larger share of new funds in 1960 and 1961
than previously, -but the-portfolio of these securities h-as grown at about
~ See Securities and Exchange Commission, ~Statistical Bulletin, June 1965 and earlier.
PAGENO="0140"
132 OLD AGE INCOME ASSURANCE~PART V
the same rate as total assets in more recent years. Multiernployer and
union funds were net sellers of U.S. governments in 1960 and 1~61. and
have added rather small amounts of these securities to portfolios s~ince.
Thus, assuming no change in the distribution of additions to port-
folios m the future, the aggregate holdings of multiemployer and union
pension funds will continue to shift toward a larger proportion of
common stock and mortgages for several years to come.
TABLE I-8.-MULTIEMPLOYER AND UNION PENSION FUNDS-SOURCES AND USES OF FUNDS, 1960-64
1960 1961 1962 1963 1964
Millions sI dollars
Sources: Net asset growth 278 308 353 392 439
Uses:
Net acquisition of financial assets 270 303 349 386 431
Cash -2 23 32 33 22
U.S. Government securities -27 -16 21 7 32
Corporate and other bonds 131 109 104 129 137
Preferred stock 5 3 2 -1 (1)
Common stock 93 108 111 94 130
Mortgages 78 71 70 91 117
Other investments (1) 5 9 33 -7
Ott;er assets (1) 5 4 6 8
Percent distribution ot uses
U.S. Government securities -9.7 -5.2 5.9 1.8 7.3
Corporate and other bonds 47.1 35.4 23.5 32.9 31.2
Common stock 33. 5 35. 1 31. 4 24. 0 29.6
Mortgages 28.1 23. 1 19.8 23.2 26.7
All other uses 1.0 11.6 13.4 18.0 5.2
1 Less than $500,000.
ATYPICAL FUNDS
Several of the very large multiemployer and union pension funds
show portfolio distributions quite dissimilar to the overwhelming
number of smaller funds. Because these larger atypical funds repre-
sent a significant proportion of all multiemployer and union fund
assets, they tend to distort the weighted average portfolios shown in
table 1-7. In order to approximate the asset distribution of the
typical or modal fund, the typical and atypical funds have been
segregated; the resulting portfolios are shown in table 1-9.
TABLE I-9.-PORTFOLIO COMPOSITION OF ATYPICAL, SELECTED, AND ALL MULTIEMPLOYER AND UNION PENSION
FUNDS, 1964
All funds Atypical funds Selected funds
Total assets (millions) $3, 040 $799 - $2, 241
Percent of book value:
Cash 6.2 10.0 4.9
Other assets 1.6 1.5 1.6
Invested assets 92.2 88.6 93.5
U.S. Government securities 14.8 19.1 13.3
Corporate and other boeds 31.8 10.7 39.2
Preferred stock 1.0 .4 1.2
Common stock 23.5 8.0 29.0
Mortgages 19.2 46.9 9.3
Other investments 1.9 3. 5 1.4
Source: NBER survey.
PAGENO="0141"
OLD AGE INCOME ASSURANCE-PART V 133
The atypical group comprises eight funds containing about 26 per-
cent of total multiemployer and union fund assets in 1964. The eight
funds cover members of five unions-the Teamsters, Ladies' Garment
Workers (two funds), Brotherhood of Electrical Workers (two
funds), Amalgamated Clothing Workers (two funds) and United
Mine Workers.
Both the Garment Workers' and Clothing Workers' funds follow
a policy of not investing in corporate stock. In addition, the Clothing
Workers exclude purchases of corporate bonds and the Garment
Workers first began acquiring these investments in 1955-56.
The two large IBEW funds and the largest Teamster fund invest
heavily in mortgages, the principal difference being that the bulk of
IBEW mortgages are Government-insured or guaranteed and on
single-family dwellings, while the Teamsters have recently favored
conventional multifamily and commercial mortgages. Over two-thirds
of the assets of these three funds are mortgage investments.
The United Mine Workers' fund reflects substantial variation in
investment policy over the years 1959-64. In the earlier years, U.S.
Government bonds were favored. This was followed by a shift to cash,
particularly bank time deposits, and in 1963 and 1964 the fund more
than doubled its holdings of eorporate stock.
The category of "selected" funds, obtained by excluding the eight
atypical funds, represents a wide range of sizes, industries, unions (the
smaller funds of the unions represented in the atypical group are in-
cluded), and portfolio policies. It is, therefore, more typical than the
aggregates heavily weighted by the extremes in portfolio policy. Fur-
thermore, the "selected" group is growing more rapidly than the atypi-
cal funds. Because of this, the aggregate portfolio of the future will
look more like the selected funds, assuming no radical change in policy
by any of the funds.
LIQUIDITY
Multiemployer and union pension funds have relatively greater cash
holdings than do corporate funds. One reason for this is that the aver-
age corporate pension fund is substantially larger than the average
multiemployer fund, and, as table 1-10 indicates, fund holdings of cash
tend to decline as a proportion of total assets as fund size increases.
TABLE 1-10.-PORTFOLIO COMPOSITION OF MULTIEMPLOYER AND UNION PENSION FUNDS BY SIZE OF FUND
1959
(In pe
rcent of book
value]
Assets
per fund (thousands)
$10,000
and over
$5,000to
$9,999
~1,000to
$4,999
$500to
$999
Under$500
Cash
3.4
6.2
6.9
17.0
31.7
Other assets
1.7
2.4
3.0
2.3
3.2
Invested assets
94.9
91.4
90.1
80.7
65.1
U.S. Government securities
38. 7
25. 4
30. 3
24. 7
24. 9
Corporate and other bonds
20. 1
45.4
37.8
35.5
20.0
Preferred stock
1.7
2.2
1.9
1.2
.9
Common stock
12.1
14.5
17.0
17.4
12.2
Mortgages
Otherinvestments
20. 8
1.4
3. 6
.3
2. 5
.8
. 4
1.6
. 9
6.2
Source: NBER survey.
PAGENO="0142"
134 OLD AGE INCOME ASSURANCE-PART V
Furthermore, rnultiemployer and union funds undoubtedly have
greater cash needs than corporate funds of comparable size. Most cor-
porate funds are level-of-benefit types, with the employer making
regular payments in order to meet qualification requirements. In the
case of rnultiemployer funds, however, the employer usually has a
fixed contribution rate and, in the short run, cannot be required to
assure the payment of specified benefits or to supply additional con-
tributions to carry the fund through occasional periods of illiquiclity.
Thus, the trustees of rnultiemployer funds probably consider the pre-
cautionary need for cash greater than the managers of corporate
funds.
The need of multieinployer funds for cash may exceed that of cor-
porate funds since most payments for administration of the system
come directly from the fund. Much of the administrative expenses of
the corporate system, such as the costs of recordkeepin~, legal and in-
vestment counsel, and actuarial advice, are usually paid directly out
of the company's accounts rather than out of the fund.
In multiemployer systems, tasks which do not exist in the corporate
systems must be performed, and these require sizable outlays of funds
at certain times. For instance, the collection of delinquent contribu-
tions may represent substantial costs to the fund, and these outlays
tend to be high when income to the fund is low because of nonpayment
of contributions.
Several of the very large multiernployer plans have high benefit
payments in relation to contributions because of age or lack of fund-
mg. A higher ratio of benefits to contributions would indicate a greater
need for liquidity for payment purposes and because of a lower
"margin of safety" of contribution inflow over benefit outgo.
It may be true that not all the cash held by multiemployer and union
funds can be attributed to greater need for liquidity. Some portion
represents certificates of deposit and savings and loan association share
accounts held for investment. Also liquidity is provided by near-cash
assets, such as Government bonds; and multiemployer and union funds
show a marked, but declining, preference for this type of asset when
compared to corporate pension funds.
OTHER ASSETS
A third factor, which, together with atypical funds and greater
liquidity needs, tends to distort portfolio comparisons between multi-
employer and corporate pension funds, is the existence of "other
assets" in the multiemployer and union funds. This category includes
accrued and delinquent contributions receivable and fixed assets, such
as office buildings and equipment. These assets would not appear on
the balance sheet of a corporate fund since the company does not have
contractual contributions which can be accrued, and it usually admin-
isters the pension system on the premises and using the facilities of
the firm.
4. COMPARISON WITH PourFoLlos OF CORPORATE PENSION FUNDS
The portfolio distributions shown in table 1-41 reflect the invest-
ment decisions of typical multieinployer and union and corporate
pension fund managers. The atypical funds of multiemployer and
PAGENO="0143"
OLD AGE INCOME ASSURANCE-PART V 135
union plans have been omitted along with cash and other asset hold-
ings. While the comparison is not perfect because some cash holdings
may represent investment decisions and some Government bond hold-
ings may reflect liquidity needs, it is superior to a comparison based
upon total assets. The portfolios of the selected multiemployer and
union funds and the corporate funds in 1964 still demonstrate signifi-
cant differences, but they are approaching each other rapidly in their
composition. As a demonstration of how fast the portfolio composi-
tion of the funds can change and how close the present portfolio of
rnultiemployer and union funds is to that of the corporate funds a
few years ago, the investment distribution of corporate funds in
1959 is also presented.
TABLE -11.-INVESTMENT PORTFOLIOS OF SELECTED MULTIEMPLOYER AND UNION PENSION FUNDS, 1964,
AND CORPORATE PENSION FUNDS, 1959 AND 1964
[In percent of book
value[
Multiemployer
and union
funds, 1964
Corpora
te funds
1959
1964
U.S. Government securities
14. 2
8. 8
5. 3
Corporate and other bonds
Preferred stock
42. 1
1.3
51.3
2.6
41.8
1.2
Common stock
31.0
30.9
42.2
Mortgages
Other investments
10. 0
1. 5
2.3
4.0
4. 5
5. 1
Note.-Figures are rounded and may not add to totals.
Source: Mu ltiemployer and union funds, NBER survey; corporate funds, calculated from data in SEC Statistical Bulletin
June 1965.
A possible explanation of the close similarity between the multi-
employer and union fund portfolio and that of the corporate funds
several years earlier is suggested by the newness of most multiemployer
funds. Investment ath~isers observe that it takes several years to change
the thinking of inexperienced trustees-both union and management-
from their natural investment conservatism and concern for the
preservation of fund assets. Meyers and Miller report, on the basis of
a survey of investment advisers, that it usually takes from 2 to 3 years
to educate fund trustees in the advantages of a flexible investment
policy, that is, one which would include liberal portions of corporate
bonds and common stock in the fund portfolio.'2 Many of the larger
multiemployer plans were established in the 1955-58 period,13 and it
would be the trustees of these funds who were making decisions to
change portfolios in the period of this survey. Since it usually takes
several years for the change in policy to be fully implemented, it can
be predicted that the shifts in overall multiemployer and union fund
portfolio composition which are evident in the 159-64 figures will
continue for several more years at least.
BANK-ADMINISTERED FUNDS
One of the most plausible reasons for the similarity between the
investment portfolios of multiemployer `and union funds and corporate
~-° F. Meyers and C. R. Miller, "Investment Policies of Bilaterally Managed Pension
Plans," Quarterly Review of Economics and Business, February 1962, p. 49.
`° Bureau of Labor Statistics, Bulletin 1326, table 1.
PAGENO="0144"
136 OLD AGE INCOME ASSURANCE-PART V
funds is that in both a strong influence on investment policy is exerted
by commercial bank trust departments.14
When a bank trust department is connected with the investment of
the fund, multiernployer pension funds show a greater proportion of
corporate bonds and common stock and a smaller proportion of Gov-
ernment bonds and mortgages than do the selected funds in which the
effect of large atypical funds is eliminated. That is, the bank-associated
funds are closer to corporate funds in portfolio distribution than are
the selected funds, which are taken as representing the average multi-
employer and union fund. This is shown in table 1-12. The portfolio
for bank funds is derived from a sample of 128 funds (having $420
million in assets in 1959) for which it could be determined that a bank
acted either as trustee, agent, or investment, adviser.
TABLE -12.-INVESTMENT PORTFOLIO OF SELECTED AND BANK-ADMINISTERED MULTIEMPLOYER AND UNION
FUNDS AND CORPORATE FUNDS, 1959
[In percent of book
value[
Multiemployer
and union funds
Corporate funds
All selected
Bank
administered
U.S. Government securities
28.1
16.3
8.8
Corporate and other bonds
Preferred stock
43.3
2.6
54. 1
3.3
51.3
2.6
Common stock
19. 4
23. 1
30.9
Mortgages
Otherinvestments
4.7
2.0
2.9
.3
2.3
4.0
Note-Figures are rounded and may not add to totals.
Source: Multiemployer and union funds, NBER survey; corporate funds, calculated from data in SEC Statistical Bulletin,.
June 1962.
VARIATIONS IN PORTFOLIO COMPOSITION BY UNION
Union policy is sometimes considered to be the most important
factor influencing the distribution of multiemployer and union pen-
sion. fund assets among the different investment media. While this is
obviously true for some unions, e.g., those which are associated with
atypical funds, it does not appear to be true for the bulk of unions
involved with multiernployer funds.
Table 1-13 shows the portfolio composition of 348 self-insured.
multiemployer and union pension funds classified by union. The cont-
pilation includes 18 national unions, or all those that had total assets
in excess of $15 million in funds covering their members in 1959. In
all, these funds mclude just over 60 percent. of all self-insured multi-
employer and union funds in the 1959 survey, and contain $1 billion,
or almost 80 percent of the assets in this type of fund.
14 Evidence of trust department influence in the investment of corporate funds Is given in
Harbrecht, Pension. Funds and Economic Power, New York, 1959, pp. 226-227. BLS data
on multiemployer funds show that the person determining investment policy was a corpo-
rate trustee or investment agent, acting either alone or In conjunction with the board of
trustees, In about one-fourth of the funds, covering about 15 percent of the workers (Bulle-
tin 1326, p. 92). These data probably understate the influence of trust departments in the
actual investment decision making since a corporate trustee is usually given broad discre-
tion In the trust agreement, and boards of trustees frequently hire banks as Investment
advisers and then abide by the bank's judgment in portfolio policy.
PAGENO="0145"
AGE INCOME ASSURANCE-PART V
137
OLD
~ ~ Ill
F-
83-200-67-pt. 5-l0
PAGENO="0146"
138 OLD AGE INCOME ASSURANCE-PART V
With the exception of the unions having atypical funds, there is a
relatively narrow range of portfolio proportions for rough categories
of assets. If we postulate a portfolio such as the following:
Cash - 10 percent or less.
Bonds 55 to 70 percent.
Stock ~0 to 30 percent.
Mortgages less than 10 percent.
the aggregate portfolios of 11 of the 18 unions (14 if the atypical ones
are excluded) with more than $15 million in pension assets fall ap-
proximately within the percentages cited. Two of the remaining seven
(or three typical) deviate from the "average" union portfolio only
slightly. In one case this reflects a preponderance of small funds with
heavy cash holdings; in the other, a liberal portfolio of common stock.
It is significant to note that the portfolio of bank-associated funds
(table 1-12) falls squarely in the middle of the "average" union's
portfolio.
FUNDS HOLDING STOCKS AND MORTGAGES
The substantial increase in the proportion of common stock and
mortgages in multiemployer and union funds raises the question: Does
this increase derive primarily from ftrnds adding to their existing
holdings of these securities at a faster rate than heretofore, or are
more funds moving into stock and mortgages for the first time?
Only four funds in the sample of 8715 bought common stock for the
first time during the period 1960-64, although it should be noted that
a large proportion (76 percent) held stock at the beginning of the
period (see table 1-14). Thus the increase in common stock in the ag-
gregate portfolio appears to have come primarily from increased
holdings by the funds already committed to this type of investment.
TABLE -14.-NUMBER OF MULTIEMPLOYER AND UNION FUNDS IN SAMPLE HOLDING, SELECTED TYPES
OF ASSETS, 1959-64
1959
1960
1961
1962
1963
1964
Total number of funds in sample
Total number of funds holding-
Mortgages
Common stock
87
27
66
87
32
68
87
33
68
87
41
69
87
49
70
87
50
70
Source: NBER survey.
By contrast, the number of funds in the sample having mortgage
investments nearly doubled between 1959 and 1964. In addition, sev-
eral large funds, for example, the Teamsters' Central States, South-
east, Southwest Areas Pension Fund, channeled the greater propor-
tion of their new moneys into mortgages. Therefore, the growing per-
centage of mortgages in rnultiemployer and union pension funds can
be traced to a large number of new entrants into the mortgage field
as well as to increased holdings by those funds already having sub-
stantial mortgage investments.
`° Although the sample used to estimate asset totals and portfolio distributions num-
bered 90 funds (see app. I). 3 of these funds did not provide detailed investment break-
downs for each of the years covered.
PAGENO="0147"
OLD AGE INCOME ASSURANCE-PART V 139
5. SUMMARY
Assets of multiemployer and union pension funds are small in com-
parison to corporate pension funds, but their rate of growth is sub-
stantially higher than that of corporate funds.
The high growth rate of rnultiemployer and union funds is a re-
flection of their younger average age.
Assets of multiemployer and union pension funds, like corporate
funds, are highly concentrated in a relatively few large funds.
Assets and coverage of multiemployer and union pension funds,
unlike corporate funds, are concentrated in nonmanufacturing in-
dustries. An exception is the large accumulation of assets in funds
covering employees in the apparel and other finished-textile products
industry.
The portfolio composition of multiemployer and union pension
funds shows significant differences when compared to corporate pen-
sion funds. However, these reflect, in part, differences in structural
characteristics and, in part, highly atypical responses to investment
choices by a few large multiemployer and union funds. The remain-
ing differences are fast diminishing because of shifts in investment
choices by the average multiemployer and union fund and because
of the slower growth rates of atypical funds. For the future, although
we can expect the two types of fund-multiemployer and cor-
porate-to become more alike in portfolio composition, it is likely that
dissimilarities will always exist because of the persisting structural
differences, that is, average size and liquidity needs, and because of
investment preferences.
Most unions do not take an active role in shaping the investment
policies of pension funds covering their members. For the most part,
this responsibility is delegated to professional investment managers,
such as commercial bank trust departments. Many of the funds that
do not delegate the function of portfolio management nevertheless
follow the pattern of investment diversification common to bank ad-
ministered pension funds.
In the funds covering members of the TCWH, IBEW, ILGW,
ACWA, and IThIW, the effect of union policy on portfolio composi-
tion is clearly discernible. In all of the other unions with substantial
pension fund assets, union policy per se appears to play little or no
role in shaping fund investment policy.
Union policy does not appear to be a factor affecting the type of
union participating in the administration of multiemployer funds.
The unions which control or jointly administer large aggregates of
pension fund assets demonstrate a wide variety of structures, leader-
ship, and approaches to unionism. The common characteristic of these
unions is that some members work in small establishments or are
included in small bargaining units attached to medium- or large-sized
companies, or that employment with a single firm in the trade or in-
dustry for a long period of time is improbable. These characteristics
are common to a wide range of unions. Since approach to unionism
does not appear to be a deciding factor influencing union involvement
in multiemployer and union pension funds, it should not be surprising
that union policy plays, in the aggregate, only a minor role in shaping
the investment of pension funds.
PAGENO="0148"
140 OLD AGE INCOME ASSURANCE-PART V
APPENDIX I: ESTIMATING PROCEDURES
Data on assets of multiemployer and union funds for 1959 are de-
rived from the survey described above. Of the 715 multiemployer and
eight union funds surveyed, 85 were insured and 577 self-insured;
information was not available on the remaining 61. The total asset
figure used in the study represents an estimate for all self-insured
funds, and the aggregate portfolio composition is that of all self-
insured funds where asset breakdowns were available (561 funds).
The proportion of insured plans in this survey is approximately the
same as that found by the Bureau of Labor Statistics.16
Data on assets and portfolio composition for subsequent years were
derived from a stratified random sample of the funds included in the
1959 survey. Information on all of the funds that had more than $5
million in total assets at the end of 1959 was obtained. These 36 funds
represented over 60 percent of the assets in all multiemployer and
union funds in the 1959 survey. In addition a sample of 64 funds in the
size classes below $5 million was drawn. (See table 1-4 for data on the
concentration of assets.) The sample size and the number of sampling
units in each size stratum were based on a system of optimum alloca-
tion which takes into account not only the number of funds in each
stratum of the PoPulation (Sh) but also the standard deviation of each
stratum (S71). The allocation of sampling units to each stratum was
proportional to the product of the number of funds in each stratum
and the standard deviation of the stratum (N7S5) .~ A sample size of
7° Bureau of Labor Statistics. Bulletin. 1326. p. 17.
When optimum allocation is used, the sample size (a) for estimating a total, e.g., total assets for all
funds, is given by the formula:
V =
l±~-7~V~S~
where
(~N0.S7~)
or the sample size uiscorrected for a finite population, and V is the desired variance in the sample estimate
and is determined l)y the formula V=(dXT)2, where d is the acceptable margin of error or the half-width
of the confidence interval (in this case .025) and T is the universe total (in this case the total assets of all
funds with less than 95 million in 1959).
;Thus,
17= (.025X$47S,240)~=142,945,936.
(114,872)2
~ 142,945,936~' ~
92.3
62.897.542 =64.1
142.945,936
or a sample size of 64.
No So
Allocated according to the ratio in each stratum of ~ the sample size in each stratum was:
Total 4ssrts Sample Size
Strut torn (thousaszds of dollars)
1 0-99 1
2 100-249 3
3 250-499 4
4 500-999 6
5 1.000-1.999 14
6 2. 000-4, 999 36
64
SOurce: William G. Cochran. Sampling Tcctmqouo, New York, 1953, pp. 73-74, 87-90,
PAGENO="0149"
OLD AGE INCOME ASSURANCE-PART V 141
64 for the funds with less than $5 million in total assets will give an
estimate of the aggregate assets in these funds, with a maximum error
of 2.5 percent. Since each of the funds with assets over $5 million was
surveyed, the resulting estimate for all multiernployer and union
funds for years subsequent to 1959 would be in error by substantially
less than plus or minus 2.5 percent.
The portfolio composition of multiemployer and union funds was
determined by weighting the sample portfolio distribution for each
stratum by the amount of assets in that stratum in 1959. The results
of this calculation are show in table 1-7.
Since the sample used to estimate aggregate assets and portfolio
distribution was drawn from the universe existing in 1959, funds
formed since that year are not included. No attempt was made to derive
an independent estimate of the assets added by newly formed funds
in the years 1960-64. This means that the aggregate assets figures
may be too low, and the size of the error may tend to increase as the
sample year moves farther from 1959.
However, there are two factors that may keep the error within
reasonable bounds. First, the growth of funds in the sample is some-
times augmented by the merger of an existing fund not in the sample.
This lends an offsetting upward bias, but probably not a very large
one, to the estimate of aggregate assets. Second, new funds are likely
to be relatively small in terms of assets for several years after their
inception, even if they cover sizable numbers of persons, because past
service liabilities are funded slowly.
On balance, it appears that the sample estimates are better for the
earlier years. However, the figures for later years are not likely to be
greatly in error.
PAGENO="0150"
PART II. PENSION FUNDS OF NONPROFIT
ORGANIZATIONS
BY ELIZABETH T. Smr~sox
6. INTRODnCTION
Nonprofit organizations comprise a diverse group, but, in the pas't~
statistics on their noninsured pension funds have been based mainly
on funds for Protestant ministers. This paper attempts to compile
more exact data for the group as a whole on total assets, their invest-
ment pattern, and the rate at which they are increasing.
The term "nonprofit organization" is used here to refer to nongov-
ernmental oi~ganizations exempt from taxation and described in section
501 (c) (3) of the Internal Revenue Code. Included are membership
corporations, community chests, funds, and foundations organized
and operated exclusively for religious, charitable, scientific, literary,
educational, or humane purposes. No part of their net earnings may
benefit any stockholder or individual, and no substantial part of their
activities may be directed toward influencing legislation. Other non-
governmental exempt organizations are also discussed briefly below
in the section "Other Nonlrofit Or~anizat.ions."
An investigation of pension pla~s for employees of nonprofit organi-
zations showed reliable statistics on the size and portfolio composition
of the noninsured funds reporting annually to the Church Pensions
Conference and on a few similar funds for colleges, groups of inde-
pendent schools, small hospitals, and other organizations. Surveys
have been made by various associations of nonprofit organizations to
ascertain the number of units with pension plans and the popularity
of different plan features. In general, however, the surveys contribute
only indirectlyto knowledge of fund assets, growth rates, and port-
folio composition.
This paper presents (by methods discussed in app. II) fairly com-
prehensive data on noninsured funds and the first estimates of reserves
of insured plans for nonprofit organizations. The latter cover the years
1958-64. Pay-as-you-go pension plans are practically unfunded and
have, therefore, been excluded.
The Standard Industrial Classification includes most nonprofit or-
ganizations under these major group headings: medical and other
health services; educational services; museums, art galleries, and
botanical and zoological gardens; nonprofit membership organiza-
tions; and miscellaneous services.1 Historically, nonprofit retirement
programs can be divided into four categories: religious bodies, edu-
cational institutions, hospitals. and other nonprofit organizations.
The Standard Industrial Classification includes religious bodies and
1 It should be noted that several of these headings also include profit-seeking organiza-
tions, but only nonprofit organizations will be included in this paper.
142
PAGENO="0151"
OLD AGE INCOME ASSURANCE-PART V 143
other nonprofit organizations in their category: nonprofit membership
organizations. Educational institutions, as used here, include the non-
profit part of SIC's educational services, museums, art galleries, etc.,
and miscellaneous services. The SIC classifies each organization under
its function rather than its affiliation; for example, all schools are
grouped together, including parochial schools, and all hospitals to-
gether, including those that belong to a university.
This study is limited to private nonprofit organizations in the United
States. In 1964, there were over 400,000 such organizations, counting
each religious congregation, each nonpublic nonproprietary school and
college, each voluntary ho~pital, and each other nonprofit organiza-
tion. They had approximately 2.7 million employees. That figure ex-
cludes members of religious orders and ordained or unordained minis-
ters who accept little or, no remuneration for their services. The for-
mer are cared for by their orders whether they are active or no longer
able to perform their duties. In a sense, they can be considered covered
by a pay-as-you-go retirement plan, but that is outside the scope of
this paper. Ministers who preach on Sundays, but depend on other jobs
for their livelihood, probably look toward the latter for pension cov-
erage. Probably between 10 and 20 percent of the 2.7 million employees
work only part time and, therefore, may not be eligible for pension
plan membership. The annual payroll for the full- and part-time em-
ployees was approximately $10 billion.
7. Siz~ OF FUNDS: VARIATIONS BY TYPE OF PLAN AND TYPE OF NON-
PROFIT ORGANIZATION
Pension funds of nonprofit organizations amounted to $3.4 billion 2
at the end of 1964, or 4 percent of total private pension funds. Almost
one-half of the nonprofit funds ($1.6 billion) were included in the'
figures on insured pension plans compiled by the Institute of Life In-
surance. In contrast, as table TI-i shows, insured funds accounted
for less than one-third of all other private pension funds.
TABLE 11-1.-PERCENTAGE DISTRIBUTION OF PENSION FUNDS BY TYPE OF PLAN, NONPROFIT ORGAN IZATION
AND ALL OTHER PRIVATE GROUPS, 1958-64
Nonprofit o
End of year
Total
rganizatinn fu
with-
nds insured
Non-
(insured
percent)
All
other private funds
Total
Insured
(percent)
Non-
insured
(percent)
TIAA-
CREF
(percent)
Agency
companies
(percent)
1958 $1. 7 33. 7 21. 7 44. 6 $39. 2 37. 5 62. 5
1959 1.9 34.1 21.7 44.2 44.8 37.0 63.0
1960 2. 2 34. 0 22. 5 43. 5 49. 9 35. 5 64. 5
1961 2.4 34.8 21.7 43.5 55.3 34.3 65.7
1962 2. 7 36. 0 20. 9 43. 1 61. 0 33. 3 66. 7
1963 3. 0 37. 5 19. 9 42. 6 67. 1 32. 5 67. 5
1964 3. 4 38. 4 19. 7 41.9 74. 1 31. 9 68. 1
Source: For derivation of funds of nonprofit organizations, see app. II. All other private funds: insured, from `Life In-
surance Fact Book," pension plans in the United States insured with life insurance companies, reserves end of year less
TIAA and agency company plans for nonprofit organizations; noninsured from Securities and Exchange Commission,.
Statistical Bulletin, assets of private noninsured pension funds, book value, less corresponding funds for nonprofit organi-
zations.
2 In terms of book value, i.e., book value for noninsured funds, `admitted assets (gen-
erally close to book value) for insured funds, including TIAA, and an estimate of book
value for CREF. The latter was computed `here by cumulating income and disbursements but
omitting changes in market value of common stock.
PAGENO="0152"
144 OLD AGE INCOME ASSURANCE-PART V
The largest part of the insured funds of nonprofit organizations
($1 billion in 1964) was carried with Teachers Insurance & Annuity
Association. TIAA is a legal reserve life insurance company, but it
*does not employ agents, and its services are restricted to nonprofit,
tax-exempt educational and scientific institutions. Other insurance
~cornpanie*s are called agency companies here. They include a few com-
panies, generally affiliated with religious bodies, similar to TIAA but
unimportant in the aggregate; also the National Health & Welfare
Retirement Association, which maintains a retirement system for
employees of nonprofit organizations engaged in charitable, health,
and welfare work. The latter's reserves of $0.2 billion in 1964 were
reinsured 100 percent with an agency company.
College Retirement Ec~uities Fund provides variable annuities for
participants in TIAA's fixed-dollar pension plans. While CREF is not
an insurance company and its funds are not included in the ILl tabu-
lations, they are included here with those of TIAA because of their
affiliation. In 1964 CREF accounted for $0.3 billion.3
The TIAA-CREF percentages are an overstatement since they
include reserves for some publicly supported universities and for
educational institutions in Canada and other foreign countries. The
agency-insured and the noninsured percentages of nonprofit organiza-
tions, on the other hand, are understated because of incomplete cov-
erage. The total figures for nonprofit organizations, however, may be
fairly accurate.4
A further breakdown of funds by purpose of nonprofit organization
is possible for 1960 and is shown in table 11-2. The largest amounts
are for educational institutions with TIAA-CREF plans and for
religious bodies with noninsured plans. Total funds for hospital pen-
sion plans have increased substantially since 1960, but they are still
probably smaller than for any of the other groups.
TABLE 11-2.-PERCENTAGE DISTRIBUTION OF PENSION FUNDS OF NONPROFIT ORGANIZATIONS BY PURPOSE OF
ORGANIZATION AND TYPE OF PLAN, END OF 1960
Religious
bodies
Educational
institutions
Hospitals
Other
nonprofit
organizations
Total
insured with-
0
1. 0
33. 5
34.5
33.8
7. 6
2. 2
43.6
0.1
3. 9
. 1
4.1
0.1
9. 9
7. 8
17.8 -
34.0
22. 5
43. 5
100.0
TIAA-CREF
Agency companies
Noninsured
Total
Note: Totals do not always add because of rounding.
Source: See app. II.
Growth rates of the dollar amounts depend on rates of increase in
pension coverage, contributions, benefit payments, and fund earnings.
Table 11-3 shows that in 1960 about a third of all units of nonprofit
organizations had pension plans and a fifth of the employees were
`In terms of market value, total pension funds of nonprofit organizations aggregated
S~.7 billion. This figure is composed of S1.fi hillIon n'~ninsuieri (comnared to hook value of
51.4 billion). 50.7 biflion agency-insisted. Si billion TIAA. and 50.4 billion CREF. Admitted
assets were used for the insured funds. hut if market values were available. in all prob-
~ibility the total would not be affected.
For sources of data asid estimating techniques. see app. II.
PAGENO="0153"
OLD AGE INCOME ASSURANCE-PART V 145
covered. It should be noted that pay-as-you-go coverage is omitted
since it would add virtually nothing to total pension funds. Within the
four groups, coverage rates varie.d from 10 to 35 percent; within sub-
groups, the variation was even greater.
TABLE 11-3.-PENSION COVERAGE RATES OF NONPROFIT ORGANIZATIONS BY PURPOSE OF ORGANIZATION, END
OF 1960
Religious
bodies
Educational
institutions
Hospitals
Other nonprofit
organizations
Total
Units:
Number (thousands)
Percent with pension plans
Employees:
Number(thousands)
Percent with pension plans
310
40
400
30
26
15
450
35
4
15
850
10
75
10
500
20
415
33
2,230
20
Note: Excludes pay-as-you-go coverage.
SOURCES
Number of units; Religious bodies, from National Council of Churches, "Yearbook of American Churches for
1961"; Educational Institutions, from Bureau of the Census and Bureau of Old-Age and Survivors Insurance,
"County Business Patterns,' lot quarter 1959 and 1962, part I, educational services, museums, art galleries,
etc., and nonprofit research organizations less correspondence and vocational schools; Hospitals, from American
Hospital Association, Hospitals, Guide Issue (Aug. 1, 1961) for voluntary hospitals. Other nonprofit organizations
from "County Business Patterns" (1959 and 1962), nonprofit membership less religious organizations.
Number of employees: Religious bodies, lay workers, from "County Business Patterns"; ministers, priests, and
rabbis: see app. II. Educational Institutions, Hospitals, and Other Nonprofit Organizations, same as for number
of units. Members of religious orders are excluded.
Coverage: From annual reports of TIAA-CREF and various noninsured funds and answers to questionnaires sent
to life insurance companies and pension boards. See app. II.
Protestant Ministers
The Church Pensions Conference, a group of pension-plan officials
that has been meeting annually for 50 years, accounts for most of the
noninsured funds of religious bodies. A group of 19 Protestant de-
nominations reporting to the conference for the 23-year period 1942-65
showed a rise in coverage of ministers from 50 to 70 percent.5 In 1965
the coverage rates of the different denominations included ranged
from 60 to 95 percent. While there were many other Protestant bodies,
they were represented by comparatively few ministers who devoted
most of their time to and derived most of their income from the min-
istry and, therefore, were potentially eligible for coverage by a church
plan. Most of the larger Protestant denominations not affiliated with
the Church Pensions Conference, even some with little or no central
organization, had plans either in operation or under study by 1965. In
general, the plans were relatively new and had low but gradually in-
creasing coverage. Less than 55 percent of all Protestant clergymen
(estimated at about 175,000 in 1960) making a career of the ministry
were covered by a pension plan. The well-established CPC plans had
an average total contribution rate of 12 percent of salary. Some of the
other plans depended mainly on special annual offerings of church
members and had low dues, but in general the tendency was to try to
pattern them after the larger CPC plans. There is, therefore, a good
chance of expansion in noninsured funds of Protestant bodies; over
the years 1960-64, there was an annual increase of almost 11 percent.
Estimates based on data given by Kenneth H. Ross In "Pension Plans of the Various
Religious Denominations." Transactdons of the Seventeenth International Congress of
Actuaries, vol. II, pp. 727-740, and on the 1965 statistical report of the Church Pensions
Conference. It sisould be noted that the 19 denominations had been reduced by merger to
14 in 1965.
PAGENO="0154"
146 OLD AGE INCOME ASSURANCE-PART V
Catholic Priests
Retirement plans for the 33,000 diocesan Catholic priests in the
United States vary among the 26 archdioceses and 112 dioceses. In 1960
more than half of the jurisdictions had an organization with a title
such as Infirm Priests' Fund or Priests' Mutual Benefit Society. These
provide medical care and, in some dioceses, a pension plan. Correspond-
ence with a sample of the larger archdioceses reporting such funds or
societies suggests that pension plans were available to about 10 percent
of all diocesan priests. In other dioceses a priest is promised a salary
or sustenance payment for life. There is ordinarily no fixed retirement
rage; the priest retires at his request and with permission of his bishop,
but usually only when he feels he is no longer able to perform his duties.
Some of the plans were funded, but not fully, and a few were insured.
Whether there is a partially funded plan or just an agreement to sup-
port the priests for life, the payments required from the general funds
of the diocese or from its parishes are proving very expensive these
days, and some dioceses are looking into funding arrangements. The
plans on which we have information generally provide a fiat-sum
benefit of $50 or $100 a month. Since no provision is made for depend-
ents and few priests ask to be retired, the funds would be expected to
be proportionately less than those for Protestant and industrial plans.
Rabbis
Over 3,000 ordained rabbis were active in religious work in the
ITnited States in 1960, a.nd almost half of them belonged to three major
rabbinical associations which reported insured pension plans. Cover-
age was estimated at almost 60 percent for this group. Information
was not available for the remaining rabbis, some of whom belong to
two other major associations or to minor groups, but their pension
coverage is estimated to be low. The high proportion of rabbis who had
elected social security coverage (92 percent in 1958) suggests a future
increase in their pension funds.
Clergy of Other Religious Bodies
Other religious groups in the United States probably had around
2,000 ministers. Pension information is not available, but the largest
group-Eastern Orthodox, with about 80 percent of total ministers-
had 71 percent social security coverage in 1957, and may, therefore,
have some pension coverage.
Lay Employees of All Religious Bodies
There were approximately 190,000 lay employees of religious bodies
in 1960. Almost 60 percent were employed by Protestant churches and
almost 40 percent by Catholic churches. Probably a large proportion
only worked part time, and some were retired from jobs in industry or
Government. Beginning in the 1930's, members of the Church Pensions
Conference started to set up plans for lay employees. In a few denomi-
nations lay workers may join ministerial pension plans, but in the
majority of the denominations the same board administers separate
though similar plans for the two groups. In general, the lay Protestant
plans are included in table 11-2 among noninsured funds for religious
bodies. In 1960 coverage was probably less than 3 percent for lay em-
PAGENO="0155"
OLD AGE INCOME ASSURANCE-PART V 147
ployees of local churches, but well over 50 percent for denominational
board and National Council employees. Also included in the funds of
religious bodies are small amounts for teachers and other employees
of church-affiliated educational institutions which could not be sub-
tracted. In general, however, church-affiliated educational organiza-
tions have joined the same plans as private schools and colleges.
Pension coverage is also very low at present for Catholic lay em-
ployees, but it received a stimulus in July 1962 with the inauguration
of a plan that will eventually cover all lay employees of the Archdio-
cese of New York (including some in education, hospitals, and other
nonprofit organizations).
EDUCATIONAL INSTITUTIONS
Colleges and Universities
In 1960 there were approximately 1,500 nonprofit, nonproprietary
colleges and universities with 240,000 teachers, administrative person-
nel, and other employees. Ninety-five percent of the teachers were
eligible for pension plans, or would be if they fulfilled certain require-
ments. Similar percentages for other types of employees were lower,
but were probably over 80 percent for all employees including teach-
ers. In contrast to this potential percentage of coverage was the actual
rate of less than 50 percent. This was mainly the result of waiting pe-
riods and voluntary participation in numerous plans, and also of high
turnover among nonacademic employees, some of whom were past
the age of plen entry. The average waiting period was from 1 to 3 years,
or until attainment of assistant professorship or equal rank, or age 30.
Participation was voluntary in plans affecting approximately 28
percent of faculty members and 44 percent of clerical and secretarial
employees. About three-quarters of the coverage was with TIAA-
CREF. Recent T1AA-CREF annual reports have mentioned that
waiting periods are being reduced and eligibility extended to new
categories of employees. Also, as an added fringe benefit, contributory
plans are being changed to noncontributory, and the change naturally
tends to increase the coverage rate.
From 1960 to 1965 there was a large increase in private colleges
participating in the TIAA-CREF retirement system. A few were new
colleges or previously had had no pension plans, but most were just
giving their employees a choice between TIAA-CREF and their other
plan, or allowing them to add a variable annuity. Although about
20 percent of the private institutions of higher education (including
junior colleges) had no pension plans, they employed only about 4
percent of all college faculty members, excluding members of religious
orders. Therefore, the establishment of plans in colleges which do not
have them at present will not have a large effect on coverage rates
and fund assets.
Increasing contribution rates should increase fund assets, and this
has been happening recently. Sixty percent of college teachers in 1959
and 75 percent in 1965 were members of plans with total contribution
rates of over 10 percent (and up to 20 percent) of annual salary.
PAGENO="0156"
148 OLD AGE INCOME ASSURANCE-PART V
Elementary a~ Secondary Schools
There were approximately 19,000 nonpublic, nonproprietary elemen-
tary and secondary schools in the United States in 1960, with 190,000
teachers and other employees. Three t.housand were independent or
private schools and 16,000 were parochial schools. Pension plans were
quite usual for teachers in private schools, including many more or less
closely affiliated with religious bodies.
Almost three-quarters of the parochial schools were run by Cath-
olic churches, but they employed just over half of all lay teachers.
Protestant parochial schools acco~mted for a quarter of the schools
and over one-third of the teachers. Approximately 2 percent of the
schools and 10 percent of the teachers served Jewish education. George
N. Shuster has pointed out that the present-day plight of Catholic
parochial schools stems from the great surge in enrollment in the
past two decades without a corresponding increase in members of
teaching orders.6 The situation has caused a large increase in the
number of lay teachers and in the cost of school operation. Since it
has not been possible to make most teachers' salaries comparable with
those paid by the public and private schools, it is not surprising that
pension plans are slow in developing. However, correspondence around
the end of 1960 with dioceses and archdioceses with the largest num-
bers of lay teachers revealed that they were considering the problem.
In 1961 the Diocese of Pittsburgh put into effect a TIAA plan for its
high schools, and the following year the Archdiocese of New York
announced a plan for all lay employees including those in its schools.
By October 1960, pension plans were available for teachers and ad-
ministrators in Jewish schools of all ideologies and in all communities
in the United States and Canada. Among Protestant parochial schools
there were various plans, including some of the Church Pensions
Conference, for the National Union of Christian Schools and for in-
dividual schools. In general, coverage was low for all parochial schools;
in fact, only about 20 percent of all private and parochial school em-
ployees were eligible.
It would appear that almost half of the coverage was in agency-
insured plans, and a quarter each in TIAA and noninsured plans. In
general, the TIAA plans were for the long-established private schools,
and the noninsured plans for Protestant parochial schools in church
plans. In 1960-65 the number of independent schools in TIAA in-
creased from 272 to 368. Coverage of independent schools may he ex-
pected to increase as college and university coverage has in recent
years. Protestant and Jewish parochial school coverage may also be
expected to rise fairly rapidly, but it is very difficult at present to
foresee how much Catholic parochial pension funds will increase.
Pension increases won by unionized public schoolteachers may have
an effect on independent and parochial schools.
Other Educational Services
There were about 90,000 employees in other educational services,
of which two-thirds were in nonprofit research organizations. The
remainder were in libraries, museums, art galleries, botanical and
zoological gardens, and schools a.nd educational services not elsewhere
"Schools at the Crossroads," Atlanflc Monthly, August 1962.
PAGENO="0157"
OLD AGE INCOME ASSURANCE-PART V 149
classified. Correspondence and vocational schools have been omitted
because they were generally proprietary. About 25 percent of the
employees in other educational services were actually covered by a
pension plan. The majority had agency-insured plans, and a goodly
number had TIAA-CREF annuities.
Greenough found that a very high proportion of the large research
organizations and foundations had pension plans.7 The waiting period
was similar to that for colleges and universities and has probably
been shortened somewhat since 1960. It was hoped that a new plan
for museum personnel would raise the number covered in that field.
HOSPITALS
The greatest possibility of pension asset growth appears to be in
the hospital field. The smallest amount of funds of any of the four
types of nonprofit organizations (table 11-2) is matched with the
largest number of employees (table 11-3). Most plans are compara-
tively new, and both coverage rates and assets per covered employee
are low.
Approximately half the hospital employees are nonprofessional,
other than clerical workers. They include nursing aids, practical
nurses, maintenance workers, housekeeping and food service em-
ployees, and laundry workers. The next most populous class-the
registered professional nurses, including those in supervisory and
teaching positions-accounted for about a fifth. Both registered nurses
and the nonprofessional group have high turnover rates and a large
proportion of part-time workers. From the Bureau of Labor Statistics
survey of hospital employees in the Nation's metropolitan areas in
mid-1960, it is estimated that pension plans other than social security
were available (or would be once certain requirements had been met)
to about 37-38 percent of full-time registered nurses and nonprofes-
sional employees in voluntary hospitals (see app. II). The rate
was about 45-46 percent for full-time professional and technical
employees other than nurses and full-time clerical workers. The latter
two groups had lower turnover and fewer part-time workers than the
former two groups. Information was not collected on executive and
administrative personnel and part-time workers. Most of the plans
are applicable to all categories of employees, but a great many are
voluntary and contributory.
Although the American Hospital Association's national retire-
ment program has been in existence since 1947, the highest coverage
rates seem to be in areas where a regional plan operates, such as those
of the Cleveland Hospital Council, Texas Hospital Association, Kaiser
Foundation Hospitals in the Far West, and Jewish federations in the
larger cities. Over 95 percent of the funds are insured with agency
companies, including about 18 percent with National Health and Wel-
fare Retirement Association.
Pension eligibility has been growing; the BLS study in mid-1963
showed 47 percent of registered nurses and nonprofessional workers
and 50 pe.rcent of professional and technical and clerical employees
`William C. Greenough, "Compensation for Foundation Staff: Salary, and Benefits," in
Henry Scum (ed.) Fifth Biennial Conference on Charitable Foundations, New York, 1961.
PAGENO="0158"
150 OLD AGE INCOME ASSURANCE-PART V
in hospitals with pension plans (see app. II). And, the rate is
likely to grow steadily in the near future. "Hospitals are * * * finding
it necessary to establish pension plans as a fringe benefit, so that they
can compete with industry and other professional fields where pen-
sion and retirement benefits have long been an accepted part of the
job." "The hospital personnel of the future will have to come from
closer to the top than the bottom of the barrel," says Ray E. Brown,
vice president for administration, University of Chicago.9 Union
activity among nurses and nonprofessional hospital workers has
created additiona.l pressure for increased pension coverage. Further-
more, prepayment plans are helping general hospitals raise salaries
and fringe benefits.
OTHER NONPROFIT ORGANIZATIONS
This category includes all of the Standard Industrial Classification
Code's group 86, nonprofit membership organization, except group
8661, religious organizations, which corresponds to our category
"religious bodies." These organizations together have approximately
500,000 employees, broken down as follows:
Charitable organizations 100, 000
Professional membership organizations 10, 000
Civic, social, and fraternal associations 205, 000
Labor organizations 100, 000
Business associations 50, 000
Political organizations 5, 000
Nonprofit membership organizations, n.e.c. 30, 000
The first two groups and subgroups of some of the others fall under
section 501(c) (3) of the Internal Revenue Code. The other groups
and subgroups are not organized for profit and are also exempt from
Federal income tax, but they promote the interests of their members
and many engage in lobbying. They are included here only because
they are probably not included elsewhere and our data are not detailed
enough to exclude them.
The majority of the charitable organizations are affiliated with re-
ligious bodies. Most Jewish agencies are included in the federation~
plans which were mentioned in the section on hospitals. On the national
level, coverage of Catholic charitable organizations is generally high;
on the diocesan level, as is the case for other diocesan lay employees,.
pension plans are being studied but few have begun operating. Cov-
erage is rather low, on the average, in Protestant charitable organiza-
tions. The American Red Cross dominates the groups not affiliated'
with religious bodies, and it has its own trusteed plan. In Cleveland
there is a plan of the Welfare Federation to which religious and secu-
lar organizations belong. Approximately half of the charitable plans
are insured, mostly with the National Health and Welfare Retirement
Association, and the other half are self-administered.
Pension plans are quite common in the other groups listed under
"other nonprofit organizations" except when the organization has very
few paid employees. Most of the plans are insured, but some of the*
8 George R. Wren. "Why Waste Scarce Manpower With a Fixed Retirement Age?"
Hospitals, Jan. 16, 1964.
"The Impact of Wages and Hospital Costs," Hospitals, July 1, 1963.
PAGENO="0159"
OLD AGE INCOME ASSURANCE-PART V 151
larger ones are not. There are also some rather informal unfunded
arrangements, which are not counted.
A study of full-time social welfare workers, including supervisors
and executives, in 1960 10 shows that two-thirds of the 40,000 in these
classifications were enrolled in a pension plan other than social secu-
rity. While most of these worked for charitable organizations, some
are employed by the Y's, Scouts, Camp Fire agencies, and so forth,
which are included by the Standard Industrial Classification Code and
our population figures among "civic, social, and fraternal associa-
tions." Those employed by hospitals and schools have been omitted.
The high average rate for full-time social workers and supervisors and
executives in the social welfare field compared with the low rate for
all workers in the other nonprofit organization category suggests that
coverage is generally very low for clerical and maintenance persom~el.
8. GROWTH RATES
High growth rates are usually associated with new pension funds,
but if there are no extraneous influences, such as war, inflation, or de-
pression, the rates gradually decline. A low growth rate would signify
a mature fund with all employees covered, little increase in the work
force or in compensation, and all pension liabilities fully funded.
The same pattern might also be expected for a large group of funds,
such as all corporate noniusured pension funds. Pension systems for
profit as well as nonprofit organizations and for governments began to
be established in the United States toward the end of the 19th century,
and an article in the Atlantic Monthly of April 1916 stated that
"practically every large and well-established industry is providing for
the disability and the old age of its employees." 11 It is estimated that
corporate funds amounted to about $50 million in 1920 and grew at an
average annual rate of almost 30 percent through the beginning of
1930. After that, depression and war brought about fluctuation of the
growth rate. Through the thirties the rate was down to 8 percent, but
with the war it rose sharply and then, as shown in table 11-4, began
to fall again.'2 The Inland Steel decision in 1949 that pensions are
subject to mandatory collective bargaining pushed the rate up again,
but since 1951 it appears to have been following the pattern.
Noninsured pension funds of nonprofit organizations may have ac-
counted for 25 to 30 percent of total private noninsured funds in 1920,
but the percentage has decreased rapidly to 9 percent. in 1945 and 3
percent in 1964. The estimates for the period before 1945 are rough,
but they suggest that growth rates were lowe.r for nonprofit than for
corporate noninsured funds. Growth rates of nonprofit funds were
higher in the 1920's than in any subsequent period, but they were not
as high as those of corporate funds. The nonprofit plans were vohm-
*tary, salaries were low, and most of the pension boards found that pro-
motion of the pension idea to employers and young employees was a
time-consuming process. They had still more difficulty in the 1930's
10 ~5alaries and Working Conditions of ~5ocial Welfare Manpower in 1960, a survey con-
ducted by the Bureau of Labor Statistics, the National Social Welfare Assembly, Inc., and
the Department of Health, Education, and Welfare.
11 Joseph H. Odell, "The Economic Crime of the Protestant Church," pp. 442-4M.
11 Data for 1920-44 from Raymond W. Goldsmith, A. Study of Saving in the United
States, Princeton, N.J., 1955, vol. I, table 1-15, p. 468.
PAGENO="0160"
152 OLD AGE INCOME ASSURANCE-PART V
TABLE 11-4.-ANNUAL GROWTH RATES OF NONINSURED PENSION
FUNDS OF NONPROFIT ORGANIZATIONS AND CORPORATIONS,
1946-64
Fiscal year
Nonprofit
organizations
Corporations
1946
8.7
22.2
1947
7.0
20.9
1948
9.9
17.8
1949
10.1
16.0
1950
8. 8
19. 2
1951
10.9
24.0
1952
9.0
21.8
1953
8.0
21.9
1954
9. 9
19. 0
1955
10.5
17.3
1956
10.1
17.1
1957
9.6
16.3
1958
9.7
14.6
1959
10.5
14.
8
1960
11.6
13.
8
1961
10.6
13.
0
1962
10.1
11.
5
1963
10.0
11.
0
1964
11.1
11.
5
Source: Nonprofit organizations, see app. II. Corporations, 1946-50, computed from Raymond W. Goldsmith, Robert
E. Lipsey, and Morris Mendelson, "Studies in the National Balance Sheet of the United States," Princeton University
Press for NBER, 1963, table Ill-Si-i; 1951-64, computed from SEC, "Corporate Pension Funds, Supplemental
Tables," table 15, and "Statistical Bulletin," June 1966.
when contributions to nonprofit organizations, and accordingly sala-
ries and fringe benefits, were Very low. Since nonprofit organizations
are exempt from income tax and the majority of their employees are
not unionized, their pension funds increased much less than those of
corporations in the early 1940's aitcl just. following the 1949 decision.
The silver lining to the nonprofit organizations' cloud is the slight
upward trend in their growth rates from 1946 through 1964, as shown
in table 11-4, compared to a downward trend for corporate funds. It.
should be noted that the two growth rates were quite similar in 1964.
Act.ually~ the nonprofit pattern is wavelike; an increase in growth
rates is followed by a decrease and then by another increase. The. in-
creases have been caused by the addition of new funds, increased cov-
erage, and improvements in contribution rates and investntent earn-
ings. Inflation has been an intportant factor in raising contribution
rates in the hope that future benefits will match cost-of-living rises. As
shown in the preceding chapter, there is still room for a. large pension
expansion in the nonprofit field. There is a question as to how much
of the expansion will be in noninsured funds, but it appears ti~at the
wavelike pattern of table 11-4 will continue for nonprofit noninsured
funds for some time.
The first insured pension policies were issued by TIAA in 1919 and
by agency life insurance companies in 1921. By 1930, reserves for all
insured pension plans amounted to $100 million. These funds seem to
have followed the model; their growth rate dropped almost. continu-
ously from over 50 percent in the last half of the 1920's to 7 percent
in the first half of the 1960's.'°
~5 The growth rate for 1925-29 is based on figures for the 5 insurance companies that
did the bulk of pension underwriting in that period, according to Murray Webb Latimer in
Industrial Pension ~~ystems in the United states and Canada, New York-, 1932, vol. II.
The 1930 figure for all insured pension plans Is from the Institute of Ll.fe Insurance's
~Pr~catc anti Public Pens6on Plans in the United To~tates~ New York, 1967.
PAGENO="0161"
OLD AGE INCOME ASSURANCE-PART V 153
Growth rates are shown separately in table 11-5 for TIAA and all
other insured pension plans (designated as agency insurance compa-
nies). The TIAA growth rate was above 30 percent in the late 1920's,
but it dropped almost continuously through 1945. The table shows that
it was lower in 1946 than the corresponding rate for agency insurance
companies. TIAA rates began to rise again in the late 1950's, while
those of agency companies continued to fall through the early 1960's.
It is uncertain whether or not the slight rise in 1963-64 means a change
in trend for the latter. The relative position of the two growth rates
had changed between the beginning and end of the table; TIAA had a
lower rate in 1946 and a higher rate in 1964 than agency companies.
TABLE 11-5.-ANNUAL GROWTH RATES OF TIAA, CREF, AND AGENCY LIFE INSURANCE
COMPANY PENSION RESERVES, 1946-64
Calendar
year
TIAA
CREF
TIAA and
CREF
Agency
insurance corn
panies for
Total
Nonprofit
organiza-
tions
All others
1946
10.6
10.6
18.1
1947
11.3
11.3
17.0
1948
11.2
11.2
19.0
1949
12.1
12.1
16.5
1950
11.4
11.4
16.3
1951
11. 1
11. 1
17.7
1952
10.3
10.6
17.4
1953
8.3
9.6
14.9
1954
8. 1
122. 5
9. 8
13. 5
1955
8. 0
67. 8
9. 8
13. 5
1956
7.7
51.8
9.7
10.5
1957
8.5
44.3
10.7
13.0
1958
9.2
43.4
12.0
10.7
1959
9. 3
40. 2
12. 6
12. 8
11. 2
12. 8
1960
9. 6
35. 9
13. 1
7. 2
17. 6
6. 9
1961
10. 1
32. 2
13. 6
7. 3
7. 0
7. 3
1962
10. 9
30. 9
14. 6
6. 6
6. 8
6. 6
1963
12.6
28.9
16.0
7.6
5.9
7.6
1964
12.3
27.0
15.7
8.2
11.8
8.1
Source: TIAA and CREF frorn TIAA-CREF annual reports. See app. II. Agency insurance com-
panies from "Life Insurance Fact Books," total reserves at end of year from table `Pension
Plans in the United States insured with life insurance companies," less reserves mentioned above
for TIAA. Plans for nonprofit organizations, see app. II.
The TIAA rise was caused by large increases in the number of par-
ticipa~ing institutions, mainly publicly supported colleges and uni-
versities, but also private nonproprietary institutions; and by reduc-
tion of waiting periods, extensions of coverage to other categories of
employees, and increasing contribution rates. The increases would
have been even larger if it had not been for the diverting of part of
TIAA annual premiums to CREF beginning in mid 1952.
Since CREF offers variable annuities, its statements are in terms
of market value and its balance sheets show no fixed reserves. The
second column of table 11-5 was obtained from cumulated funds,
omitting increases in market value of common stock. For the available
years CREF growth rates follow the model perfectly. When the data
are combined for TIAA and CREF, `the fall is reduced and the follow-
ing rise magnified. TIAA-QREF `and noninsured funds for nonprofit
organizations both show wavelike patterns with an upward trend.
The growth rates for pension plans carried by agency life insurance
companies are practically the same as for all insured pension plans,
83-200-67-pt. 5-ii
PAGENO="0162"
154 OLD AGE INCOME ASSURANCE-PART V
which, as we have seen, follow the model rather closely. Only since 1958
has it been possible to break agency-insured pension funds down
into reserves for nonprofit organizations and for all others. The
period is rather short for detecting trends,'but all funds except those
`for nonprofit organizations include such a large proportion of all
agency-insured funds that one would expect the former to follow the
trend of the latter. In the 6-year period that expectation appears to
come true. In 4 of the 6 years insured funds for nonprofit organiza-
tions show quite different growth rates from those of all agency-in-
sured or all insured pension plans. This suggests that growth rates of
plans for nonprofit organizations, whether self-administered (non-
insured) or insured, and whether insured with TIAA-CREF or agency
companies, tend to have a wavelike pattern and probably will continue
for some time to show an upward trend. This is in `contrast to corporate
funds, whether insured or noninsured.
9. Ponmor~io Co~IPosITIoN
The investment portfolios of noninsured pension funds of nonprofit
organizations and corporations are distributed as shown in tables 11-6
and 11-7. Table 11-6 shows that in nonprofit funds the proportions of
common stock and mortgages increased and Government bonds and
preferred stock fell over the period 1951-64. The proportion of non-
Government (corporate and other) bonds increased through 1960, but
has since decreased slightly.'4
Comparison of tables 11-6 and 11-7 suggests that, on the average,
the investment managers of nonprofit and corporate funds have fol-
lowed similar policies, although the former have been slower in in-
creasing their holdings of common stock and decreasing their hold-
ings of non-Government bonds. In general, the nonprofit funds have
professional investment counsel, and their trustees include business-
men and bankers. However, ministers, Y secretaries, social workers,
and others are also on their boards, and some of~ them have conserva-
tive leanings. In addition, some of the funds must observe investment
restrictions that can only be changed at statutory me~tings of the
parent body. The proportion of the portfolio in mortgages has been
larger for nonprofit than for corporate funds. The latter, however,
have recently been increasing this proportion more rapidly than non-
profit funds.
Table Il-S shows that, in general, the. ratio of market to book
value was as high for. the nonprofit organization funds as for the
corporate funds.
Before 1962 the assets behind insured pension plans were entirely
commingled with all other assets of life insurance companies ancL
therefore, were subject to the same investment regulations. Greater
14 These distributions were obtained from aggregate figures. Individual funds show wide
variations, but the argregate does not seem to be dominated by large atypical funds. In
1964, approximately SO percent of the number of funds fell, within the following ranges:
11.5. Government securities 3-15 percent.
Corporate and other bonds 20-60 percent.
Preferred stock Less than 3 percent.
Common stock 10-40 percent.
Mortgages Less than 20 percent.
PAGENO="0163"
OLD AGE INCOME ASSURANCE-PART V
155
TABLE 11-6--PERCENTAGE DISTRIBUTION OF PORTFOLIOS OF NONINSURED PENSION FUNDS OF NONPROFIT
ORGANIZATIONS, SELECTED YEARS, 1951-64
1951
1955
1960
1964
Book value, end of year:
Cash and deposits
U.S Government securities
1.2
26.6
1.7
18.3
1.5
11.1
0.7
8.2
Corporate and other bonds!
Preferred stock
37.6
12.6
43.9
7.5
47.4
4.1
45.4
1.7
Common stock
15.6
20.4
21.1
26.2
Mortgages
Other assets
5.7
.7
7.6
.6
11.6
3.2
13.2
4.6
Total assets:
Percent
100.0
100.0
100.0
100.0
Amount (billions of dollars)
.4
.6
1.0
1.4
Market value, end of year:
Cash and deposits
U.S. Government securities
1.2
25. 5
1.5
15. 7
1.4
10. 0
.6
6. 7
Corporate and other bonds 1
Preferred stock
36. 1
12.3
38.6
6.9
40.6
3.3
39.0
1.4
Common stock
18.6
30.0
31.2
36.7
Mortgages
Other assets
5.6
. 7
6. 7
. 6
10.6
2. 9
ii. 5
4. 1
Total assets:
Percent
100.0
100.0
100.0
100.0
Amount (billions of dollars)
. 4
.7
1. 0
1. 6
Source: See app. II.
1 Includes World Bank, State and local government, eleemosynary, and foreign government and corporate as well as
U.S. corporate.
TABLE 11-7.-PERCENTAGE DISTRIBUTION OF PORTFOLIOS OF NONINSURED CORPORATE PENSION FUNDS,
SELECTED YEARS, 1951-64
1951
1955
1960
1964
Book value, end of year:
Cash and deposits
U.S. Government securities
4. 1
31.4
2.4
17.8
1.4
7.1
1.5
5.2
Corporate bonds
Preferred stock
44.7
4.0
49.9
3.6
48.4
2.3
41.1
1.2
Common stock
12.3
21. 2
33. 5
41.6
Mortgages
Other assets I
Total assets:
1
~
1. 4
3 7
3. 0
4. 3
4. 4
5.0
.
Percent
100.0
100.0
100.0
100.0
Amount (billions of dollars)
Market value, end of year:
Cash and deposits
U.S. Govoroment securities
Corporate bonds
Preferred stock
Common stock
Mortgages
Other assets
Total assets:
7. 2
(1)
(1)
(1)
(1)
(1)
(1)
(1)
14. 9
2. 1
15.6
43.7
3.3
30. 7
1.2
3.3
30. 3
1. 3
6. 2
40.2
1.9
43. 9
2.7
3. 8
47.3
1.2
4.2
32.4
1.0
53. 4
3.6
4.2
Percent
Amount (billions of doIIors)..~
(1)
(1)
100.0
16.7
100.0
34. 1
100.0
50.1
Source: "SEC Statistical Bulletin," June 1966, and SEC release of July 1964, "Corporate Pension Funds, Supple-
mental Tables."
1 Not available.
latitude in investments of pension funds kept in separate accounts is
now permitted in practically all States, but. only 0.3 p'cent of all
insured pension funds were in separate accounts by the end of 1964.
Therefore, the portfolio distribution of all life insurance assets is
used for pension funds of agency insurance companies in table 11-9,
PAGENO="0164"
156 OLD AGE INCOME ASSURANCE-PART V
TABLE II-8.-RATIOS OF MARKET TO BOOK VALUES FOR SECURITY INVESTMENTS OF NONINSURED PENSION
FUNDS OF NONPROFIT ORGANIZATIONS AND CORPORATIONS, SELECTED YEARS, 1951-64
1951
1955
1960
1964
Nonprofit organizations, end of year:
U.S. Government securities
Corporate and other bonds
Preferred stock
Common stock
Corporations, end of year:
U.S. Government securities
Corporate bonds
Preferred stock
Common stock
0. 97
. 98
.99
1.21
(1)
(1)
(1)
(1)
0. 97
1. 00
1.05
1.67
.98
.98
1. 02
1. 62
0. 98
. 94
.89
1.62
.99
.93
- 92
1. 47
0. 94
.98
.98
1.60
.99
.97
1. 02
1. 58
Source: Same as tables 11-6 and 11-7.
I Not ovailable.
and is comparable with the book-value panels of tables 11-6 and 11-7.
The main difference between insured and noninsured portfolios is in
common stock and mortgages. In 1964, agency insurance companies
held 3.6 percent of their assets in common stock and 36.8 percent in
mortgages, and the corresponding percentages were 26.2 and 13.2
f or noninsured funds for nonprofit organizations. The discrepancy
between insured and corporate noninsured was even greater. Informa-
tion is not now available for the separate accounts of insured pen-
sion reserves; but, it is probable that if their portfolio were substituted
for that of all insurance company assets, the discrepancy between
insured and noninsured pension fund portfolios would be smaller.
TABLE 1-9.-PERCENTAGE DISTRIBUTION OF PORTFOLIOS OF INSURED PENSION FUNDS, AGENCY INSURANCE
COMPANIES, AND TIAA-CREF, SELECTED YEARS, 1951-64
1951 1955 1960 1964
Agency life insurance companies, end of year:
Cash and deposits 1. 6 1. 4 1. 1 1. 0
U.S. Government securities 16. 2 9. 5 5. 4 3. 8
Corporate and other bonds 1 41. 9 43. 3 43. 6 41. 7
Preferred stock 2.1 1.9 1.5 1.7
Common stock 1. 2 2. 1 2. 7 3. 6
Mortgages 28. 2 32. 5 34. 8 36. 8
Other assets 8. 8 9. 3 10. 9 11. 4
Total assets 100. 0 100. 0 100.0 100. 0
TIAA-CREF, end of year: 2
Cash and deposits . 7 . 6 . 4 . 2
U.S. Government securities 1. 5 1. 2 . 7 . 3
Corporate and other bonds I 40. 5 41. 0 34. 9 29. 9
Preferred stock 1. 6 . 8 . 3 (3)
Common stock (3) 5. 2 16. 3 27. 0
Mortgages 52. 9 48. 4 43. 9 39. 7
Other assets 2. 8 2. 8 3. 5 2. 9
Total assets 100. 0 100. 0 100. 0 100. 0
Source: Based on Life Insurance Fact Books and annual reports of TIAA-CREF. See app. II.
1 See table 11-6, note 1.
2 CREF was established in 1952; therefore, 1951 figures include only TIAA.
3 Less than 0.05 percent.
Teachers Insurance & Annuity Association is a limited-eligibility
life insurance company concentrating on insured pension plans. Its
companion organization, College Retirement Equities Fund, was es-
tablished in 1952 to provide variable pension annuities, and its assets
are, therefore, all invested in common stock. The 1951 distribution is
for TIAA alone and shows a considerably larger proportion of assets
PAGENO="0165"
OLD AGE INCOME ASSURANCE-PART V 157
in mortgages and a smaller proportion in Government bonds than
the corresponding distribution for agency companies. TIAA varied its
portfolio distribution only slightly over the years 1951-64, but, with
the addition of increasing amounts in CREF, the lower panel of table
11-9 shows a large growth in common stock and decreases in nongov-
ernment bonds and mortgages. These trends are expected to continue,
especially since a change in TIAA-CREF rules allows as much as 75
percent of each pension premium due on or after January 1, 1967, to
be allocated to CREF and the remainder to TIAA. Previously not
more than 50 percent could go to CREF. The TIAA-OREF portfolio
for 1964 shows 27 percent in common stock, practically the same as
the corresponding book-value distribution of noninsured funds of non-
profit organizations. The main difference between the two funds is
TIAA-CREF's larger holdings of mortgages and lower holdings of
bonds.
In table 11-10, all the insured and noninsured pension funds of non-
profit organizations are combined. The table shows that over the years
1958-64 Government securities, other bonds, and preferred stock fell
slightly while common stock rose markedly and mortgages showed lit-
tle change. In spite of these changes, in each of the years shown, the
three most important items remained in the same order: Corporate and
other bonds, mortgages, and common stock.
TABLE 11-10.-PERCENTAGE DISTRIBUTION OF PORTFOLIOS OF TOTAL PENSION FUNDS FOR NONPROFIT
ORGANIZATIONS, YEAREND, 1958-64
1958 1959 1960 1961 1962 1963 1964
Total assets (billions of dollars) 1.7 1.9 2.2 2.4 2.7 3.0 3.4
Cash and deposits 1. 1 1. 1 1. 0 .9 .9 . 7 .6
U.S. Government securities 6. 3 6. 7 6. 3 6. 2 4. 9 4. 3 4. 3~
Corporate and other bonds 44.1 43.5 42.3 41.4 40.9 40.2 38.5
Preferred stock 3.4 2.8 2.2 2.0 1.6 1.2 1.0
Common stock 12.8 13.7 15.4 17.0 18.9 20. 4 22.0
Mortgages 27. 8 27. 7 27. 8 27. 4 27. 2 27. 5 28. 1
Other assets 4.5 4.5 5.0 5.1 5.6 5.7 5.5
Source: Based on combined portfolio for (1) noninsured funds of nonprofit organizations plus (2) funds insured with,
TIAA-CREF and (3) agency-insured funds for nonprofit organizations. See app. II.
Table TI-il shows that the funds were large purchasers of each of
these three types of securities. In fact, they purchased $100 million or
more of each type in 1964, more than a half billion dollars each of
nongovernment bonds and common stock, and only slightly less of
mortgages over the 6-year period.
TABLE Il-il-USES OF TOTAL PENSION FUNDS BY NONPROFIT ORGANIZATIONS IN TIlE CAPITAL MARKETS,
1959-64
[In millions o
f dollars]
1959
2960
1961
1962
1963
1964
Cash and deposits
U.S. Government securities_....
Corporate and other bonds ---
Preferred stock
Common stock
+1
+20
+74
-4
+43
+2
+9
+87
+71
-1
+14
+77
(1)
+76
+3
-21
+99
+96
-3
-1
+103
+103
-3
+15
+100
(1)
+136
Mortgages
Other assets
Total
+53
+10
+197
+74
±22
+260
+57
+15
+238
+67 +92
- +27 +19
+ii~~~ - +305
±127
- ±15
+390
Source: Based on the same data as table 11-19.
1 Less than 0.5.
PAGENO="0166"
158 OLD AGE INCOME ASSURANCE-PART V
Similar data are given for multiemployer and union funds in table
1-8, above, for the years 1960-64. In general, the nmltiemployer and
union funds are newer than the nonprofit funds, so they would be ex-
pected to increase more rapidly. Actually, the multiemployer and
union funds increased from $1.3 billion in 1959 to $3 billion in 1964.
The corresponding figures for nonprofit organizations are $1.9 1)1111011
and $3.4 billion. The uses of funds for the two types of pension funds
for the combined period 1960-64 are as follows (in millions of
dollars)
rmltinmployer Nonprofit
and union organization
fundo fundo
Caoh and depoolts
U.S. Government oecuritins
Corporate and othor bonds
Preferred otnck
Common stock
Mortgages
Other assets and isvestments
±108
±17
±610
±9
±536
±427
±63
-2
±16
±466
-18
±482
±417
±98
Total
±1,770
±1,459
The nonprofit organization pension funds have less need for cash
since their expenses are usually paid by the parent organization. Other-
wise, the two distributions are fairly similar.
RATE OF RETURN ON INVESTMENTS
To measure investment performance, most fund managers compute
a. rate of return on investment, usually the ratio of interest and clivi-
dends to mean invested assets (including cash) less half the invested
income. Some use gross investment income; others subtract out amorti-
zation, depreciation, mortgage service fees, and investment manage-
ment fees. Another difference is whether the ratio is based on book
or market value, or some combination of the two.
Table 11-12 presents comparable rates for pension funds of non-
profit organizations and other pension arrangements. It shows that
the highest mean yield was realized by TIAA, which had the largest
proportion of bonds and mortgages and the lowest proportion of
stock. At the other extreme is corporate pension funds with the lowest
yield, highest proportion of stock, and lowest proportion of bonds
and mortgages. Noninsured firnds of nonprofit organizations show a
slightly higher mean yield than corporate pension funds, a somewhat
lower portfolio proportioll in stock, and a somewhat higher pro-
portion in bonds and mortgages.
The table bears out Dietz's statement: "A measure of performance
based only on ordinary income `~ is misleading when trying to compare
two or more funds. The fund invested in equities would have been un-
duly penalized during the 10-year period of this study (1953-62) be-
cause equities generally produced a lower rate of present return (com-
pared to bonds), with the expectation of a future increase in value." 16
Table 11-13 and chart 11-1 present. rates of return computed ac-
cording to Dietz's preferred formula. He defines investment income as
uI.e.. interest and dividends.
~ See Peter 0. Dietz, Pension Funds: Measuring Investment Performance, New York,
19Gi;, i 49.
PAGENO="0167"
OLD AGE INCOME ASSURANCE-PART V 159
TABLE 11-12.-RATIO OF GROSS INCOME FROM INTEREST AND DIVIDENDS TO BOOK VALUE
OF TOTAL ASSETS FOR NONINSURED PENSION FUNDS OF CORPORATIONS AND NON-
PROFIT ORGANIZATIONS AND FOR AGENCY INSURANCE COMPANIES AND TIAA, 1958-64
[In percent]
Noninsured p
ension funds of-
~
Agency insur-
ance companies
TIAA
Corporations
Nonprofit
organizations
1958
1959
1960
1961
1962
1963
1964
3. 91
4.01
4. 08
4.08
4. 06
4.08
4.17
3. 94
4.05
4. 17
4.33
4. 27
4.36
4.36
4. 12
4.34
4. 48
4.57
4. 72
4.85
4.97
4. 26
4.51
4. 96
5.12
5. 26
5.36
5.42
Mean
4. 06
4. 21
4. 58
4. 98
Note: The denominator of the ratis is the arithmetic mean of total assets at the beginning and
end of the given year less half the corresponding gross income from interest and dividends.
Source: Corporatisos: Computed from "SEC Statistical Bulletin," June 1964 and 1965and SEC
release of July 1964, "Corporate Pension Funds, Supplemental Tables."
Nonprofit Organizations: See app. II.
Agency insurance companies and TIAA: Competed from Annual Reports of the Superintendent
of Insurance, State of New York, vol. I-A, 1958-65, New York State companies and companies of
other States licensed to do business in New York. Balance sheets were changed to include book
value rather than market value of stock.
CHART 11-1.-Ratio of Gross Income from Interest and Dividends Plus Realized
and Unrealized Gains and Losses to Market Value of Total Assets for Non-
insured Pension Funds of Corporations and Nonprofit Organizations and
CREF, 1958-64
Percent
45
40-\
Corporations
- \ Nonprofit organizotions -
CREF
30- -
25- -
20- \ -
I'
15-
* I..- I
/
i:
0
`-..--. I
-5- \ I -
-10- -
V
-15 I I ._ ________ I
1958 `59 `60 `61 `~2 `63 `64
Source: Table 11-13.
PAGENO="0168"
160 OLD AGE INCOME ASSURANCE-PART V
"ordinary income plus realized and unrealized gains and losses" or as
M2 -M1 - C, where M1 is market value of the fund at the beginning
of the period, M° is market value at the end of the period, and C is
net contributions to the fund. The formula naturally uses market
value also in the denominator (L + 1/2C) of the ratio.
TABLE 11-13.-RATIO OF GROSS INCOME FROM INTEREST AND DIVI-
DENDS PLUS REALIZED AND UNREALIZED GAINS AND LOSSES TO
MARKET VALUE OF TOTAL ASSETS FOR NONINSURED PENSION
FUNDS OF CORPORATIONS AND NONPROFIT ORGANIZATIONS AND
CREF, 1958-64
[In percent[
Noninsured pen
sion funds of-
Nonprofit CREF
organizations
Corporations
1958 13.70 9. 72 41. 52
1959 5.86 5.79 14.04
1960 5. 87 6. 94 4. 63
1961 14.79 10.86 17.91
1962 -2.68 1. 07 -12. 41
1963 11.17 9.33 18.19
1964 10. 61 7. 18 12. 37
Mean 8.47 7.27 13.75
Note: The formula is shown in the text. Net contributions (C) is the
difference between total receipts from dues, donations, annuity agreements,
etc., and all payments, grants, and expenses.
Source: Same as for table 11-12 for corporations and nonprofit orga-
nizations. CREF: Computed from TIAA-CREF annual reports.
Table 11-13 omits the insurance companies because they do not pre-
sent data on market value of all bonds and stock. It includes CREF,
which was omitted from table 11-12 for lack of data in terms of book
value. The highest mean ra.tio in table 11-13 was obtained by CREF,
whose portfolio was invested almost entirely in common stock. The
noninsured funds of nonprofit organizations realized the lowest mean
rate of return of the three groups shown in the table. They had the
smallest proportion of portfolio in stock and the largest proportion
in bonds and mortgages. If data were available for TIAA and agency
insurance companies, their mean rates of return would in all proba-
bility be considerably lower tha.n that of noninsured funds of non-
profit organizations because they were invested more in bonds and
mortgages and less in stock than the latter.
Chart 11-1 shows the variability of returns over the years 1958-64.
This is commonly considered a measure of risk. Comparison of the
mean yields from table 11-13 with the chart shows that the higher
the mean rate of return, the higher the risk. Thus CR-EF, with the
highest mean rate (13.75 percent), showed a loss rate of 12.41 per-
cent in 1962 and a gain rate of 41.52 percent in 1958. The noninsured
pension funds of nonprofit organizations, on the other hand, with the
lowest mean rate (7.27 percent), showed the lowest variability. As
mentioned above, TIAA would probably have a still lower mean rate
of return. The observed relationship between mean rate of return and
year-to-year variability in the rate suggests that TIAA would also
have a lower variability or risk than any shown on the chart. Thus
participants in CREF who must also participate in TIAA have prob-
PAGENO="0169"
OLD AGE INCOME ASSURANCE-PART V 161
ably had a lower mean rate of return and risk of not receiving such
a return annually than shown for CREF. How much lower these rates
would be would, of course, depend on how they had chosen to divide
their contributions between the two companion organizations. It is
possible that the return to the average participant in nomnsured pen-
sion funds of nonprofit organizations and to the average participant
in TIAA-CREF might be similar. The return to the average partici-
pant in pension plans carried with agency insurance companies would
probably be lower unless their reserves were kept in separate accounts.
The series for noninsured pension funds of nonprofit organizations
shown in table 11-13 and the chart are aggregates of numerous funds.
Managers of the individual funds may wish to apply the formula to
their data and compare the results with those shown in the table.
10. CoNcLUsioNs
Although many of the pension funds of nonprofit organizations
have been in existence 40 to 50 years or longer, there are good reasons
for believing that the group as a whole will continue to show a fairly
substantial growth rate. This is in contrast to the normal pattern as
shown by corporate pension funds. The latter have been increasing,
but at a consistently declining rate.
There are two major reasons for the expected steady growth in non-
profit pension funds: first, only about one-third of all units of non-
profit organizations had pension plans at the end of 1960 and only
about one-fifth of the employees were eligible for coverage; second,
once some individuals have the prospect of a small income after retire-
ment, they realize they need more. As pointed out by Cagan,17 econo-
mists are aware of the tendency of group pension plans and GI insur-
ance to cause certain individuals to increase their saving in other forms.
Employees of nonprofit organizations other than ministers only be-
came eligible for OASI coverage in 1951, and ministers in 1955. Most
employees are now covered, also a large proportion of Protestant min-
isters, rabbis, and some Catholic priests. For those covered by OASI,
but not by a private plan, it is not difficult to see that income after
retirement will probably be ~ow compared to needs. In general, direc-
tors of nonprofit organizations are aware of this fact and are trying
to establish pension plans or raise low benefits through increased pre-
mium asse~ments.
The groups for which pension funds are expected to expand
markedly are lay employees of religious bodies; lay teachers and other
employees of parochial schools and private schools; hospital workers,
especially registered nurses and nonprofessional employees other than
clerical workers and nonprofessionals in Catholic and Protestant chari-
table organizations. There are also indications of substantial future
growth in funds for retirement or support of aged Catholic priests.
It must be noted that the expected growth in pension funds of non-
profit organizations will not all show up in the figures on private non-
insured funds, since over half the funds were insured in the years
17 See Phillip Cagan, The Effect of Pension Plans on Aggregate Saving: Evidence from a
Sample Survey, Occasional Paper 95. New York, National Bureau of Economic Research,
1965, ~p. 5f.
PAGENO="0170"
162 OLD AGE INCOME ASSURANCE-PART V
1958-64. While in the past some of the plans insured with agency
companies have changed over to noninsured funds, and this trend is
likely to continue, when smaller organizations set up plans they will
probably be insured. Also, TIAA and CREF have such a large propor-
tion of the higher educational field and the advantage of portable
pensions that few if a.ny of their funds are likely to be transferred to
noninsured funds.
A combined portfolio of all pension funds of nonprofit organiza-
tions amounted to $3.4 billion at the end of 1964, with 39 percent in-
vested in corporate and other bonds, 28 percent in mortgages, and 22
percent in common stock. It should be noted that these figures include
noninsured funds at book value and CREF at an estimate of book
value computed only in this paper. When market values are substituted
for the two series, the total is $3.7 billion, with 30 percent invested in
common stock; 35 percent in bonds, excluding U.S. governments; 26
percent in mortgages. In that year the combined funds purchased $136
million in common stock, $127 mfflion in mortgages, and $100 million
in corporate and other bonds. The expected sustained rate of growth
in total pension funds of nonprofit organizations suggests a continued
flow of funds to the securities markets.
PAGENO="0171"
OLD AGE INCOME ASSURANCE-PART V 163
APPENDIX II. SOURCES AND LIMITATIONS OF DATA
UNITS, EMPLOYEES, AND PAYROLL
Estimates of the population of employees of nonprofit organizations
were mainly taken from the 1959, 1962, and 1964 editions of the U.S~
Bureau of the Census County Business Patterns. These publications
show number of employees as of mid-March pay period, the taxable
payroll for January-March, and number of units reported under the
Federal Insurance Contributions Act for the old-age, survivors, and
disability insurance program. The data are presented in the Stand-
ard Industrial Classification three-digit code. All Government employ-
ees are excluded. While data for nonprofit organizations are included
only if they are covered under the elective provisions of the act, the
1959 publication says that nearly all eligible employees are now cov-
ered by the program.
Certain modifications, which had to be made for the purposes of
this report, are shown below. Units (in nonmanufacturing indus-
tries) are counted only once in each county. Therefore, units for re-
ligious bodies and for elementary and secondary schools had to be
compiled from other sources.
Religious Bodies
The number of units comes from the Yearbook of American
Churches, published annually by the National Council of Churches of
Christ in the U.S.A. The number of employees shown in County Busi-
ness Patterns for religious organizations refers to lay employees only.
Estimates for ministers, priests, and rabbis were obtained from Year-
book of American Churches and an interview with its editor, Benson
Y. Landis; from The Official Catholic Directory, published annually
by P. J. Kenedy & Sons; and from H. S. Linfield, The Rabbis of the
United States, published by the Jewish Statistical Bureau in 1957.
The Linfield data are for 1954, but estimates of rabbis active in re-
ligious work were made for 1960 on the basis of the increase since a
previous study (1927) made by the same author.
A breakdown of lay employees into major religions bodies served,
was estimated from a study reported by F. Ernest Johnson and J.
Emory Ackerman in The Church as Employer, Money Raiser, and
Investor (New York, 1959), from a conversation with an official of a
large Catholic archdiocese and information in The Official Catholic Di-
rectory on Catholic population and number of parishes, and from data
on the number of Jewish congregations in the Yearbook of American
Churches and other information in The Rabbis of the United States.
Educational Institutions
Correspondence and vocational schools were subtracted from educa-
tional services on the assumption that the majority were proprietary
organizations; museums, art galleries, botanical and zoological gar-
dens, and nonprofit research agencies were added.
A list of nonpublic senior colleges and universities was compiled
from Retirement and Insurance Plans in American Colleges by Wil-
liam C. Greenough and Francis P. King (New York, 1959), and non-
public junior colleges were added from the TVorld Almanac, 1960. For
Catholic colleges the number of lay teachers was added from The
PAGENO="0172"
164 OLD AGE INCOME ASSURANCE-PART V
Official Ccit/iolic Directory, and for all other colleges, total teachers
from the World Almanac. Colleges with no lay teachers or no infor-
mation on number of teachers were omitted. The following results
were obtained:
Colleges
Teachers
Nonprofit
Proprietary
1,061
36
113.603
1,400
Total
1,097
115,000
From County Business Patterns, it would appear that the total nuin-
ber of private colleges should be between 1,500 and 1,600 and total
employees around 245,000. It is estimated. therefore, that in 1960
there were about 1,500 private nonprofit colleges with the following
breakdown of total employees : 18
Faculty 130, 000
Administrative officers 27, 000
Nonacademic employees 83, 000
Total 240, 000
The number of elementary and secondary school units and lay
teachers was compiled from The Official Catholic Directory. 1960: the
January 1962 issue of the. Lutheran Church-Missouri Synod's
Paris/i. Education Bulletin; Why a Pension. Pro qram for Teachers
and Administrators in Jewish Schools. November 1957, pamphlet. of
the American Association for Jewish Education; and Pr-irate Inc/c-
per-dent Sc/tools, 1960. published by James E. Bunting and The Hand-
booh of Private Se/tools, 1960, by Porter Sargent.
Hospitals
Proprietary hospitals were excluded from the County Bwsiness
Patteui?s data by using information in the August guide issue of
Hospitals, the. journal of the American Hospital Association.
Other Son profit Organizations
All data are from County Business Patterns: Nonprofit member-
ship organizations excluding religious organizations.
Annual Payroll
The taxable payroll for the first. quarter of the year was multiplied
by 4. During the period of study, the first $4,800 of wages was taxable
under OASDI. Therefore, the annual payroll figures are too low only
for employees with annual wages of over $19,200. The average annual
wage for employees of nonprofit organizations in 1964 appears to be
less than $4,000. This includes an amount for ministers, priests, and
rabbis estimated from the 1964 National Council of Churches survey of
ministers' salaries in 15 Protestant denominations. The survey showed
a median cash salary of $5,158.'° This is probably too low for the
denominations covered, since it does not. include certain fringe benefits
18 Some help in breaUng down nonafaculty employees was obtained from H. Robert Bolcel-
man, Higher Education: Planning and Management Data, 1957-58, U.S. Department of
Health, Education, and Welfare. Circular No. 517, 1958.
in Reported in tine 1964 Annual Report of the Annuity Fund for Congregational Ministers,
p. 14.
PAGENO="0173"
OLD AGE INCOME ASSURANCE-PART V 165
such as rent. Also a mean figure would probably `be higher. On the
other hand, it is probably too high for certain other denominations,
especially those with many part-time ministers. The total annual
payroll for all nonprofit organizations is $10.2 billion 1± the $o,158
figure is used for ministers, priests, and rabbis, or $10.5 billion if
$6,500 is substituted.
PENSION COVERAGE INFORMATION
Protestant Mini sters
Pension information for Protestant religious bodies was obtained.
from the statistical reports, minutes, and other materials distributed
at the annual meetings of the Church Pensions `Conference and from
annual reports and other information sent by the member pension
boards. Nonmember Protestant denominations were checked against
the Yearbook of American Churches and questionnaires sent to 18
reporting 1,000 or more pastors. One was returned by the Post Office,
five did not answer, and one reported no plan. The remaining 11 sent
some information on modest funds. Several small funds have since
been discovered for denominations with less than 1,000 ministers. It
appears that any undiscovered funds would probably not raise our
figures much. The larger funds generally gave information for the
years 1957-64. Data back to 1945 were obtained from Church Pensions
Conference reports and some earlier data from the New York (State)
I'nsurance Reports.
Cat holic Priests
Father Joseph M. Becker. S.J., professor at St. Louis University, on
a visit to the National Bureau advised the author that priests are cared
for by their orders even when they are no longer active, but that retire-
ment arrangements for diocesan priests vary among the 112 dioceses
and 26 archdioceses. A check of The Official Catholic Directory. 1960,
showed that 72 jurisdictions with 18,600 diocesan priests listed an
organization with a title such as Infirm Priests' Fund or Priests'
Mutual Benefit Society. Questionnaires were sent to such organizations
in 18 dioceses and archdioceses with 9,400 priests. In reply, five dio-
ceses with 1,435 priests reported pension plans and five others with
4,443 priests reported plans for medical care and hospitalization but
not for retirement. In other words, almost a quarter of the priests
were in dioceses with plans. On the assumption `that a unit with a plan
is more likely to report than one without, we estimate pension coverage
at about 20 percent in dioceses with clerical funds but at only 10 to 11
percent in the country a~s a whole. Three of the reporting plans had
funds of less than $100,000 each; the other two were unfunded. This
suggests that despite the small size of our sample, the total probably
is not greatly affected.
Rabbis
Correspondence with five national rabbinical associations brought
three replies, each reporting an insured pension plan. One reported
coverage of 20 percent, another 80-85 percent, and the third did not
answer the question.
PAGENO="0174"
166 OLD AGE INCOME ASSURANCE-PART V
Clergy of Other Religioz&s Bodies
The two largest denominations in other religious groups each
reported 400 or fewer clergymen to Yearbook of American Churches
in 1957. Replies to our questionnaire were not received from either
of those denominations.
Lay Employees of All Religious Bodies
Lay members of pension funds appear in the statistical reports of
the Church Pensions Conference, but there is no breakdown between
church secretaries, sextons, organists, and so on, and teachers and
other employees iii church-affiliated schools, colleges, hospitals, and
homes. Some discussion of la.y coverage is given in the annual reports
of the denominational pension boards.
Correspondence brought information on no diocesewide pension
plans for lay employees of the Roman Catholic Church except the one
started by the Archdiocese of New York in 1962.
No data were available on pension coverage of Jewish lay employees,
but their number is small compared to total lay employees of religious
bodies; so the lack of information is probably not serious.
Colleges and Universities
Retirement and In-surance Pio'n.s in American. Colleges by William
C. Greenough and Francis P. King (New York, 1959) lists 4-year
colleges and gives detailed information about retirement plans avail-,
able for their faculty members, administrative officers, and nonaca-
demic employees. The majority of the private nonprofit colleges were
listed as having TIAA-CREF plans. Other plans were with agency in-
surance companies (see agency life insurance data below), included
with funds for Protestant ministers and other lay employees, or self-
administered plans usually for individual colleges.
Letters were written to all 30 private colleges listed as having
self-administered plans, including some for which the noninsured plan
was for nonacademic employees only or was an alternative to an in-
sured plan, and also to the four colleges with 200 or more faculty mem-
bers for which information was not available in Greenough and King.
Replies were received from all four of the latter colleges and from 20
of the colleges with 80 percent of the teachers in the noninsured-plan
group. Two of the plans for nonacademic employees were insured, two
others had been dropped, and another two were unfunded. One un-
funded plan was also reported for faculty and administrative officers
of a. small college. The other respondents sent information on total
assets and in some cases on portfolio distributions. Again it would
appear that the missing data for funded plans were probably small
compared to the given information.
Data for junior colleges come from Francis P. King's article, "In-
sured Staff Benefit Plans in the Jlmior Colleges," in the Junior
College Journal, September 1960. Some junior colleges were also listed
in Greenough and King, among church plans, and in the list of par-
ticipating institutions in the TIAA-CREF annual reports.
Elementary and Secondary Schools
The TIAA-CREF annual reports also list school participants.
Other information was obtained from correspondence with Catholic
diocesan school boards, religious orders that run private schools in
PAGENO="0175"
OLD AGE INCOME ASSURANCE-PART V 167
more than one diocese, the American Association for Jewish Educa-
tion, the National Union of Christian Schools, the Board of Support
and Pensions of the Lutheran Church-Missouri Synod, and the Na-
tional Council of Independent Schools. Schools and their employees
are also included in the agency life insurance data (see below), but no
breakdown of educational services is given.
Other Educational Services
Most of the pension information for other educational services
comes from the agency insurance company data (see below) and
TIAA-CREF reports. We have also had correspondence with the
American Association of Museums and the American Association for
State and Local History, both of which conducted surveys among their
members around 1960-61 on pension and insurance provisions for their
employees.
Hospitals
Numerous investigations have been made of hospital pension cover-
age. The study has profited from data in the American Hospital Asso-
ciation's Hospital Salary Survey, 1956, the survey conducted by Louis
S. Reed in New York State in November 1958,20 and the periodic
studies carried on by the Bureau of Labor Statistics.2' Included in the
BLS studies in 15 large metropolitan areas (16 in 1956-57) are per-
centages of workers employed by nongovernment 22 with
pension plans (other than OASDI) for employees of given occupa-
tional category. If a pension provision applied to at least half of the
workers in the given category in the given hospital, all such workers
were included; otherwise, none were included. The 1963 survey also
gives percentages for the total of all standard metropolitan statistical
areas in the country. On the basis of the number of employees in the
given categories in 1963 and the relationship between 1963 coverage
rates for total areas and for the 15 studied, an estimated 1960 U.S.
coverage rate was computed for each category.
Correspondence with hospital associations in areas with high cover-
age gave some names of hospitals with noninsured plans and names of
agency life companies carrying hospital pension plans. A few teaching
hospitals are included in the TIAA-CREF 1960 report.
Other Nonprofit Organisation~
Pension data on other nonprofit organizations come from the agency
life insurance company data (see below); annual reports of pension
funds for certain well-known charitable organizations; correspondence
with organizations listing 100 or more employees in Jay Judkins, Na-
tional Associations of the United States, U.S. Department of Com-
merce, 1949; and also correspondence with Jewish federations and
Catholic chanceries. The TIAA-C1REF report for 1960 shows about
40 professional, technical, historical, civic, and charitable organiza-
tions; but in general, they have few employees.
Agency Life Insurance Data
A list was compiled of agency insurance companies mentioned by
Greenough and King and by various nonprofit organizations as in-
20 See "Where Hospitals Stand on Employee Benefits," Hospitals, Sept. 1, 1959..
~` Bulletins No. 1210 (Earnings and Supplementary Benefits in Hospitals, 1956-57), No.
1294 (Mid-1960), and No. 1409 (Industry Wage Survey-Hospitals, 1~fid-1963).
Only hospitals with 100 or more employees were included.
PAGENO="0176"
168 OLD AGE INCOME ASSURANCE-PART y
surers of their pension plans. Questionnaries were sent to the 33 coin-
panies on the list and to nine companies not reported by any of our
respondents, but shown as having substantial group annuity operations
or group permanent contracts for funding retirement plans in the
April and June 1962 issues of Em~ployee Ben4t Plan Review. Infor-
mation was requested as of December 1960 on number of contracts,
number of employees covered, and amount of reserves for nonprofit
organizations in the SIC two-digit categories: medical and other
health services; educational services; museums, art galleries, and so
forth; and nonprofit membership organizations. Unfortunately, we
did not ask for the three-digit category, nonprofit educational and
scientific research agencies, which is included under the two-digit code
miscellaneous services. Thirty-five companies (83 percent) replied,
but two of them had no plans tha.t fitted the definition, nine said they
did not have the data in the form requested, and two were able to
send totals but not breakdowns into the four categories. In general,
the companies that were mentioned most often reported the largest
totals. That fact is encouraging, since some nonprofit organizations
reported insured pension plans but did not name the insurer. The ques-
tionnaire was also returned by the National Health and Welfare
Retirement Association.
Life insurance companies have been obliged to include, in the income
tax returns for taxable years beginning after Dece1nber 31, 1958, the
amount of pension plan reserves in four categories at the beginning
and end of the taxable year. Category D is for employees of 501(c) (3)
organizations. V\Te were unable to get the totals of these reserves from
the Internal Revenue Service in 1965. Therefore, the same companies,
queried in 1960, were asked to send us data they reported for the years
1958-64. Fourteen companies sent data for 1958 or 1959 to 1964. The
figures were raised to include 12 companies with reserve data on only
1 or 2 years in the period and three companies that reported number
of employees covered but not size of reserves.
The life insurance companies were most cooperative and their corn-
ments were helpful even when they could not send figures. In many
cases, insured pension plans for 501 (c) (3) organizations had been
qualified as trusts described in section 401 (a), and it was not possible
for them to separate out their reserves. Also, in some cases plans
funded by individual policies are excluded. In general the companies
that sent data for the end of 1960 made special surveys in order to
collect the data. Therefore, the figures for other years have been re-
vised on the basis of the 1960 figures. For the later years of the period,
some companies were unable to deduct reserves for public schoolteacher
plans qualified under section 403 (b).
PENSION FUND ASSETS AND RESERVES
Total pension funds for nonprofit organizations include:
1. Funds insured with TIAA-CREF (from annual reports). This
is the sum of (a) TIAA reserves for life annuities and other periodic
payment plans (previously shown separately as reserves for annuities
and for settlement plans) and (b) CREF cumulations of periodic and
single premiums, accumulations transferred from TIAA, and divi-
dends on common stock less operating and investment management
PAGENO="0177"
OLD AGE INCOME ASSURANCE-PART V 169
fees, payments to participants, and organizational expense. Increases
in market value of common stock holdings are excluded, except when
noted that market values are used.
2. Funds insured with agency companies. This is the NBER tabula-
tion of pension reserves for nonprofit organizations on replies to ques-
tionnaires. See Agency Life Insurance Data. The questionnaires were
sent mainly to insurance companies reported by nonprofit organiza-
tions as insurers of their pension plans. In 1964 five of the companies
accounted for 80 percent of the total reserves.
3. Noninsured funds. This is the book value (market value when
noted) of total assets of all noninsured pension funds for nonprofit
organizations that could be obtained.
DATA FOR TABLES
Reference to appendix II in table notes are listed here:
Table IT-i. Nonprofit funds, see "Pension Fund Assets and Reserves,"
above.
Table 11-2. See "Pension Fund Assets and Reserves," above. All
TIAA-CREF funds were assigned to educational institutions ex-
cept small amounts estimated for a few teaching hospitals and
approximately 40 professional, historical, charitable, etc., organiza-
tions most of which had -few employees. Agency-insured funds were
assigned as reported except that those for nonprofit membership
organizations had to be divided between religious bodies and other
nonprofit organizations. The former were estimated from letters
and reports from such bodies mentioning insured funds and cover-
age rates.
Table 11-3. Ministers, priests, and rabbis, see "Units, Employees and
Payroll-Religious Bodies," above. Coverage, see "Pension Cover-
age Information," above.
Table 11-4. Nonprofit organizations, see "Pension Fund Assets and
Reserves," section 3.
Table 11-5. TTAA-CREF, and agency insured funds for nonprofit
organizations, see "Pension Fund Assets and Reserves," sections 1
and 2.
Table 11-6. See "Pension Fund Assets and Reserves," section 3. Com-
posite portfolios in terms of book and market value were compiled
for each year for the funds for which both sets of data were avail-
able. The proportion of market to book was then applied to total
book value to get total market value.
Table 11-9. Agency life insurance companies, assets of U.S. life in-
surance companies (from Life Insurance Fact Books) less TIAA
(from Annual Reports).
TIAA-CREF: Sum of (1) TTAA: proportions of total port-
folio applied to reserves for annuities and settlement plans and
(2) CREF: portfolios corrected by subtracting (a) "deposits"
and "due after yearend for stocks purchased" from cash (if
negative, use zero and see (c)) ; (b) other liabilities from other
assets; and (c) cumulated increase in market value and excess
of subtrahend over minuend in a from common stocks at
market value.
S3-200-67-pt. 5-12
PAGENO="0178"
170 OLD AGE INCOME ASSURANCE-PART V
Table 11-10. See "Pension Fund Assets and Reserves," above, also
rfI~_CREF under note to table 11-9. Agency-insured funds of
nonprofit organizations were distributed according to proportion
of total portfolios of United States life insurance companies (from
Life I~surance Fact Books) less TIAA (from. Aimual Reports).
Table 11-12. Noninsured funds for nonprofit organizations, see "Pen-
sion Fund Assets and Reserves," section 3. The proportion of funds
in book value for which income data were available in different
years ranged from 81 to 90 percent.
PAGENO="0179"
COMMENTS OF AMERICAN LIFE CONVENTION, BY Dr.
ARTHUR 5. FEFFERMAN and LIFE INSURANCE ASSOCIATION
OF AMERICA, n~ Dr. JAMES J. O'LEARY
The American Life Convention and the Life Insurance Association
of America welcome the opportunity to comment on the issues raised
by this Print, which was prepared by staff of the Joint Economic
Committee. We believe that a subject of such importance to the welfare
of the Nation merits the most careful consideration. The objective is
to secure the best possible retirement program for our older population,
consistent with our economic resources and with our national goals of
growth and economic stability. Our two organizations are directly
concerned with this objective.
The main thrust of our comments will relate to private pension fund
saving and investment and their implications for economic growth
and stability (part II of this paper). As an introduction to those com-
ments, we will say a few preliminary words, which are not intended
to be exhaustive, about our present retirement system and the achieve-
ments of the private `pension system (part I of this paper).
I. OUR PRESENT RETIREMENT SYSTEM AND THE ACHIEVEMENTS OF
THE PRIVATE PENSION SYSTEM
We believe that `any thoughtful examination of the retirement area
should give recognition to the great achievements of our present retire-
ment system. Essentially, that system is a mixed one, provided in part
by the public sector and in part by the private sector of the economy.
The social security system historically has provided a floor of pro-
tection; private pension plans and `individual savings for retirement
supplement this basic floor of protection.
Private pension plans offer unique advantages in this retirement
system. Built on top of the floor of protection offered by social security,
they possess a flexibility and a capacity for meeting individual retire-
ment needs that cannot be matched by any mass public plan. They
provide a means by which private enterprise, working through a
voluntary system, can experiment and make adjustments to suit the
individual circumstances of particular groups of employees in different
firms, industries, and geographical locations. In general., they relate
retirement benefits to the particular financial needs of individuals in
different' circumstances. Pension plans also have the advantage of
being financed out of private funds, without the use of public funds
derived from taxes. And, as will be developed later, they have the great
virtue of helping to supply the savings and capital required for a
dynamic growing economy.
Because private pension plans have these advantages, the life in-
surance business believes that they should play an expanding role in
the provision for retirement income in the future. To achieve this, two
factors are necessary.
171
PAGENO="0180"
172 OLD AGE INCOME ASSLRAXCE-PART V
First, the social security system and private pension plans must be
kept in proper balance. A social security system which provides a
floor of protection is vita.l for retirement purposes. But inordinate in-
creases in the social security wage base and in social security benefit
levels could impair the private pension system by invading areas which
can be better serviced by the private sector. We believe that the average
earnings of regularly employed male workers represents an appropriate
dividing line between the area in which the Government should have
primary responsibility to provide basic economic security and the area
in which the individual and his employer should have full respon-
sibility to provide security through private media such as pension plans.
Once basic retirement needs are met, it is undesirable to compel
individuals to finance additional retirement benefits under the social
security system by paying higher taxes. Instead, at this point the in-
dividual should retain freedom of choice to determine both the extent
to which lie wishes to set aside additional funds to provide higher
retirement income for himself as well as the particular form that such
retirement protection should take. Such supplementary savings for
retirement can best be achieved through voluntary means in the private
sphere of the economy.
Second, it is important that all interested parties continue to search
for ways to accelerate the growth and improvement of private pension
plans. In recent years, there has been considerable interest in this area.
The hearings on pensions which the Joint Economic Committee held
last year, the present document on Old Age Income Assurance pre-
pared by the Committee's staff, and the 1965 Report of the President's
Committee on Corporate Pension Funds-all represent part of an ex-
amination of pension plans, and a search for ways to improve them.
The life insurance business welcomes this examination since, as one of
the major funding media for pension plans, we have a direct concern in
their improvement. However, we urge that this examination be placed
in proper perspective by taking cognizance of the fact that the whole
history of pension plans has been one of continued improvement over
the years-in coverage, in vesting, in funding, a.nd in the level of bene-
fits. The President's Committee on Corporate Pension Funds, though
recommending far-reaching changes in pension plans, explicity rec-
ognized the value of these plans by stating". . . public policy should
continue to provide appropriate incentives to private plan growth...
THE GROWTH OF THE PRIVATE PENSION SYSTEM
The vitality of the private pension system is shown by its rapid
growth. By the end of 1965. there were over 150,000 private pension
plans qualified under the Internal Revenue Code. These plans covered
more than 25 million employees-about one-half the total private non-
agricultural work force. Over 66,000 of these plans, covering 6.3 million
employees, were insured with life insurance companies.1
In 1965, total employee and employer contributions to private pen-
sion plans reached $7.8 billion, and 2.7 million beneficiaries received
$3 billion in pension payments. At the end of 1966, the assets of pu-
1 Institute of Life Insurance, PrIvate and Public Pension Plans in. the United States~
March 1967.
PAGENO="0181"
OLD AGE INCOME ASSURANCE-PART V 173
vate pension plans totaled $93.4 billion-more than four times the re-
serves of the social security system. The reserves of insured pension
plans alone amounted to $29.4 billion.
By 1965 the number of employees covered by pension plans was
2˝ times as large as in 1950, the number of beneficiaries six times the
195~0 total, payments to beneficiaries eight times the 1950 total, and
reserves seven times the 1950 total.
According to the President's Committee on Corporate Pension
Funds, by 1980 an estimated 43 million employees will be covered by
private pension plans, the number of beneficiaries of such plans will
increase to 6.6 million, and annual benefit payments to about $9 bil-
lion. By 1980 total contributions to private pension plans are expected
to rise to $10.9 billion a year, and total reserves to $225 billion.2
THE PRIVATE PENSION SYSTEM BENEFITS THE RANK AND FILE OF EMPLOYEES
It is worthy of note that the rapid increase in private pension plans
over the past few decades has consisted overwhelmingly of plans
which meet Internal Revenue requirements designed to insure that
the plans will benefit the rank and file employees and not merely a
few highly paid employees. Since 1942, the Internal Revenue Code has
contained provisions which prohibit qualified pension plans from
discriminating as to coverage or benefits in favor of highly paid em-
ployees. The Internal Revenue Code seeks to induce compliance with
these nondiscriminatory requirements by a carrot-and-stick ap-
proach-by giving better tax treatment where plans comply with these
provisions and by withdrawing this treatment where plans do not
comply.
More specifically, in order to qualify under the Internal Revenue
*Code, a pension plan must cover a specified percentage of employees3
or, as an alternative, cover employees under a classification found by
the Commissioner of Internal Revenue not to discriminate in favor
of employees who are officers, shareholders, supervisory employees, or
highly compensated employees. Similarly, the contributions to the plan
or benefits paid out by the plan cannot constitute a larger percentage
of pay for upper income employees than for lower income employees.
THE BULK OF PENSION PLANS ARE FUNDED
It is also noteworthy that the bulk of the present pension plans are in
varymg degrees funded; that is, contributions made by the employer
and the employees are set aside either in a trust or with an insurance
company to pay benefits at a later date. The Internal Revenue Code
recognizes the desirability of funding in its provisions dealing with
qualification. The Code does not specify minimum funding standards.
However, as a matter of administrative practice, the Internal Revenue
Public Policy and Private Pension Programs, A. Report to the President on Private
Employee Retirement Plans by President's Committee on Corporate Pension Funds and
other Private Retirement and Weif ace Programs, app. A, table 1.
To qualify on this basis, `the plan must cover 70 percent or more of all the employees,
or SO percent or more of all the employees who are eligible to benefit under the plan if 70
percent or more of all the employees are so eligible, excluding In each case employees who
have been employed not more than a minimum period prescribed by the plan, not exceeding
5 years, employees whose customary employment is for not more than 20 hours In any
i week, and employees whose customary employment is for not more than S months in
any calendar year (sec. 401 (a) (3) (A) of the Internal Revenue Code).
PAGENO="0182"
174 OLD AGE INCOME ASSURANCE-PART V
Service has approved plans under which the employer's total pension
contributions over the years, on the basis of assumptions adopted by
film, cover the current costs of the plan plus the interest on unfunded
accrued liabilities.
The life insurance business favors adequate funding of private pen-
sion plans. Funding provides the employer with an orderly method of
making provision for his pension costs. It enables him to plan ahead
to meet these costs, and, by spreading out the contributions, to avoid
the severe financial burden of paying the bulk of the costs within a
relatively short period of time to a maturing labor force. Further, and
`very important, a funded plan provides greater assurance to a covered
employee that the money will be available to pay him a. pension when
he retires.
The fa.ct that private pension plans are funded makes it possible for
them to pay substantially larger pensions than would otherwise be
possible. This is so because the investment income accruing on the
funds is devoted to the payment of benefits. The increased benefits re-
sulting from such investment income represents the counterpart of the
increase in the productivity of the economy, which pension plans make
possible by enlarging the supply of savings and capital.
Moreover, although the funds in pension plans represent. savings set
aside for the retirement of covered employees, they do not inhibit indi-
vidual savings for retirement and other purposes. On the. contra.ry,
there is strong evidence that coverage under pension plans encourages
individuals to increase their personal savings.4
In the past few years, proposals have been made by a number of
groups, including t.he President's Committee on Corporate Pension
Funds, to require qualified plans to fund current. and past service
costs over a specified period of years. Proposals of this type warrant
study and discussion by iiiterested part~es. It must be kept. in mmcl,
however, that accelerated funding increases costs, and a program
should not be adopted which would disc.ourage the growth and devel-
opment of prjvate pension plans.
Proposals t.o guarantee the pension rights of' employees have also
been suggested. Adequate funding would be a necessary concomitant to
any such program, however, and to the extent that funding is im-
proved the need for a guarantee program te.nds to disappear. More-
over, it does not seem to us feasible to provide an equitable method of
guaranteeing the widely varying risks characteristic of different kinds
of pension plans-including the contingency that the employer may
go out of business and terminate the plan before the funding is com-
pleted, or that the administrator might make unwise investments with
a resu.lting loss of pension funds.
VESTING
The history of vesting provisions in pension plans has been one of
continuing improvement. Many plans grant liberal vesting rights at
the time they are adopted. Others provide only moderate or no vesting
See Private Pensions and Individual Saving by George Katona. Survey Research Center.
Institute for Social Research. the Tjniversity of Michigan, lOGo. and Tue Effect of Pension
Plans on Aggregate Savings by PhilIip Cagan, Nahonal Pur~aU of Economic Roieureh, 196o.
PAGENO="0183"
OLD AGE INCOME ASSURANCE-PART V 175
at the time they start but then continually improve the vesting pro-
visions as they go along. By the winter of 1962-63, according to a
Bureau of Labor Statistics study of 16,000 pension plans covering over
15.6 million employees, two out of three private pension plans covering
three out of five workers provided vesting.5 Present vesting is undoubt-
edly even more widespread in view of the marked trend toward more
liberal vesting.
The life insurance business is in favor of vesting. Vesting helps to
assure that covered employees will actually receive pensions and will
not lose their benefits through termination of employment. Moreover,
although the impact of vesting on the mobility of labor requires
further study, there appears to be a presumptive relationship between
vesting and labor mobility.
Recently, there have been suggestions that some degree of vesting
should be required as a condition to qualification of plans under the
Internal Revenue Code. Again, we believe that such suggestions should
be studied and discussed. It must be recognized, however, that vesting
affects costs and, hence, must be considered in relation to other pos-
sible improvements in pension plans, and to the ability of new or small
employers to initiate such plans.
PENSION PLANS ARE FINANCED OUT OF PRIVATE FUNDS
An important advantage of private pension plans is that they are
financed out of private funds, not public funds. Since they are paid
for by voluntary contributions made by employers and employees,
they do not involve compulsory tax payments or the direct expendi-
ture of public funds.
It is sometimes alleged that, although private pension plans are
financed by private funds, they are "subsidized" by the Government
in that present law grants them favored tax treatment. This view is
based on three points: (1) The employer gets a tax deduction within
certain limits for contributions to a qualified plan at the time the
contribution is made; (2) covered employees are not taxed on em-
ployer contributions made on their behalf until they actually receive
pension benefits; and (3) investment earnings on funds set aside in
qualified trusteed and insured plans are free of tax until received by
the employees.
It is questionable, however, whether the three items just cited can
be said to amount to a "subsidy."
Allowing an employer to. take a tax deduction for his contribution
to a qualified pension plan does not constitute special tax treatment.
In order to secure the deduction under present law, the employer
must irrevocably part with the contributed funds when he hands
them over to the pension trust or the life insurance company for pen-
sion purposes. The deduction received for such contributions, there-
fore, merely recognizes that the employer has incurred a business
expense similar to compensation paid to employees in. other forms-
for example, cash wages.
Labor Mobility and Private Pension Plans-A study of Vesting, Early Retirement and
Portability Provisions, U.S. Department of Labor, Bureau of Labor Statistics, B.L.S.
Bufl~tin, No. 1401, p. 11.
PAGENO="0184"
176 OLD AGE INCOME ASSURANCE-PART V
Moreover, it is questionable that an employee covered by a pension
plan actually receives special tax treatment because he is not taxed
when his employer makes the contributions, but instead pays tax when
he receives the benefits. A good case can be made in equity for impos-,
ing the tax at the later date. This is the time when the employee
actually enjoys the use of the pension income. It would be unfair to
tax the employee at the time his employer makes the pension con-
tributions, since even an employee who has vested rights generally
cannot receive any benefits prior to retirement. Furthermore, he might
die prematurely and never receive any benefits.
For the same reasons, it is equitable that tax on the investment
earnings on l)eflsiofl funds be paid only when they are received by the
employee as pension benefits.
Finally, there would appear to be considerable justification for en-
couraging the private pension system even at the cost of some tax
revenue. The social objective to be gained-the encouragement of
private provision for retirement needs without the direct expenditure
of public funds-would be well worth the cost.
II. PRIVATE PENSION FUND SAVING AND INVESTMENT AND THEIR
IMPLICATIONS FOR EcoNo~IIc GROWTH AND STABILITY
Probably the most provocative idea. advanced in the Joint Com-
mittee Print is the suggestion that the flow of saving into investment
through private pension funds may be contrary to the public interest
in that it may inhibit full employment and economic growth and con-
tribute to economic instability. This suggestion conflicts with the views
of the vast majority of economists, who are convinced that the healthy
expansion of saving through private pension funds is vital to the
sound growth of the American economy in the years ahead.
This section of our statement will discuss: (1) Some of the specific
issues raised by the print in this regard; (2) the theory of saving,
investment, and economic growth; (3) the facts about private pension
fund saving and investment; (4) the urgent need for a high rate of
saving and investment in the years ahead; (5) pension savings and
longTun investment, planning; (6) pension savings, investment, and
risk capital; and (7) recommendations in other countries for the en-
couragement of pension savings.
THE IS5UE5 RAISED BY THE JOINT COMMITTEE PRINT
An introductory paragraph of the print asks: "May not pension
plans . . . interfere unnecessarily with the exercise of free choice
in employment and in saving, induce an excessive rate of saving * * *
and hinder the productive deployment of wealth." 6Later, in discussing
the "economics of pension saving," the print states:
The large volume of pension plan saving is to some no source of great con-
cern. it being argued that since saving increases the capital stock we are all
better off therefore. The difficulty with this argument is that the premise is
false. Saving does not increase the capital stock. Investment increases the cap-
ital stock. Given the demand for capital goods, increased saving reduces the
rate of interest. A fall in the rate of interest may induce additional investment;
6 Joint Committee Print. p. 1.
PAGENO="0185"
OLD AGE INCOME ASSURANCE-PART V 177
again it may not. The factors which prompt businessmen to invest are diverse
or at any rate poorly understood, but there is little evidence that the avail-
ability of funds or the rate of interest are of much significance except on occa-
sion in exerting restraint. Were the economy operating at rates of growth which
tended chronically to be excessive, there would be a case for devising ways to
induce additional saving and a highly fragmented pension system might be as
good a way as any. However, the economy is not subject to chronic excess de-
mand; rather, there is a more or less continuing problem of keeping demand
adequate. In such circumstances, a high rate of saving does not add to but sub-
tracts from the capital stock by depressing demand for output and the motive
to invest.7
The ideas thus expressed resurrect the "stagnation thesis," which
enjoyed popularity in the depressed 1930's. The assumptions of this
thesis are wholly irrelevant to the economic conditions of the past
two decades in which: (1) The U.S. economy has generally operated
at high levels of employment; (2) the general price level has dis-
played a persistent tendency to move upward; (3) cyclical down-
turns have been moderate; and (4) the level of long-term interest
rates has trended upward (despite cyclical fluctuations) carrying the
average yield on long-term Government bonds, for example, from
2.19 percent in 1946 to 4.65 percent in 1966.
The Print's assertion that "there is little evidence that the avail-
ability of funds or the rate of interest are of much significance (in
influencing investment spending) except on occasion in exerting re-
straint" is highly questionable. There is clear-cut evidence that in the
period 1961-65 the ready availability of financing by institutions ag-
gressively seeking to increase their loans and investments at declining
interest rates was a highly important contributor to economic expan-
sion. In fact, the Print's statement is contrary to all of the assump-
tions in the .January, 1967, Economic Report of the President which
counts heavily upon increased availability of funds and lower interest
rates to stimulate a rising rate of residential construction.
In a discussion of the economics of pension fund saving, the Print
goes on to state:
Now pension saving, in contrast to personal and corporate saving, is relatively
unresponsive to changes in economic fortune. Pension saving, governed by rules
for funding plans, increases rather steadily in good times and bad. It is, therefore,
a good thing in good times and a bad thing in bad times. Corporate saving,
especially, but also personal saving to a lesser degree are automatically stabiliz-
ing. To whatever extent corporate and personal saving are smaller because pen-
sion saving is greater, something in the tendency of the economy to stabilize
itself through compensating changes in savings rates is lost.8
This again reflects the stagnation thesis. Economists who are seek-
ing today to project the longer run future of the American economy
are assuming that appropriate fiscal and monetary policies by Gov-
ernment will be combined with dynamic factors in the private econ-
omy-such as rapid technological change and a pronounced rise in
the rate of family formations-to assure a strong rate of economic
growth.9 Built into their models is a high rate of capital formation
and a high rate of saving. If this is, indeed, the realistic prospect, the
contractual nature of private pension saving has great advantages.
Because of its contractual nature, pension saving is better able to resist
7Ibid., p. 18.
S Ibid., p. 18.
° See, for example, the studies cited and discussed at length, beginning on p. i70. infra.
PAGENO="0186"
178 OLD AGE INCOME ASSURANCE-PART V
the great competitive pressures for the consumer dollar which are
tending to whittle down the rate of personal saving. Because of the
steady flow of pension savings, longer run planning Qf capital spend-
ing by business and industry becomes more feasible.
Unless private savings are available for investment, the only pos-
sible alternatives are monetary expansion and Government savings.
The Joint Committee Print comes close to saying that the Govern-
ment should provide funds for investment, thus dispensing with much
of the need for private savings. This would seem to call for an uimeces-
sary encroachment of Government, to the detriment of the free enter-
prise system.
Another issue raised by the Joint Committee Print is whet-her an
alleged failure of private pension funds to assume investment risks
has curbed the economic growth of the country. The issue is stated
as follows:
Very generally speaking, reductions in pension plan funding requirements
probably confer upon society a boon in productive efficiency. Transferring these
funds to business in general through the intermediation of pension funds means
that they will be invested on the average more cautiously `than if the transfer
had not taken place. What good does it do, we may ask, to consider all the
alternatives if one consistently chooses safe, low-yield investments? The reduc-
tion over several years in the rate of economic growth from this diversion of
funds at risk could easily exceed the cost of all pension benefits paid.1?
Again, this is an erroneous concept. Safety of principal must, of
course, always be an important consideration in the bond and mort-
gage investments of pension funds. During the past 20 years, however,
through direct placement of corporate bonds and mortgages, the in-
stitutions administering pension funds have taken reasonable and
measured risks, and have been compensated for such risks by a better
rate of investment return. Also, noninsured pension funds have for
many years been invested to- a substantial degree in common stocks, and
the proportion of insured pension savings invested in common stocks is
rising ma-rkedly. It is not surprising, therefore, that pension fund
investments have contributed strongly to increased productivity a.nd
growth. -
After having argued that pension fund investments are too con-
servative, the Print then immediately proceeds to argue just the
reverse, as follows: - - -.
We have stressed the likelihood `that pension funds will be managed with an
excess of caution. -`The rate' at which pension funds have acquired- corporate,
shares during recent years does not suggest caution. Rather it suggests that
pension trustees from an interest in reduced contributions and perhaps from lack
of experience in financial management may be putting their plans in a position to
suffer a disaster such `as that which overtook mutual funds in the early 1930's-
for much the same reason. Pension plan purchases of corporate shares now, just
as mutual fund purchases then, are a primary factor accounting for a chronic
excess demand for shares which explains a long, rapid rise in share, prices * *
It is a misfortune that banks and insurance companies, competing for pension
funds. have got caught up in this process. Through the exercise of monetary
and fiscal policy the economy can be protected from widespread unemployment
but not against stock market debacles. There is, therefore, some prospect that
much of the value of pension funds may sometime disappear. What will the
plans do then-asks the Federal Government-to make good their pension
promi~es?'1
10 Joint Committee Print, pp. 19-20.
11 Ibid., p. 20.
PAGENO="0187"
OLD AGE INCOME ASSURANCE-PART V 179
As we have already noted, a high proportion of noninsured pension
savings has for some time been invested in common stocks, with a
limited but growing share of the insured funds going into corporate
shares. There are, of course, some risks in this policy. But, as the
American economy has expanded strongly during the past two decades,
with the general price level moving upward, and with the cyclical
downturns moderate and of brief duration, the policy has proved suc-
cessful. Looking to the future, stock investment by pension funds
has a role to play, particularly if fiscal and monetary policies are to be
directed toward maintaining full employment and stimulating faster
economic growth.
The comments above highlight some of the basic defects in the eco-
nomic thinking reflected by the Print. We turn now to a fuller devel-
opment of our views with respect to the great importance in the
years ahead of a vigorous growth in private pension savings.
THE THEORY OF SAVING, INVESTMENT AND ECONOMIC GROWTH
The objectives of TJ.S. economic policy, as set forth in the Employ-
ment Act of 1946, and in subsequent annual economic reports of Presi-
dents of both parties, are: (1) The maintenance of full employment
of our labor force and other resources; (2) a stronger rate of eco-
nomic growth (currently the target seems to be 4-5 percent compared
with an historical trend figure of about 3 percent) ; (3) the avoidance
of serious fluctuations in the economy; (4) reasonable stability of the
general price level; and (5) restoration and maintenance of reasonable
balance in the international payments position of the United States.
All of these objectives are to be sought by means designed to. preserve,
at a maximum degree, a free market-oriented economy.
At the present time, the U.S. economy is operating at very close to
its full employment potential. The only way to grow faster than the
present attainable growth rate is to increase the potential for growth.
Walter Heller recognizes this in the following statement:
So future fiscal dividends must be declared with an eye not just to the uses
but to the sources of growth. In the past few years part of our rise to the top
of the growth ladder has been accomplished by closing the GNP, or production
gap. With the economy operating at or near its potential, our realized growth
in the future will depend chiefly on the rate of increase in that potential. We can
no longer pad the figure, so to speak, by taking up economic slack.
Rising productivity will be the key. In part this will require continued meas-
ures to maintain high levels of private investment in plant and equipment. In
part, also, it calls for measures to improve efficiency and hence productivity by
adjustments-many of them politically painful-in our policies for transportation,
manpower allocation, agriculture, and the like.12
Charles L. Schultze, Director of the Bureau of the Budget, in his
book National Income Analysis, cites two main sources of economic
growth-labor input and the growth of productivity.13 The labor in-
put of any economy depends upon (1) the number of people of work-
ing age, (2) the "participation rate"-the percentage of people of
working age who choose to enlist in the work force, and (3) the num-
~ Walter W. Heller, New Dimensions of Political Economy, Harvard University Press,
Cambridge, Mass., 1966, p. 108.
13 Charles L. Schultze, National Income Analysis, Prentice-Hall, Inc., 1964, pp. 113 if.
PAGENO="0188"
180 OLD AGE INCOME ASSURANCE-PART V
ber of hours worked per year. As Dr. Schultze points out, "in the long
run, rising living standards flow from an increase in output per
worker." Since ~Torld War II output per man-hour has increased at
an annual rate of about 3.2 percent.. The ways to increase output
ler man-hour, as set forth by iDr. Sehultze, are: (1) Increases in the
stock of physical capital; (2) improvement in technology-advances
in scientific and teclmica.l knowledge; (3) the spread of education;
and (4) an expansion of markets which foster productive specializa-
tion.
What is the role of saving and investment in rising prothictivity
and economic growth? Dr. Sc.hultze answers as follows:
Rising productivity involves an increase in our ability to harness the forces
of nature, for supplying power, transporting goods, and manipulating materials.
This, in turn, requires not merely scientific and technical knowledge, but the
provision of expensive plant and equipment. It requires, in other words, capital.
And capital goods are not "free". They must be produced using labor, materials,
and machines that might otherwise have been utilized to turn out consumer
goods. Providing the labor force with an increasing supply of machinery and
equipment involves saving-the Nation must refrain from consuming part of
its output so that resources may be free to produce capital goods.14
Dr. Schultze then goes on to point out. that growth in procinctivity
depends not only upon workers having more capital goods at their
disposal but better capital goods. The latter depends upon technologi-
cal progress. He notes:
Thus, although the accumulation of capital and technological progress can be
thought of as two distinct sources of economic growth. they are intimately
related. Without the other, each could make only a limited contribution to
growth. Much of technological progress is embodied in new capital. Much of
the contribution of new investment to economic growth is due not to the addi-
tional capital per se, but to the new technology that it carries with it and that
could not be introduced without it.15
The above quotations arc in the classical tradition in stressing the
great importance of saving and investment in economic growth. TJncler
conditions of full employment such as exist today, and as will continue
if our national economic objectives are to be achieved, increasing the
money supply, beyond the monetary expansion needed simply to carry
out transactions on a. rising national income basis, would go largely
to produce an increase of the general price level. In a slack economy,
such as in 1961-64 when the country had idle labor and plant capacity,
capital formation could be financed out. of an increase in the money
supply without precipitating an increase in the general price level.
But, under conditions of full emp1o~ment, the only sound way to fi
nance capital formation is by means of saving. An even more rapid
rate of real economic growth than we have hitherto enjoyed, which
would permit the country to raise its objectives, would require an in-
crease in the "full employment savings" which the economy generates.
THE FACTS ABOUT PRIVATE PENSION YtXD SAVING AND INVESTMENT
1\That are the facts about. private pelision fund saving and invest.-
ment? Using the national income accounts prepared by the U.S. De-
partment of Commerce, the following picture of the importance of
pension fund saving emerges:
`~ Sehultze, p. 118.
~ Ibid., p. 122.
PAGENO="0189"
OLD AGE INCOME ASSURANCE-PART V 181
[Dollars in billions[
Average p
er annum
1946-50
1951-55
1956-60
1961-65
Personal income'
$203.0
$283.4
$366.0
$471.2
Personal saving'
Pension saving 2
Personal saving as percent of personal income
Pension saving as percent of personal income
$11.7
$2. 7
5. 8
1. 3
$17.2
$4. 6
6. I
1. 6
$20.0
$7. 2
5. 5
2. 0
$22.6
$9. 9
4. 8
2. 1
Note: If personal saving is expressed as a percentage of disposablepersonal income (after Federal income tax payments),
the savings ratios for the four 5-year periods are 6.4, 6.9, 6.2, and 5.5 percent, respectively. The use of the ratio of saving
to personal income is considered more relevant here since most of the pension saving is, in effect, from pretax income;
i.e.,employer contributions to private and public pension plans are not treated as taxable income to the employee.
i National income accounts of the U.S. Department of Commerce Survey of Current Business.
2 Flow of funds accounts of the Board of Governors of the Federal Reserve System. This figure measures the increase in
assets held by private pension plans, pension programs administered by State and local government units, and the Federal
employee and railroad retirement benefit programs. It includes both insured and noninsured plans. It excludes OASDI.
As shown in the table, the annual amount of saving through private
pension funds and those administered by State and local government
units has increased markedly since 1946, both in absolute dollar amount
and as a percentage of total personal income. It is significant that total
personal saving as a percent of personal income has declined during
the past 15 years from an average of 6.1 percent in 1951-55 to an
average of 4.8 percent in 1961-65. Thus the increase in the rate of
pension saving has been an important factor in holding up the ag-
gregate rate of saving.
The savings accumulated through private pension funds and through
pension funds administered by State and local government units have
been invested in corporate securities, mortgages, State and municipal
bonds, and U.S. Government obligations. They have thus provided
financing for the construction of industrial plant and equipment,
single-family homes, apartment buildings, commercial properties of
all kinds, public utilities, transportation and communication facilities,
roads and other public facilities, and many other kinds of capital
projects. The capital expenditures made possible by pension savings
have provided increasing job opportunities in our economy and have
contributed heavily to improved productivity and thus higher living
standards.
Studies conducted by the National Bureau of Economic Research
indicate that in coming years the rate of increase of private pension
saving is likely to slacken because, with a rising percentage of our
people moving into the retirement age group, benefit payments are
likely to be increasing at a more rapid rate than the sum of contribu-
tions and investment earnings.bG This is a disturbing prospect in view
of the need to finance the heavy demands for capital formation which
the country faces in the years ahead. This is an additional reason to
encourage expansion of the private pension plan system.
THE URGENT NEED FOR A HIGH RATE OF SAVING IN THE YEARS AHEAD
In looking to the future, the majority of economists are projecting a
stronger rate of economic growth-say 4 to 5 percent-than we have
iO Daniel M. Holland, Private Pension Funds: Projected Growth, Occasional Paper 97
National Bureau of Economic Research, New York, N.Y., 1966.
PAGENO="0190"
182 OLD AGE INCOME ASSURANCE-PART V
experienced in the past. This expectation is based on the following
reasoning:
1. There are a number of factors in our economy which will
provide a favorable climate for strong growth such as (a) a sharp
increase in the labor force; (b) a marked increase in the rate
of family formations; (c) a very rapid rate of technological in-
novation abetted by a large increase in research and development
expenditures; (d) the need for heavy and rising public and pri-
vate expenditures to meet tile problems of our cities-urban re-
newal and rehabilitation, air and water pollution, transportation,
and the like; and (e) tile drive for stronger economic growth in
other countries and the leverage whjch this will exert upon our
own desire and capacity to grow.
2. Both political parties and tile public at large agree that fiscal,
monetary, and other G-overmnent policies must be directed toward
encouraging full employment and faster economic growth.
To achieve a stronger rate of economic growth there will have to be
a very high rate of investment spending, and, of course, a very high
rate of saving.
To illustrate the above points, it. will be helpful to review three re-
cent projections of the growth of the U.S. economy in the next decade:
(1) U.S. Economic Growth to 1975: Potentials and P'roblem-s: a study
prepared for the Subcommittee on Economic Progress of the Joint
Economic Committee of the Congress; (2) Albert T. Soimners, "The
Economy in tile Next Decade," The Conference Board Record, Decem-
ber 1965; and (3) Leonard A. Lecht, Goals, Priorities, and Dollars-
The Next Decade; a study prepared under the auspices of the National
Planning Association. These studies are typical of the thinking of most
economists about the prospects for growth of the American economy
in the years ahead.
Tables I and 2 present the projections of C-NP and national income
developed in the study made by the staff of the Joint Economic Com-
mittee. These projections have built into them a number of assump-
tions about Federal fiscal policy aimed at maintaining full employ-
ment.' "Projection A" in the tables is based Oil the assumptldn that.
during tile next decade the real G-NP will grow at a 4.5-percent annual
rate, and also assmnes an average unemployment rate of 3 percemlt.
"Projection B" is based on the assumption that- during tile next decade
the real C-NP will grow at a 4-percent annual rate, and assumes an
average unemployment rate of 4 percent.
It is significant to note, ill table 1. that projection A (ill current
dollars) calls for a rise in gross private domestic. investment, from
~106.6 billion in 1965 to $142.9 billion in 19T0 and to $201.4 billion m
1975, an increase of nearly 90 percent by 19Th. Excluding the change
in business inventories, the increase ill fixed investment would be from
$97.5 billion in 1965 to $185.7 billion in 1975, or a.n increase of just about
90 percent. These figures certainly do not suggest that. there will be any
lack - of demand for capital funds generatedl by savings during the
next dlecadle. Tile figures ill projection B, aithough somewhat lower,
also suggest an enormous rise in captial demands.
`~ For a discussion of these assumptions, see U.S. Economic Growth to 1975: Potentials
ani Problems, pp. 29-47.
PAGENO="0191"
OLD AGE INCOME ASSURANCE-PART V
Billions of 1958 dollars
183
TABLE 1--PROJECTIONS OF GROSS NATIONAL PRODUCT, WITH MAJOR COMPONENTS, ADJUSTED TO ILLUSTRATE
AN EQUILIBRIUM FULL-EMPLOYMENT POSITION I
Projection
Actual, 1965
1970
A
Projection B
1975
1970 1975
Gross national product 681. 2 950. 0
431.5 601.7
Billions of current dollars
66. 1
190.6
174.8
106.6
69.7
27.8
9. 1
7. 0
39.0
32.0
1,310.0
815.8
125.6
315.7
374. 5
201.4
113. 0
72. 7
15. 7
10.8
69.2
58. 4
92. 7
247.9
261. 1
142.9
87. 4
46. 0
9. 5
9.2
52. 1
42.9
920.0 1,205.0
583.4 753.9
87.5 113.5
241.5 292.9
254. 4 347. 5
137.5 183.5
85. 0 105. 5
44.3 66.0
8.2 12.0
9.3 12.1
51.0 66.2
41.7 54.1
136.2 197.8 282.0 189.8 255.5
66. 8 80. 9 99. 3 78. 5 93. 1
69.4 116.9 182.7 111.3 162.4
Personal consumption expenditures
Durable goods
Nondurable goods
Services
Gross private domestic investment
Nonresidential fixed investmeat
Residential structures
Change in business inventories
Net exports of goods and services
Exports
Imports
Government purchases of goods and
services
Federal
State and local
Gross national product 614. 4
Personal consumption exependitures_ - - - 396. 2
Durable goods 66.4
Nondurable goods 178. 2
Services 151.6
Gross private domestic investment 97. 8
Nonresidential fixed investment 64. 9
Residential structures 24. 1
Change in business inventories 8. 8
Net exports of goods and services 6. 3
Exports 37.3
Imports 31.0
Government purchases of goods and
services 114. 1
Federal 57.8
State and local 56.3
Ratio of nonresidential fixed investment
to real GNP in 1958 dollars (percent) -- 10.6
770.0 960.0 760.0 925.0
505.0 631.1 498.3 606.0
92.5 124.6 88.0 114.5
213.5 253.0 211.5 243.0
199. 0 253. 5 198. 8 248. 5
117.7 150.7 115.5 143.6
- 75.5 91.2 74.5 87.9
34. 5 48. 0 34. 2 46. 5
7.7 11.5 6.8 9.2
7.9 8.5 8.0 9.9
49.6 64.7 48.5 61.9
41.7 56. 2 40. 5 52. 0
139. 5 169. 7 138. 2 165. 5
60. 0 64. 3 60. 0 64. 5
79.5 105.4 78.2 101.0 -
9.8 9.5 9.8 9.5
Source: Department of Commerce, Office of Business Economics, and staff, Joint Economic Committee.
I Projections in tables 19-28 are illustrative only and portray 1 possible method of achieving equilibrium at full
employment by the use of Federal fiscal policy.
Shifting to table 2, it is significant to note that both projections call
for amarkedincrease in the aggregate of personal saving.'8 Recogni-
10 The authors of the Joint Economic Committee study explained the personal savings
rates In table 2 asfohlows. The average personal saving rate for the years 1901-65 was 5.5
percent, the rate used for the B projections. This seemed reasonable in view of the fact
that the B set assumes a 4 percent unemployment rate. The lower savings rate used for
the A set of projections was chosen because more of the employed labor force would be
in the low-income group at a 3 percent rate of unemployment than would be the case were
the unemployment rate to average 4 percent. It is a characteristic of low-income families
to save less as a proportion of their income than those in higher income classes (see p. 9
of U.S. Economic Growth to 1975: Potentials and Problems).
PAGENO="0192"
184 OLD AGE INCOME ASSURANCE-PART V
tion by the Joint Economic Cormnittee staff of the importance of
personal saving to the growth projections can be found in the following
quotation:
Expansions in social security, private pensions, disability and unemployment
insurance, hospital insurance, and medicare may all act to maintain, if not reduce,
the consumer's propensity to save on his own account.
It should be recognized that there will likely be strong forces in the remainder
of the 1960's which could tend to shift the saving rate to a lower level. The high
birth rates of World War II and early postwar years will be reflected in higher
rates of new family formation. These individuals will be in a stage of life when
automobiles, household furnishings, and other durable goods typically are
acquired for the first time.
Hence, public attention may at some point in the decade need to be focused on
the desirability of encouraging private saving to promote private-oriented invest-
ment, risk bearing, and entrepreneurship.19
TABLE 2-GROSS NATIONAL PRODUCT AND NATIONAL INCOME, WITH MAJOR COMPONENTS ADJUSTED TO
ILLUSTRATE AN EQUILIBRIUM FULL EMPLOYMENT POSITION
[In billions of dollars)
Actual, Projection A Projection B
1965
1970 1975 1970 1975
Gross national product
Less-
Capital consumption allowances
Indirect business tax and nontax liability
Business transfer payments
Statistical discrepancy
Plus subsidies less current surplus of Government enterprises. -
681.2
59. 6
62.7
2. 6
-1.6
1. 0
950.0
78. 2
83.7
3. 3
. 8
1,310.0
101. 6
115.8
4. 1
. 4
920.0
76. 6
81.4
3. 3
. 8
1,205.0
96. 2
108.2
4. 1
. 4
Equals national income
Less-
Corporate profits and IVA
Contributions for social insurance
Plus-
Government transfer payments to persons
Interest paid by Government (net) and consumers
Dividends
Business transfer payments
559. 0
74.2
29. 2
37. 1
20. 6
19. 2
2. 6
785. 6
102.7
49. 5
62. 1
29. 0
24. 5
3. 3
1, 088. 9
139. 4
69. 3
89. 3
38. 5
30. 7
4. 1
759. 5
97. 0
48. 4
60. 9
28. 5
24. 1
3. 3
996. 9
118. 2
66. 0
83. 3
36. 5
29. 8
4. 1
Equals personal income
Less-
Personal tax and nontax payments
Tax reductions
Adjusted personal tax and nontax payments
535. 1
66. 0
752. 3
105. 6
4. 3
101. 3
1, 042. 8
171. 8
13. 7
158. 1
730. 9
100. 9
3. 6
97. 3
966. 4
152. 8
10. 6
142. 2
Equals disposable personal income
Less-
Personal outlays
Personal consumption expenditures
Other personal outlays
469. 1
443.4
431. 5
11. 9
651. 0
619.6
601. 7
17.9
884. 7
841.8
815. 8
26. 0
633. 6
600.8
583. 4
17.4
824. 2
777.9
753. 9
24. 0
Equals personal saving
Personal saving rate (percent)
National income
Compensation of employees
25. 7
5. 5
559.0
392. 9
31. 4
4. 8
785.6
559. 0
42. 9
4. 8
1,088.9
792. 0
32. 8
5. 2
759.5
546. 3
46. 3
5. 6
996.9
734. 4
Wages and salaries
358. 4
504. 5
712. 7
492. 9
661. 9
Private_
General government
Government enterprises
289. 1
62. 3
6. 9
399. 2
95. 8
9. 5
560. 1
139. 5
13. 1
392. 2
91. 5
9. 2
520. 4
129. 1
12. 4
Supplements to wages and salaries
34. 5
54. 5
79. 3
53. 4
72. 5
Private
General government
Government enterprises
28.5
5. 5
. 6
45.8
8. 8
. 7
66.7
11. 7
. 9
44.9
7. 8
.7
60.8
10. 8
.9
in U.s. Economic arowtlv to 1975: Potentials and Problems, p. 46.
PAGENO="0193"
OLD AGE INCOME ASSURANCE-PART V 185
TABLE 2.-GROSS NATIONAL PRODUCT AND NATIONAL INCOME, WITH MAJOR COMPONENTS ADJUSTED TO
ILLUSTRATE AN EQUILI BRIUM FULL EMPLOYMENT POSITION-Continued
[In billions of dollarsj
Actual, Projection A Projection B
1965
1970 1975 1970 1975
Proprietors' income
55. 7
71. 0
84. 0
65. 0
75. 0
Business and professional
Farm
40. 7
15.1
55. 0
16.0
68. 0
16.0
50. 0
15.0
60. 0
15.0
Rontal income of persons
Corporate profits and IVA
18. 3
74.2
20. 9
102.7
23. 5
139.4
20. 2
97.0
22. 3
118.2
Profits before tax
75.7
105.7
142.4
99.5
120. 7
Profits tax liability
Profits after tax
31. 2
44. 5
44. 1
61. 6
59. 9
82. 5
41. 5
58. 0
50. 8
69. 9
Dividends
Undistributed profits
19.2
25.3
24.5
37. 1
30.7
51.8
24.1
33.9
29.8
40. 1
Inventory valuation adjustment
-1. 5
-3.0
-3.0
-2. 5
-2.5
Net interest
17.8
32.0
50.0
31.0
47. 0
Source: Department of Commerce, Office of Business Economics, and staff, Joint Economic Committee.
The projection of GNP in 1975 prepared by Albert T. Sommers
of the National Industrial Conference Board is shown in tables 3 and
4. The Sommers model assumes an average unemployment rate of 4.5
and a 4.35 percent increase per annum in real output. Again, it is inter-
esting to observe in table 3 that gross private domestic investment in
current dollars would rise sharply from $92.9 billion in 1964 to $171.9
billion in 1975, an increase of about 85 percent. Table 4 likewise shows
the marked increase in personal saving projected by Sommers, from
$26.3 billion in 1964 to $48.5 billion in 1975, with the rate rising from
6 to 6.2 percent of disposable personal income.
Commenting upon his projection, Sommers stated:
Up ahead, nevertheless, appears to lie a further vigorous expansion in the rate
of investment. Indeed, vigorous growth in investment is an essential requirement
if the substantially faster rate of growth of employment envisioned here is to be
accompanied by an appropriate increase in the total stock of capital.
The annual rate of new fixed investment in 1975 may approach $118 billion, in
prices then prevailing; this is almost double the gross investment rate of 1964.00
In his concluding remarks about his projection of the growth of the
American economy in the next decade, Sommers states:
In the first instance, the challenge posed by the trillion-dollar potential of the
U.S. economy is job creation. The pumper dimensions of the 1975 potential arise
largely because the American economy of 1965 is on the threshold of a great wave
of growth in its human resources, unequalled since the days of heavy immigration
to America many decades ago. The great challenge in the next decade is to create
the jobs to fully engage this surging tide of manpower.
The creation of jobs means, of course, investment; and investment, of course,
depends upon saving. Given the increments to the labor force that seem to lie
ahead, it would be hard to envision any 10-year projection of the American econ-
omy that would not focus on a compelling need to achieve and sustain a high rate
of capital formation, to equip the increments to the labor force, and to improve
the efficiency of the labor force as a whole.°1
20 Albert T. 5ommers, "The Economy In the Next Decade," The Confereace Board Record,
December 1965, p. 8.
21 Ibid., p. 23.
83-200---67-pt. 5-13
PAGENO="0194"
186 OLD AGE INCOME ASSURANCE-PART V
TABLE 3.-GROSS NATIONAL PRODUCT, 1964 ACTUAL AND PROJECTIONS TO 1975'
Item
GNP (b
1964
In
1964
dollars
illions of dollars)
1975
~-
In In
1964 1975
dollars dollars
Annualgrowth
rate (in percent)
-
Based Based
on on
constant current
dollars dollars
Annual
price
increase,
1964-75
Perce
---
1964
In
1964
dollars
nt distribution
1975
-~
In In
1964 1975
dollars dollars
Total GN P
628. 7
1, 001.6 1, 138. 9
4. 35 5. 55
1. 2
100. 00
100. 00 100. 00
Personal consumption
expenditures
398. 9
644. 3 713. 9
4. 5 5. 5
1. 0
63. 45
64. 33 62. 68
Durable goods
Nondurable goods
Services
58.7
177. 5
162. 6
112. 8 117. 0
253. 3 274. 7
278. 1 322. 0
6. 2 6. 5
3. 4 4. 1
5. 1 6. 4
. 3
. 7
1. 3
9. 34
28. 23
25. 86
11. 26 10. 27
25. 29 24. 12
27. 77 28. 27
Gross private domestic
investment
New construction
92.9
48. 7
151. 1 171.9
79.2 90. 1
4.6 5.8
4. 6 5. 8
1.2
1. 2
14.78
7. 75
15.09 15.09
7. 91 7. 91
Residential nonfarm.~ -
Other
27. 0
21. 7
39. 5 44. 9
39.9 45. 3
3. 5 4. 7
5. 7 6. 9
1. 2
1. 2
4. 29
3. 45
3. 94 3. 94
3. 98 3. 98
Producers durable
equipment
Change in business
inventories
39.4
4. 8
64.3 73.1
7. 6 8. 7
4.6 5.8
4. 4 5. 6
1.2
1. 2
6.27
. 76
6.42 6.42
. 76 . 76
Net export of goods and
services
Exports
Imports
8. 6
37.0
28.5
9. 4 10. 6
(2) (2)
(2) (2)
.7 1. 9
(2) (2)
(2) (2)
1. 2
(2)
(2)
1. 37
5.89
4.53
. 94 . 93
(2) (2)
(2) (2)
Government purchases of
goods and services
Federal
128. 4
65. 3
196. 7 242. 9
84. 8 104. 6
4. 0 6. 0
2. 4 4. 4
2. 0
2. 0
20. 42
10. 39
19. 64 21. 32
8. 47 9. 18
National defense
Other
49. 9
15. 4
56. 0 69. 0
28. 8 35. 6
1. 0 3. 0
5. 9 7. 9
2. 0
2. 0
7. 94
4. 45
5. 59 6. 06
2. 88 3. 1
State and local
63. 1
111. 9 138. 3
5. 4 7. 4
2. 0
10. 04
11. 17 12. 14
Source: 1964-U.S. Department of Commerce; 1975-The Conference Board.
1 This tablo is consistent with thu new national accounts data released in August 1965. The figures are tentative trans-
lations of findings drawn in part from the old national accounts, and are subject to significant revisions in a final report
now in preparation.
2 Not estimated separately.
A similar conclusion was reached by Leonard A. Lecht, a member of
the staff of the National Planning Association, in his study Goals,
PTiorWes, and Dollai's-T/ie ~Vee~t Decade. Lec.ht's projection calls for
a GNP of close to $1 trillion in 1975, expressed in 1962 dollars. He as-
sumes a 4-percent increase per annum in real GNP, based on full
employment and a 3-percent per annum increase in productivity of
labor. Lecht has studied what the "costs" (in terms of GNP growth)
would be to fulfill the goals which have been advanced by the achnin-
istration in such fields as urban development, health, education, hous-
ing, and the like. To satisfy these goals, he has estimated that by 1975
G-NP would have to rise to $1.5 trillion (in 1962 prices). Concerning
this output, he comments:
Creating sufficient output to realize the aspiration standards for the 16 goals
would entail a GNP growth rate approaching 5.5 percent a year between 1962
and 1975 (measured by GNP in constant prices from the full capacity level of
production in 1962). This compares with an increase in GNP averaging 3 percent
a year in the past generation and 3.5 percent between 1948 and 1962. Sustaining a
5.5 percent annual growth rate for over a decade would require technological
changes leading to massive increases in productivity, considerably beyond the
3-percent growth in GNP per manpower anticipated in the next 10 years.22
25 G~Kzls, Priorities, and Dollars-T lie ivext Decade, p. 43.
PAGENO="0195"
OLD AGE INCOME ASSuRANCE-PART V
TABLE 4.-DETAILS OF SELECTED INCOME ACCOUNTS'
[Dollar figures in billions of current dollars[
187
1959
1963
1964
1975
1. Capital consumption allowances
$41. 4
$52. 8
$55. 7
$97, 5
2. Corporate depreciation charges
23. 5
32. 0
34. 0
66. 3
3. Other allowances
17.9
20.8
21.7
31.2
4. Corporate profits before taxes 2
5. Corporate taxes
52. 1
23. 7
58. 6
26. 0
64. 8
27. 6
118. 5
6. Corporate profits after taxes 2
28. 5
32. 6
37. 2
47. 4
71. 1
7. Dividends
12.6
15.8
17.2
8. Retained earnings
15. 9
16.8
19.9
38.7
9. Corporate cash flow
10. Personal income
52. 0
383. 5
64. 5
464. 8
71. 2
495. 0
137. 4
894. 9
11. Personal taxes
46.2
60.9
59.2
12. Disposable personal income
337. 3
403. 8
435. 8
111.9
783. 0
13. Personal outlays
14. Personal saving
318. 2
19. 1
383. 4
20. 5
409. 5
26. 3
734. 5
48. 5
15. Personal saving rate (percent)
5. 7
5. 1
6. 0
6.2
Selected Percentage Relationships
16. Personal income as percent of-
17. GNP
79.3
78.9
78. 7
18. National income
95.9
96.6
96. 2
78.6
19. Corporate profits before taxes as percent of-
20. GNP
10.7
9.9
96.7
21. National income
12. 9
12. 1
12. 5
10.4
22. Dividends as percent of-
12.
23. Corporate profits after taxes
24. Corporate cash flow
44. 2
24. 2
48. 6
24. 5
46. 4
24. 2
45. 6
23; 6~
Sources: 1959, 1963, and 1964-U.S. Department of Commerce; 1975-The Conference Board.
iThis table is consistent with the new national accounts of data released in August 1965. The figures are tentative
translations of findings drawn ri part from the old national accounts, and are subject to significant revisions in a final
report now in preparation.
2 Excluding invartory valaation adjustment.
3 Corporate profits after taxes plus corporate depreciation charges.
Including inventory valuation adjustment.
Aside from the "aspiration" level of GNP (i.e., $1.5 trillion),
Lecht's projection of a $1 trillion GNP provides that private expendi-
tures for business and industrial plant and equipment would have to
rise from $48.9 billion in 1962 to $102.3 billion in 1975 (all expressed
in 1962 dollars). This more than doubling of capital expenditures by'
business and industry is a measure of the great need for savings in the
years ahead if capital formation is to be financed soundly.
The most comprehensive study of capital requirements in the United
States in the years ahead has been made by Harvard professor Simon
Kuznets. 23 In his monumental study, Kuznets concluded that during
the preceding 20 years (1941-60) the rate of economic growth of the
Capital in. the American Economy-Its Formation and Financing, a study by tha
National Bureau of Economic Research, published by Princeton University Press, 1961.
The Icusnets monograph was the summary volume in a project financed with a grant of
$460,000. Other monographs, all published by the Princeton University Press, were: (1)
Capital Formatioav in Residential Real Estate: Trends and Prospects (Leo Grebler, David
M. Blank, and Louis Winniek) ; (2) Capital in Agriculture: Its Formation and Financiaag
since 1870 (Alvin Tostlebe) ; (3) Capital in Transportation, Communicctioaos, and Poeblic
Utilities: Its Formation and Fiosanciasg (Melville Ulmer) (4) Capital in Manufacturing
avid Mining: Its Formation and Financing (Daniel Creamer, Sergei P. Dobrovoisky, and
Israel Borenstein) ; (5) Trends in Government Financing (Morris A. Copeland) ; and (ul~
Financial Intermediaries in the American Economy ~S'ince 1900 (Raymond Goldsmith).
PAGENO="0196"
188 OLD AGE INCOME ASSURANCE-PART V
United States had been held considerably below its potential because of
too low a rate of saving. He found that the demands for capital financ-
ing will be very large in corning years, and that high levels of consump-
tion are likely to continue. Therefore, he concluded, the supply of
savings in the years ahead will be inadequate., and lie strongly recom-
mended measures to encourage a higher rate of saving. Kuznet's prin-
cipal conclusions are well summarized in the following quotation:
The above treatment of conditions for the future, far too brief and dogmatic,
may carry a sense of firmer conviction than is intended. The discussion reflects
conclusions suggested by the record of this country's economy-an economy geared
for a longrun rise to increasingly high levels of consumption per capita, and one
in which savings and capital formation, though large and sustained, nevertheless
have been kept within moderate proportional limits by the secularly high propeil-
sity to consume. It is also an economy in which the recently increasing diversion
of product to current consumption by governments, combined with high levels of
consumer demand, has limited capital formation and savings proportions and
brought about, under conditions of full employment, rising price levels which
have persisted even through the 1058 recession. Against this background, consid-
eration of the prospective large rise in population numbers-particularly of new
familyinakers, entrants into the labor force, and of the school-age groups-the
prospective acceleration of potential technological change and the prospect of
continued international competition and strain suggest the following prospects.
First, the demand for capital over the coming two and a half to three decades is
likely to be large. Second, drains upon the national product for current consump-
tion by governments will continue to be proportionately sizable and may well rise.
~-` Third, high levels of consumption and the high secular propensity to consume
by individuals and households are likely to continue. Fourth, under the cir-
cumstances, the supply of voluntary savings may not be adequate. Finally, infla-
tionary pressures may well continue, with the result that part of the savings
needed for capital formation and government consumption will be extracted
through this particular mechanism. Yet, extrapolation of inflationary pressures
over the next 30 years raises a specter or intolerable consequences, making
the policy solutions adopted critically important; and those solutions, in turn.
will affect the structure and pattern of financial intermediaries and their role
in flnancing.~
The views of Sommers, Lecht, and Kuznets, as well as those in the
Joint Economic Committee study, are characteristic of the great ma-
jority of economists today. The American economy has great potential
for vigorous growth in the years ahead. Public policies will be directed
to maintaining full employment and faster growth. If growth is to
be realized soundly and without a sha.rp rise in the general price level,
the rate of saving will have to be high. It follows, therefore, that pen-
sion plan savings should be encouraged-not discouraged, as some of
the passages in the joint committee print would seem to suggest.
PRIVATE PENSION SAVINGS AND LONGRUN INVESTMENT PLANNING
Economists have long recognized that cyclical fluctuations in capital
spending have played a major role in t.he general business cycle. A
fundamental objective of public policy must be to introduce greater
regularity and sust.ainability in the flow of investment expenditures.
Pension savings serve this purpose well, for two main reasons: (1)
they are contractual in nature and hence flow with regularity and pre-
dictability into the institutions administering the funds; and (2) a
high proportion of pension savings are invested on a forward commit.-
ment basis which facilitates longrun planning by business and in-
dustry of their capital expenditures.
~ Kuznets, pp. 459-460.
PAGENO="0197"
OLD AGE INCOME ASSURANCE-PART V 189
Very little needs to be said about the contractual nature of pension
saving. Under normal circumstances, the cash flow for investment of
institutions administering pension funds is predictable with only a
small margin of error. Consequently, these institutions are in a posi-
tion to enter into forward investment commitments with business and
industrial concerns, with public utilities, with homebuilders, and with
State and local government units to buy their bonds and mortgages.
Under such forward commitments the actual disbursement of the loan
funds is in many instances spread over a period as long as 2 or 3 years.
Under the forward commitment process, a corporation desiring
funds to expand its plant and equipment can, through the route of a
direct placement of its bonds with a pension institution, obtain financ-
ing on terms fixed at the time the loan was negotiated. Disbursement
of the funds will occur as the funds are needed to complete the con-
struction project. Similarly, a homebuilder planning a housing devel-
opment can obtain his financing on the same forward commitment
basis. Every type of construction project lends itself to the forward
commitment process, ranging from gas pipelines, jet airplanes, and
office buildings to superhighways and toll bridges.
Tables 5 and 6 show the outstanding forward investment commit-
ments as of January 31, 1967, of a group of life insurance companies
holding about 70 percent of the assets of all TJ.S. life companies.25
Only 44 percent of these commitments were expected to be taken down
within 6 months. A considerable part of the remainder was committed
for 1,2, or even 3 years ahead.
TABLE 5.-Outstanding Commitments on Jan. 31, 1967, of Reporting Life
Insurance U~mpanies 1
MimoSa
Total commitments ~9, 511. 7
Securities, total 3, 262. 0
Business and industrial bonds 2, 736. 2
Public utility bonds 152. 2
Railroad bonds 50. 0
State, municipal and local securities 13. 6
All other securities 310. 1
Mortgage loans and real property, total 6, 249. 7
Business and industrial mortgages 3, 363. 0
Farm mortgages 188. 6
Nonfarm residential mortgages 2 2, 231. 4
FHA insured 213. S
VA guaranteed 99. 7
Conventional 1,917. 8
Real property for lease or rental 424. 3
All other real property 42. 5
Nova-Because of rounding, components may not add to totals shown.
1 Comnanies covered in this report hold 70 percent of the assets of all U.S. life insurance
companies.
2 Memo:
1 to 4 family 429. 2
Multifamily 1, 747. 2
This detail by property is not provided by all reporting companies.
21 The assets of these companies are not, of course, all pension fund assets, but the pat-
torn would he the same for commitment of funds derived from pension reserves as well as
regular life insurance reserves.
PAGENO="0198"
190 OLD AGE INCOME ASSURANCE-PART V
TABLE 6.-EXPECTED TAKEDOWNS OF COMMITMENTS OUTSTANDING JAN. 31, 1967, OF REPORTING LIFE
INSURANCE COMPANIES'
Outstanding commitments (mil- Percent of outstanding com-
lions of dollars) expected to mitments expected to be
be taken down within- taken dawn within-
1 month 2 months 6 months 1 month 2 months 6 months
Total cammitments $810.3 $1, 561. 5 $4, 198. 4 8. 5 16. 4 44. 1
Securities, total 347.4 660.9 1,567.5 10.6 20.3 48.1
Business sod industrial bonds 280.9 527.8 1,308.6 10.3 19.3 47.8
Public utility bonds 14.0 30.7 68.5 9.2 20.2 45.0
Rallrowi bonds 10.4 19.1 31.7 20.8 33.3 63.4
Stats, municipsi, and local securities 3.9 7 4 11.2 24.8 55.0 82.7
All other securities 38.1 75.9 147.5 12.3 24.5 47.6
Mortage loans and real property, total 462.9 900.6 2,631. 0 7.4 14.4 42.1
Business and industrial mortgages 203.7 420. 8 1, 272.2 6.2 12. 5 37. 8
Farm mortgages 38.5 68.5 134.7 20.4 33.3 71.4
Nonfarm residestial mortgages ` 198. 2 369. 1 1, 098. 1 8. 9 16. 5 49. 2
FHA insured 28.9 57.2 120.3 13.5 23.7 56.2
VA guaranteed 9.5 23.2 69.0 9.5 23.2 69.2
Conventional 159.8 288.8 908.8 8.3 15.1 47.4
1-4 tamily2 (64.2) (128.9) (338.6) (15.0) (30.0) (78.0)
Multifamily2 (127.9) (231.2) (742.1) (7.3) (13.2) (42.5)
Real property for lease or rental 16. 5 40.4 97.4 3.9 9. 5 23.0
All other real property 1.0 1.9 28.6 2.2 4.5 67. 2
Note: Because of rounding, components may not add to totals shown.
Companies covered in this report hold 70 percent of the assets of all U.S. life insurance companies.
2 Memo: This detail by property is not provided by all reporting companies.
Accordingly, rather than being a destabilizing force, as the joint
committee print suggests, the contractual nature of pension savings
and the investment of such funds on a forward commitment basis are
powerful forces for stability in the economy and for sustainable long-
term growth.
PENSION SAVINGS. INVESTMENT, AND RISK CAPITAL
As noted earlier, the joint committee print asks whether the invest-
ment of pension savings has not been too conservative and thus not
conducive to vigorous economic growth. This question appears to
reflect a misunderstanding.
It is true, of coursea that pension fund investments must be made
with due regard to the safety of the principal of the investment. But,
this does not. mean that the funds are not placed in highly productive
uses. It. does not mean that the funds cannot. be invested with imagina-
tion and with a dedication by the investment managers to find the most
productive outlets.
The greatest proportion of corporate bonds purchased by msured
pension funds today-and for the past 15 years-have been acquired
by means of direct placements. A high proportion of the corporate
bonds purchased by noninsured funds are also direct placements. By
dealing directly with the borrowing firm, pension fund officers are
enabled to appraise risk and growth potential. They are in a position
to underwrite loans and to be paid for any investment risks taken in
the financing of firms in new industries and firms with a high growth
potential. This is also true of mortgage loans to business and industry.
PAGENO="0199"
OLD AGE INCOME ASSURANCE-PART V 191
In short, pension fund investing has been a powerful force for
vigorous growth in the American economy. Perhaps the best way to
illustrate this point is to cite some examples of investments made by
life insurance companies; the funds involved would, of course, come
both from regular life insurance and insured pension savings.
1. In the early post-World War II period the long-distance natural
gas transmission industry was still relatively small. In the 20 years
since that time, pipelines have been built to connect virtually every
State of the Union with the reserves of the South and Southwest to
bring the benefits of natural gas to the whole country. This tremendous
growth has helped to make the natural gas industry the sixth largest
industry in the United States. Life insurance companies have invested
billions of dollars in gas pipelines. For example, m the period since
1946, one life insurance company has purchased a total of $1,717 mil-
lion of bonds for the construction and expansion of natural gas
pipelines.
2. The development of the great commercial airline fleets, first
with the propeller-driven planes, then with jets, and now with the
superjets, has been a big factor in our economic expansion since
World War II. Life insurance companies have provided a great deal
of the financing. For example, since 1946 one life insurance company
has invested $485 million in the airline industry, and another more
than $300 million.
3. An important step in prolonging the life of the great Mesabi
iron ore range was the development of a method for processing taco-
nite (low-iron-content rock) and pelletizing it into iron concentrate
for direct use by the steel mills. Life insurance companies have pro-
vided much of the financing for the development of taconite. One life
company invested a total of approximately $145 million in the senior
obligations of two mining companies to finance such projects. The
same life company also played a large part (together with other life
companies) in a further solution of the problem of the depletion of
the primary iron ore sources in the United States by financing two
Canadian projects for the development of iron ore mining and proc-
essing facilities in Labrador. This company invested a total of
$98,400,000 in the senior securities of these two projects. The same
company also purchased a total of $42,500,000 of bonds to help fi-
nance the construction of a hydroelectric power development in Lab-
rador to furnish power to one of these projects.
4. In 1950, a life insurance company loaned an aluminum company
$20 million. Over the years the same life company purchased $105
million of the $474 million of bonds sold by t.his company. Other life
insurance companies purchased most of the remainder. This tre-
mendous volume of funds was required to finance a major expansion
program. The aluminum company entered the aluminum field in 1946.
Now, it also produces various chemicals and refractories. Growth over
the past two decades has made the company one of the largest domes-
tic producers of aluminum, with worldwide operations. The company's
aTuminum sales have increased more than tenfold between 1950 and
1~66, and the number of employees duringthis period has increased
from about 5,000 to 27,000.
5. A life insurance company issued three recent commitments cover-
ing loans on new medicenter properties in Boston, Fort Worth, and
PAGENO="0200"
192 OLD AGE INCOME ASSURANCE-PART V
Raleigh, N.C. The commitments are to a. relatively new corporation
that is undertaking nursing care to provide postoperative treatment
for ambulatory patients. This will be at lower costs than available in
hospitals and should reduce the overall cost of medical care. It is
anticipated that a sizable percentage of patients will come under the
Federal medicare program. The medicenters should also free use of
hospital beds which are presently in such short supply. The life insur-
ance company's commitments are presently $1,140,000 in Boston and
$900,000 each in Fort Worth and Raleigh.
6. A life insurance company in 1946 made a loan of $565,000 to a
company engaged in processing soybeans and related products. The
purpose of the loan was to help the company finance working capital
and plant expansion. Total sales at that time were $56 million. An
additional loan of $10 million was made in 1951 when the company's
sales were $100 million. Today the company's sales are approximately
$560 million. Its active research program has provided valuable in-
formation on protein and has helped in the development of an eco-
nomic protein additive which may be a partial answer to malnutrition
in many parts of the world.
7. A life insurance company made several loans to a company in
the forest products industry between 1959 and 1963 to finance plant
expansion and acquisitions into broader product areas. In 1964, these
various loans were refinanced into one total issue of $110 million. Dur-
ing this period sales had grown from $73 million to $331 million. The
company has become an important factor in the development of nat-
ural resources of the United States and has been a major supplier of
raw materials to the building industry.
8. In 1963, a company headquartered in Baltimore approached a
life insurance company with the idea of buying enough land to build
a city of 110,000 people. This was the origin of the city of Columbia,
in Howard County, Md. The life insurance company subsequently ac-
quired 15,000 acres and now has about 20 different projects under con-
struction, plus roads, utilities and other public facilities. This will
be a complete city offering not only a place in which to work but also
all of the features normally found in a large metropolitan area. It is
contemplated that there will be jobs for 30,000 people in the city. Op-
portunities will range for people to be employed or to express them-
selves in the performing arts, medical care, and a variety of industries
and recreational activities. The life insurance company has so far com-
mitted a total investment of $26 million and other investors have com-
mitted an additional $25 million.
9. In 1957, when atomic power for electric utilities was still in its
infancy sta.ge, life insurance companies agreed to purchase $20 million
of first mortgage bonds of an atomic powerilamit. This financing rep-
resented 35 percent of the cost of a $57 million project to build a new
134,000 kilowatt atomic-powered electric generating plant in Rowe,
Mass. While construction of atomic powered generating plants has
become rather commonplace today, 10 years ago the concept of atomic
fuel for electric generation was new. There had been no private financ-
ing of any atomic plant at the time the life insurance companies
agreed to purchase these bonds. In fact., to date, there still have been
only three privately financed atomic generating plants, of which the
one under discussion was the first.
PAGENO="0201"
OLD AGE INCOME ASSURANCE-PART V 193
10. In December 1938, a life insurance company purchased directly
from a textile manufacturing company an entire issue of $1,875,000
of debentures due in 1948. The li-fe company has since purchased
additional issues in their entirety in 1939, 1940, and 1951 and partici-
pated (together with other insurance companies) in subsequent direct-
placement transactions in 1960, 1963, and 1964. At December 31, 1966,
the life company held $26 million of the textile company's note obli-
gations. The remarkable growth of the textile company since the date
of the first financing is reflected by the fact that between 1938 and
1966 its net sales increased from about $27 million to about $1.4
billion and its total assets from about $16 million to about $1 billion.
Although part of this growth was due to acquisitions, it may be of
interest to note that the number of employees increased from about
8,000 on January 1, 1939, to some 69,000 on October 1, 1966.
These are but a few examples of the way in which life insurance
company investments, including pension fund investments, have con-
tributed to the economic growth of the United States. More are in-
cluded in an appendix heret.o. If additional examples are desired, a
multitude can be made available by the Life Insurance Association
of America and the American Life Convention.
RECOMMENDATIONS FOR THE ENCOURAGEMENT OF PENSION SAVINGS
ABROAD
It seems ironic that a question should be raised concerning the
desirability of encouraging private pension saving here in the United
States at the very time when foreign countries are urging increased
pension saving in order to permit the sound financing of economic
growth. For example, in the recent report The Development of a
European Capital Market, which was prepared by a group of experts
appointed by the European Economic Community (EEC) Commis-
sion, it is pointed out that:
The way in which financing requirements are met depends to a large extent
on the composition of the funds available. It is this composition rather than
a shortage of savings which explains the structural imbalance to be seen in
certain sections of the capital markets in the member states. The supply of
funds reaching these markets reflects primarily the following points: (i) Savers
generally prefer to hold cash or short-term assets and it would be difficult to
alter this liquidity preference radically in the short-term; (ii) Savings which
take the form of deposits will remain a major factor in the supply of capital.
The financial institutions which collect this form of savings should, therefore,
be able to offer medium- and long-term loans and to use their funds on the
capital market; (iii) Saving through institutional investors, particularly con-
tractual savings, is inadequate in all member countries except the Netherlands.
This inadequancy in part explains why markets are not able to function
satisfactorily.26
The report went on to add:
This state of affairs does not seem likely to undergo a spontaneous change,
especially now that income redistribution to the advantage of wage earners
has raised the savings potential of social classes still unprepared to take a direct
interest in the capital market. Consequently, several member states are stimu-
lating the consolidation of savings and the various forms of contractual
saving.27
26 The Development of a European Capital Market, Summary of the Rep'ort of a group
of experts appointed by the EEC Commission, Brussels. November 1966, pp. XV-XVI.
~ Ibid., p. XVI.
PAGENO="0202"
194 OLD AGE INCOME ASSURANCE-PART V
Again, on p. XXI of the report, in discussing incentives to saving,
the EEC Commission experts point to the need to give "top priority"
to steps to encourage "contractual savings-in particular those linked
with employment contracts."
Early this year the Business and Industry Advisory Committee to
OECD and the Atlantic Institute jointly sponsored a major con-
ference on the European capital markets. This conference, held in
Cannes, January 19-22, 1967, concluded:
Increased investment is required to assure a rapid increase of production and
productivity. With monetary stability and a high level of employment, this
brings higher real wages for all. This sequence is the essence of sound economic
growth. Both governments and private enterprise require even greater quan-
tities of investment capital as a consequence of the growth of population and
the quickening pace of technical progress. At the same time O.E.O.D. member
countries ought to increase their flow of capital to developing countries.
This growing demand for capital is not being met by comparable increases in
supply. To meet the additional needs, measures must be taken to improve capital
markets. Moreover, recourse must be had to more effective use of 1)udgetary
policy and adequate self-financing for public and private enterprise.2S
In order to improve the capital markets and to increase the supply
of capital in OECD countries, a number of recommendations were
advanced by the conference. Recommendation II called for action to
encourage private contractual savings as follows:
II. To stimulate and expand private contractual savings should be one of the
primary policy objectives of the governments. In order to implement this recom-
inendation, the following proposals are made:
A. Pension funds. insurance, and other forms of contractual savings should
be stimulated by tax and other inducements, and by further enabling legis-
lation in those countries where it is now lacking.
B. Governmental regulations restricting institutional purchases in capital
markets should be relaxed as much as possible.
C. In the field of insurance, the use of funding techniques should be
encouraged.~
Thus, as European countries strive to maintain full employment and
a st.rong rate of economic growth, it is becoming more and more clear
that they must develop and enlarge their capital markets. They ap-
preciate that a high rate of capital formation under full employment
conditions must be financed from savings rather than from monetary
expansion. It is not surprising, therefore, that they have recognized
that to enlarge the flow of savings for capital formation they must
take steps to encourage contractual savmg through private pension
funds and insurance. The American system of private pension saving
is the envy of countries around the world which arc striving to find a
sound basis for financing economic growth.
SUMMARY AXD CoNcLUsioNs
The main Points ircsentedl in this PaPer may be summarized as
follows:
1. The private pension system has a record of outstanding accom-
plishment and has clearly demonstrated its capacity for growth and
improvement. Private, pension plans now cover about 25 million em-
2S RecommencThtions from ilip Conference o~ Copital .~fai7:ets, Business and Industry
Advisory committee to the OECD and the Atlantic Institute. Cannes. January 10-22. 1067.
p. 1.
nIbicl., p. 1.
PAGENO="0203"
OLD AGE INCOME ASSURANCE-PART V 195
ployees, and there is a strong trend toward increased coverage and
more rapid funding and vesting.
2. Pension plans benefit the rank and file of employees. The bulk
of such plans qualify under provisions of the Internal Revenue Code
designed to insure that they do not discriminate as to coverage and
benefits in favor of highly paid employees as compared with employees
with modest incomes.
3. A private pension system, which continues to grow and continues
to improve, is essential for achieving the best possible retirement pro-
tection for our population. Pension plans offer unique advantages for
this purpose, in view of their flexibility and ability to adjust to the
individual circumstances of particular groups of employees in dif-
ferent firms, industries, and geographical locations.
4. The social security program is a basic ingredient in our system
of providing retirement protection. But it is essential to keep a proper
balance between private pension plans and the social security system.
The latter should not be expanded in wage base and benefit levels to
the point where it takes over retirement functions which can be per-
formed better by the private sphere. At the same time it is important
to continue to improve private pension plan coverage, vesting, and
funding so that pension plans which are now doing a good job can
do an even better job. The objective is to develop new pension plans
and to improve existing ones so that the maximum number of indi-
viduals can benefit from them.
5. The Joint Committee Print's concern that pension saving may
have a depressing effect upon the rate of economic growth of the
United States is unrealistic. It flies in the face of the experience of
the past two decades. If the rate of saving has been excessive, as the
Print suggests, how can we explain the persistent upward trend of
long-term interest rates during the past 20 years? If effective demand
for goods and services has been chronically weak, how can we explain
the upward drift of the price level since World War II? The U.S.
Government is committed to pursuing fiscal and monetary policies
designed to maintain full employment and strong economic growth,
with stability in the value of the dollar. Such policies will require a
very high rate of saving and capital formation, as is assmned in all
of the projections of the growth of the American economy in the years
ahead. Viewed in this light, and not in the shadow of the "stagnation
thesis," the Print's concern about private pension saving is without
justification.
6. The Print's concern that pension saving, because of its contrac-
tual nature, tends to be a destabilizing force in the economy is also un-
founded. The fact is that the contractual nature of pension saving is
highly advantageous from the standpoint of economic stability. Inas-
much as the cash flow for investment of pension funds is regular and
predictable, institutions administering pension funds have been able
to make forward investment commitments which aid business and
mdustrial firms to plan their capital expenditures on a long-run basis.
In an economy in which business and industry expects appropriate
fiscal and monetary policies, longrun planning of capital expenditures
has become realistic and has been encouraged and facilitated by the
contractual nature of pension funds and the forward investment corn-
initment process.
PAGENO="0204"
196 OLD AGE INCOME ASSURANCE-PART V
7. The Print's assertion that institutions administering pension
funds do not contribute to vigorous economic growth through their
investments is entirely at odds with the facts. Generally speaking,
private pension savings have been directed into highly productive
outlets. The examples which we have presented are typical of the way
pension funds are invested-with imagination and with high poten-
tial for economic growth. At the same time, these investments have
been made with safety, as the record of virtually no investment losses
in the past two decades attests.
8. In the decade ahead-indeed, for the foreseeable future-there
will be an urgent need for a high rate of saving if we are to achieve our
i~ational goals of full employment and faster economic growth with
reasonable stability of the value of the dollar. To achieve the rate of
saving necessary for growth, we must have a healthy expansion of
private pension saving. As Kuznets and others have pointed out so well,
there have been powerful forces operating in the past to lower the rate
of saving, and these forces will persist. It is even more necessary, there-
fore, to encourage the growth of contractual savings such as those
accumulated through private pension ftmcls.
9. Finally, other countries in the free world-notably in Europe-
are so convinced that contractual savings are essential to economic
growth that they are urging Government measures to stimulate such
savings. The really pertinent questions about private pension savings
are: In view of the fact that a very high rate of private pension saving
will be sorely needed in coming years to aid in financing sound eco-
nomic growth in the TJnitecl States, Is the Government doing enough
to encourage pension saving? What further steps can be taken to
strengthen the flow of pension savings? These are the significant ques-
tions to be asked as we look to the future.
PAGENO="0205"
OLD AGE INCOME ASSURANCE-PART v 197
APPENDIX
Additional examples of the way in which life insurance company
investments, including pension fund investments, have contributed to
the economic growth of the United States are:
1. Some years ago, a firm was established to manufacture high fre-
quency electronic tubes with emphasis on klystron tubes, which had
been invented in 1938. A life insurance company loaned this firm
$2 million, in 1956, with the funds being used to retire bank debt and
expand the plant. With the help of this loan, the firm grew to the
point where its gross revenues amounted to $145 million in 1966, com-
pared with $7.2 million in 1955. The company is thriving and has
become a diversified electronic manufacturer with foreign operations.
2. A new company was established to engage in the application of
paint and other synthetic finishes to continuous coils of steel, alumi-
num, and other metals. The company was a pioneer in the industry. A
life insurance company helped this innovative process by lending the
firm $800,000, in 1959. Additional loans of $400,000 and $740,000 were
made in 1962 and 1965, respectively. All of these loans were used to
finance additional productive capacity. The company's sales rose five-
fold between 1958 and 1966, with a proportionate increase in income
and employment.
3. Tn 1958, a life insurance company made a loan of $115,000 to a
farm wagon manufacturer. In addition to farm wagons, the company
had a mobile hydraulic crane under development. There have been five
additional loans made to the company since that time and the total
outstanding in 1966 was $1,900,000. The loan funds facilitated the
growth of the wagon manufacturing firm and made possible the
development of the mobile hydraulic crane. During this period sales
and employment of the firm have increased about fifteenfold.
4. Tn 1958, a life insurance company made a $360,000 loan to a
California manufacturer of metal containers for shipping and hous-
ing electronic components. Loans to this firm have been increased
since 1958 and now total $2,800,000. During this period the sales of
the manufacturer have risen from $3 million to $17 million and net
worth has increased 51/2 times. Employees now number 1,200, com-
pared with 180 in 1958.
5. A small Wisconsin company borrowed $151,000 from a life in-
surance company in 1964. The company makes portable hydraulic
loaders for trucks and tractors which are used primarily to load bulky
items such as logs and cement blocks. The business had been formed
only about 9 years earlier and it is very doubtful that it could have
obtained funds in the public market. There were two additional loans
made by the life insurance company and the total investment, in 1966,
was $400,000. In 1964 the manufacturer had 86 employees and last
year it had 220.
6. In 1960, three insurance companies purchased $16,800,000 of first
mortgage bonds of a pipeline in Alaska. This represented 86 per-
cent of the cost of building a 78-mile natural gas transmission pipe-
line from the Kenai Peninsula to the city of Anchorage, Alaska. Two
oil companies who owned and developed the natural gas reserves on
the Peninsula advanced the equity moneys of $2,600,000. The building
PAGENO="0206"
198 OLD AGE INCOME ASSURANCE-PART V
of this pipeline from the peninsula, across the open water to the city
of Anchorage brought for the first time to the 70,000 people in An-
chorage a cheap source of fuel available for both heating and the
generation of electric power. It also permitted the continued develop-
ment of the gas fields offshore by the two oil companies. Prior to the
delivery of natural gas to the city of Anchorage, the only source of
fuel were high-cost coal and oil. Completion of the pipeline also
resulted in the expansion of the gas distribution system serving
Anchorage, operated by a wholly owned subsidiary of the pipeline
company. Prior to the delivery of the natural gas, the distribution
company in Anchorage had been a very small company delivering gas
obtained as a byproduct of burning coal and oil.
7. In 1943, a life insurance company loaned $2 million to a rubber
manufacturing company having sales aggregating about $14.1 mil-
lion. In 1966, this company reported annual sales of $106.1 million.
During the intervening 23 years, the life insurance company has made
12 different loans to the rubber company, substantially contributing
to its growth. The life insurance company's present investment in
the company is $15 million.
8. During the period 1951-67, a life insurance company purchased
or participated in the purchase of 15 debt obligations issued by an
aluminum company and two of its affiliates. These purchases aggre-
gated $260,810,000 (including a present commitment of $30 million).
The life company's investment in such debt obligations at December 31,
1966 (including the commitment), was $170,794,000. The dynamic
growth of this major integrated aluminum producer during this
period is shown by the fact that between 1951 and 1965 its net sales
increased from about S216 million to about $740 million, and its total
assets from about S319 million to about $1.2 billion.
9. In 1946, a life insurance company invested $300,000 in a company
which nmnufactures chemical coating. The finishes are used on a
multitude of Products including appliances, furniture, wall paneling,
automotive equipment and a variet of other items. The company re-
cently entered the field of high-pressure laminated plastic sheets for
use on furniture and in residential and commercial building construc-
tion for counter tops~ wall paneling, etc. At. the time of the original
investment, the manufacturing company's assets were $1.5 million.
The life company's investment in this company has grown to over $4.1
million and the manufacturing company's assets are now over $30
million.
10. A life insurance company's investment in an industrial gas com-
pany has risen from $3 million, in 1961, to over $1.9 million. The gas
company's assets are now about $50 million, almost five times what
they were 6 yea.rs ago. Its maj or operation involves the recovery and
sale of industrial gases for use primarily in welding, oil well servicing~
and certain chemical processes. In 1962, the gas company constructed
one of the. first completely automated nitrogen gas plants in the
country. The company serves a market principally in Texas and
Louisiana, and, more recently, California and Florida, through new
plants in these States.
PAGENO="0207"
GUARANTY FUND FOR PRIVATE PENSION
OBLIGATIONS
BY DAN M. MCGILL*
CONTENTS
Page
THE SETTING 200
Nature of the Employer's Undertaking 201
Methods of Financing 201
HAZARDS TO BENEFIT FULFILLMENT 203
APPLICABILITY OF INSURANCE CONCEPTS TO A PENSION GUARANTY FUND~ - 206
Large Number of Homogeneous Risks 206
Objective Determination of Occurrence and Amount of Loss 207
Randomness of Loss 208
Low Probability of Loss 208
Significance of Loss 209
Absence of Catastrophe Hazard 209
RELEVANT PRECEDENTS 210
Governmental Programs 210
Private Programs 212
Plans Jointly Administered by Governmental and Private Agencies - - 213
Swedish Pension Guaranty Fund 215
ISSUES 217
Administering Agency 218
Event Insured Against 218
Obligation of the Guaranty Fund 222
Assurance of Ultimate Payment of Guaranteed Benefits 222
Completion of the Employer's Funding Commitment 224
Plans Covered 225
Benefits Covered 227
Implementation of the Guaranty 229
Determination of the Guarantor's Liability 229
Manner in Which the Guaranty Would Function 230
Financing the Guaranty 232
A MINIMUI~i PROGRAM 235
Administering Agency 235
Event Insured Against 235
Obligation of the Guaranty Fund 237
Plans Covered 238
Benefits Covered 239
Implementation of the Guaranty 239
Financing the Guaranty 240
SUMMARY 241
Within the last few years, considerable interest has developed within
certain quarters in some type of cooperative arrangement that would
assure the fulfillment of legitimate benefit expectations under private
pension plans, irrespective of the financial status of the plans or their
cWharton School of Finance and Commerce, LTniversity of Pennsylvania.
199
PAGENO="0208"
200 OLD AGE INCOME ASSURANCE-PART V
sponsors. The idea was given great impetus and a measure of respecta-
bility when the President's Committee on Corporate Pension Funds
suggested that serious study should be given to the possibility of estab-
lishing "a system of insurance which, in the event of certain types of
termination, would assure plan participants credit for accrued bene-
fits." 1 Later the National Commission on Technology, Automation,
and Economic Progress, in its report to the President and the Con-
gress, under the heading, "Protecting the Earned Benefit Rights of
Displaced Employees," stated that:
We favor whatever legislative or administrative measures may be necessary
to promote greater equity and security in the establishment and administration
of private pension plans. Specifically, we recommend that careful study be given
to a legislative system of reinsurance for private pension plans similar to the
reinsurance [sic] of bank deposits through the Federal Deposit Insurance
Corporation.
More recently, Nelson McClung and his fellow staff economists pre-
pared a document for the Joint Economic Committee of Congress in
which they espoused the cause of a pension guaranty fund, not only
in the interest of benefit security but as a means of accommodating a
lower level of funding.2
The concept has been embodied in various legislative proposals, the
earliest of which was a bill introduced by Senator Vance Hartke of
Indiana, in 1964, and reintroduced with minor modifications in each
subsequent session of Congress. It is contained in the omnibus bill,
S. 1103, introduced by Senator Javits during the current session of
Congress.
This proposition cannot be evaluated without at least a skeletal
description of the milieu in which private pension plans operate.
THE SETTING
The primary purpose of a pension plan is to provide old-age income
to retired workers and their widows, the income to continue throughout
the remaining lifetime of the individuals involved. The plan may, and
frequently does, provide benefits in the event of the employee's death,
extended disability, or voluntary withdrawal from the working force
prior to retirement. Under a noncontributory plan, the employer (or
group of employers, in the case of a multiemployer plan) assumes sole
responsibility for providing the benefits cont.emplated under the plan,
although some or all of the cost of the plan may ultimately be shifted
to the employees (through lower wages), the consumers (through
higher prices), or the tax-paying public (through tax deductions).
Even under a contributory plan, the employer generally assumes the
principal burden of financing the benefit structure. Thus, for the sake
of simplicity, the objectives envisioned for the pension plan will be
referred to as the emplo~'er's undertaking, even though the covered
employees may share in the financing of the plan and participate
(through their elected representatives) in its administration.
1 "Public Policy and Private Pension Programs", Report to the President on Private
Employee Retirement Plans (Washington, D.C.: Government Printing Office, January
198~l. p. 80.
2 Old-Age Income Assurance: An Outline of Issues and Alternatives, 89th Cong., 2d sess.,
1966, pp. 29-30.
It may be argued that under a collectively bargained plan, the employees consciously
absorb the estimated cost of the plan by accepting lower cash wages. By the same line of
reasoning, one could assert that these employees would bear the cost of a guaranty program.
PAGENO="0209"
OLD AGE INCOME ASSURANCE-PART V 201
NATURE OF THE EMPLOYER'S UNDERTAKING
The obligation assumed by an employer in establishing a pension
plan may take one of two forms: an undertaking to set aside funds
on a specified basis or an undertaking to provide benefits in accord-
ance with a specific schedule. Under the first approach, the employer
does not undertake to provide retirement a.nd other benefits in accord-
ance with a fixed scale of benefits, although the scale of contributions
will normally be set in the light of an anticipated level of benefits. The
contribution commitment may be stated as a percentage of the compen-
sation of covered employees, as an amount per hour or day of work,
or in terms of some productivity factor. Regardless of how the com-
mitment is expressed or determined, the employer's obligation to the
plan is deemed to be completely fulfilled when he pays over the appro-
priate sums, even though the assets of the plan eventually prove insuf-
ficient to provide the projected level of benefits.
Under the second approach, the employer ostensibly commits him-
self to contribute whatever sums are necessary to provide the benefits
payable pursuant to a formula or schedule set forth in the plan. The
plan may call for a unit of benefit for each year of credited service,
a composite benefit equal to a specified percentage of compensation (for
the entire period of employment, or the years immediately preceding
retirement), or a composite benefit expressed as a fiat dollar amount.
The plan usually recognizes, and gives credit for, some or all of the
service performed for the employer in question prior to the inception
of the plan, and subsequent benefit liberalizations are generally made
retrospective as to years preceding the plan amendment, both prac-
tices giving rise to an unfunded accrue,d liability that would be the pri-
mary source of loss to any guaranty arrangement. Except for COl-
lectively bargained plans, the employer reserves the right to alter,
modify, or terminate the plan at. any time-and for any reason.4 More-
over, he generally reserves the right to suspend, reduce, or discontinue
contributions to the plan, whether or not previous contributions have
been sufficient to provide all benefits credited to date. Finally, it is
customary for the employer to limit his obligation, in the event of
plan termination, to contributions already made to the plan. In other
words, if the plan terminates, the participants and pensioners must
look to the accumulated assets for the satisfaction of their claims. In
a relatively small number of cases, largely confined to the petroleum
and steel industries, the employer guarantees to provide all accrued
benefits (sometimes only vested benefits) in the event of plan termina-
tion, in effect pledging the corporate resources to the satisfaction of
accrued pension claims.
METHODS OF FINANCING
In order to meet the benefit commitments, explicit or implicit, gen-
erated under a pension plan, the employer generally sets aside funds
During the first few years of a plan's existence, it can be terminated without retro-
active tax penalties only for reasons of "business necessity."
R3-200-67-pt. 5-14
PAGENO="0210"
202 OLD AGE INCOME ASSURANCE-PART V
with a bank or insurance company in amounts and at times roughly
commensurate with the rate at which the pension costs accrue. In other
words, the employer, through an independent funding agency, usually
accumulates funds for the payment of pension and related benefits in
advance of the dates when the payments are due, a practice known as
funding. The general procedure is to fund currently accruing benefits
during the accounting period in which they accrue. Benefits credited
for service prior to inception of the plan are funded, if at all, over an
extended period, usually ranging from 12 to 40 years. If the employer
is to get a current tax deduction for these contributions (or advance
paymei~ts). the moneys must be held under an arrangement that en-
sures their use for the exehTsive I)enefit of the plan participants or their
beneficiaries. This requirement is met through transfer of the funds
into an irrevoe~ble trust. generally administered by a bank, or through
the purchase of insurance or annuity contracts from an insurance com-
pany, including those contractual arrangements under which the funds
are not allocated to specific individuals until retirement. In some cases,
only the benefits of individuals in a. retired status are funded, a practice
known as terminal funding.
Under a. relatively small percentage of plans, the employer pays
the benefits directly to retired employees, with no advance funding
through intermediaries. Some of these plans are large but the great
majority are associated with small firms. Occasionally, the basic pen-
sion plan of an employer will be ftmded, with additional benefits being
provided on a current disbursement or pay-as-you-go basis. This
method of financing would create complications for a pension guaranty
arra.n~einent.
Un~ler existing law, an employer is under no legal obligation to
fund his accuring pension costs, and he may deduct as an ordinary and
necessary business expense reasonable payments to retired employees
or their beneficiaries. However, if he wants to enjoy the tax treatment
accorded "qualified" pension plaims, which would include exemption
from current. taxation of investment earnings on plan assets, he must
as a minimum fund the normal (or current) cost of the Plan Plus in-
terest on the initial supplemental liability at the rate assumed in the
calculation of the liability. This requirement, which is believed by
Some authorities to be applicable only during the first 10 years of a
plan's existence. is designed to discourage a firm from establishing a
pension plan with the deliberate intention of terminating it within a
few years. The Internal Revenue Service does not inquire into the
adequacy of the cost estimates underlying the funding payments, its
chief concern being protection of the Federal revenue against excessive
tax deductions.
Traditicnaliy, the actual contributions to a pension plan (or, under
a current disbursement plan, the actual payments to benefit. recipients)
have been construed to represent. the cost of the plan for both tax and
general accounting purposes. That is, pension costs have, been accounted
for on a cash basis rather than the acc?~u.ai basis associatedi with most.
business transactions.. In November 1966, the Accounting Principles
Board of the American Institute of Certified Public Accountants, fol-
lowing several years of intensive study and discussion of the account-
PAGENO="0211"
OLD AGE INCOME ASSURANCE-PART V 203
ing aspects of pension plans, took the position that beginning Janu-
ary 1, 1967, pension costs should be accounted for on an accrual basis.5
The costs are to be computed in accordance with one of the generally
acceptable actuarial cost methods (described in Appendix A of Opinion
No. 8), using reasonable actuarial assumptions and reflecting the
other conclusions of the Opinion.6 The Opinion stated that the annual
provision for pension cost should not be less than the total of normal
cost and an amount equivalent to interest on any unfunded prior serv-
ice cost, plus, if necessary, an additional sum calculated in such man-
ner as to insure the full accrual over a 20-year period of the costs
associated with vested benefits.7 The annual provision for pension
cost should not be greater than the total of (1) normal cost, (2) 10
percent of the initial past service cost (until fully amortized), (3) 10
percent of the amounts of any increase in prior service cost arising out
of an amendment of t.he plan (until fully amortized), and (4) interest
equivalents on the difference between pension charges and amounts
funded.8 As intimated by item (4) above, the appropriate pension
charges are to be made whether or not corresponding payments are
made to the funding agency (bank or insurance company). If contri-
butions to the plan are less than the amounts charged to expense,
whether attributable to normal cost or prior service costs, the differ-
ence is to be shown on the liability side of the balance sheet and de-
scribed in a manner to make clear that it does not constitute a legal
obligation. An excess of contributions over expense charges are to be
reflected on the asset side as a deferred charge.
The employer's income tax deductions are still based on his contri-
butions to the pension plan rather than expense charges per se.
The new accounting rules apply to all types of pension plans-
defined contribution as well as defined benefit plans and unfunded as
well as funded plans. It is especially significant that future pension
cost accruals under an unfunded plan must be reflected in the employ-
er's balance sheet, which may induce the employer to adopt the practice
of funding his pension obligations. The accountants' recommendation
that unfunded pension charges under any type of plan be reflected in
the balance sheet is likely to cause all employers to fund at a rate at
least equal to that at which the pension costs accrue. It may, of course,
lead to the practice of computing cost accruals at the lowest acceptable
level, thus inhibiting funding. In any event, it should be recognized
that in the typical case the employer's legal liability, if the plan should
terminate, is limited to the amounts already contributed, even though
the balance sheet may show an excess of expense charges over funding
payments.
HAZARDS TO BENEFIT FULFILLMENT
The primary source of security for accrued pension benefits is a
fund of assets, including life insurance and annuity contracts, ir-
revocably placed beyond the control of the employer and committed
See Opinion No. 8, Accounting for the Cost of Pension Plans, pars. 16, 17, and 15. This
Opinion supersedes Accounting Research Bulletin No. 43, ch. 13, sec. A, Compensation:
Pension Plans-Annuity Costs Based on Past Services and Accounting Research Bulletin
No. 47, Accounting for Costs of Pension Plans.
Ibid., par. 24.
1 Ibid., par. 17. However, the additional sum need not exceed the amount that would be
necessary to amortize the Initial past service liability over a 40-year period.
Ibid., par. 17.
PAGENO="0212"
204 OLD AGE INCOME ASSURANCE-PART V
to the payment of benefits (and attendant expenses) in accordance
with the terms of the plan. Thus, broadly speaking, any factor or cir-
cumstance that interferes with the accumulation of plan assets at ap-
proximately the same pace as the benefits accrue is a threat to the ulti-
mate fulfillment of benefit expectations. A deficiency in plan assets
as of any given time may be the result of inadequate past contributions
or the shrinkage of asset accumulations through unfortunate invest-
ments, decline in market value of sound investments, or, in rare cases,
misappropriation.
Past contributions may be insufficient to support the current struc-
ture of accrued benefits for a number of reasons. First of all, the pro-
jected cost of benefits and expenses may have been inaccurate. Cost
projections reflect assumptions as to mortality among the participat-
ing employees and their beneficiaries; investment earnings on plan
assets; employee withdrawals without vested benefits, the pattern of
retirement; and the expenses of operation. If benefits are based on
compensation, especially earnings during the years immediately pre-
ceding retirement, future compensation levels must be predicted. These
cost projections are normally prepa.red by actuaries after consulta-
tion with the employer and other int.erested parties. Wide areas of
judgment exist with respe.ct to most of these cost factors and actuaries
of unquestioned professional skill, judgment, and integrity can come
up with sharply different answers even when working with the same
basic data.
The rate at which the total estimated costs are assumed to accrue is
determined by the actuarial cost method employed by the actuary.
There are several basic actuarial cost methods, with numerous varia-
tions and refinements. One family of methods assumes that the cost
of the plan accrues at precisely the same rate as the benefits accrue,
the cost per $1 of benefit with respect to any particular individual
increasing year by year as the employee nears retirement. The other
family of methods projects the total cost of benefits for the covered
group, individually or collectively, and assumes that the cost is spread
evenly over the total employment period or the remaining years to
retirement. Under each of these two basic approaches, the cost of
benefits associated with years of service prior to inception of the plan
(referred to hereafter as supplemental cost) may be separately cal-
culated and accrued at a rate different from that applica.ble to the
normal cost. These cost methods have no influence on the ultimate true
cost of the plan (other than their impact on investment earnings),
but they have a significant effect on the rate at which pension costs
are charged and presumably funded.
Any of the commonly accepted actuarial cost methods, when em-
ployed as a guide to funding policy, can, under the right conditions,
and in time, lead to the accumulation of assets equal to or in excess of
the actuarial value of accrued benefits. Most employers do pattern
their funding contributions after one of the commonly recognized ac-
tuarial cost methods and undertake as a minimum to meet the normal
cost of the plan plus interest on the initial supplemental cost, if any.
There is no legal compulsion to fund the supplemental cost of the plan,
and if the cost method in use produces a supplemental cost, the em-
ployer may decide not to fund it. In that event, the accumulated as-
PAGENO="0213"
OLD AGE INCOME ASSURANCE-PART V 205
sets as of any given time would under many circumstances be less than
the actuarial value of accrued benefits.9
Even if the employer should undertake to fund the supplemental
cost (usually associated with past service benefits or retrospective bene-
fit increases), he will usually spread the funding over an extended
period of time, possibly as long as 40 years. I-Ic is unlikely to fund
more than one-tenth of the supplemental cost in any one year, since he
would be unable to deduct currently any greater sum for income tax
purposes. Retrospective benefits increases and other plan liberaliza-
tions add to the supplemental cost and usually prolong the period
needed to achieve a fully funded status. In fact, many plans will never
achieve a fully funded condition purely because of periodic plan lib-
eralizations. Clearly, as long as any portion of the cost of accrued
benefits under the plan remains unfunded, a termination of the plan
will result in the loss of benefits by some participants.
The achievement of a fully funded status does not assure the pay-
ment of all benefits in full, even if the plan terminates while that con-
dition exists. The cost estimates of many plans take into account antic-
ipated withdrawals among the employee group and the funding pay-
rnents are reduced by the amount of estimated nonvested withdrawals.
Allowance may also be made for retirements expected to take place
beyond the normal retirement age. However, when a plan termi-
nates, these cost reducing factors become inoperative and the benefits
that have been credited to that point in time may well exceed the assets
that have been accumulated.
The contributions to a pension plan are invested by the funding
agency and exposed to the hazards of the capital market. If the fund-
ing agency is a life insurance company, the funds must be invested in
accordance with the standards set forth in applicable investment stat-
utes. These statutes severely limit the amount of common stock that
can be held in the general investment account of an insurer but the
limitations are not applicable to special (or separate) accounts set up
under group deposit administration annuity contracts. Banks and in-
dividual trustees are nominally subj ect to fiduciary investment stat-
utes, which can be very restrictive, but the instrument creating the
trust can and generally does grant broader investment powers to the
trustee than those provided by applicable State law, especially in
respect to common stock holdings. At the present time the Federal
Government imposes no standards of investment conduct, apart from
the prohibition against certain types of transactions that would im-
properly benefit the employer. Legislation now pending before Con-
gress would impose additional standards. Pension funds administered
through the general investment accounts of life insurance companies
are invested predominantly in high-grade corporate securities and real
estate mortgages, while the funds held in separate accounts are invested
largely, if not wholly, in common stocks. Funds held by trustees are
invested mainly in high-grade corporate securities and common stocks,
the proportion of common stocks in some portfolios being well over 50
° When the participant group is immature and growing (or even stationary), and a
projected benefit cost method is used, the funding of normal cost plus interest on the
unfunded actuarial liability, augmented with actuarial gains, can lead to the accumulation
of assets equal to or In excess of the value of all accrued benefits, including those attrib-
utable to past service and retrospective benefit increases.
PAGENO="0214"
206 OLD AGE INCOME ASSURANCE-PART V
percent, especially at market value. Apart from the inherent risks as-
sociated with common stock holdings and the unlikely possibility of
forced liquidation of assets, pension plan assets, in general, appear
to be relatively well protected against investment, risks.
Accumulated pension assets ma.y also be lost through the fraudulent
or dishonest behavior of plan administrators. Fortunately, there has
been little evidence of this type of behavior thus far, and loss of assets
from this source has been infinitesimal. Moreover, a number of bills
now before Congress would impose a fiduciary status on all persons
exercising control over pension assets and would vest in designated
Federal agencies certain investigative and enforcement powers de-
signed to protect pension funds against mismanagement.
If the employer pursues a policy of paying the pension plan bene-
fits directly to the eligible claimants, without using a funding agency,
the security of the benefits rests entirely on t.he ability and willingness
of the employer to continue his support of the program. The fulfill-
ment of benefit expectations is subject to all the economic vicissitudes
that beset the business enterprise itself. The participants can look to
no fund of assets insulated from the hazards of the business. Under
the action of the American Institute of C.P.A., to which earlier refer-
ence wa.s made, the employer must set up a. balance sheet reserve equal
to the assumed value of the benefits accruing hereafter. Irrespective
of whether the proper amounts are credited to the reserve, the assets
offsetting the account are not earmarked for the exclusive benefit of
the plan participants and their beneficiaries. If the employer should
decide to discontinue the payment of retirement benefits, the pen-
sioners and active employees would have no claim against the assets
representing the reserves, unless a court were to hold that in the light
of all the circumstances the employer had in effect guaranteed the
payment of pension benefits. In that event, the pension claims would
not necessarily be limited to the value of the pension reserve. On the
other hand, if the employer should become insolvent or bankrupt, the
accrued pension rights of employees a.nd participants would be given
no preference in the allocation of assets and might not be recognized
at all.
APPLICABILITY OF INSURANCE CONCEPTS TO A PENSION GUARANTY FUND
A mechanism to assure the payment of accrued benefits under pri-
irate pension plans is, in essence, an insurance arrangement, and its
feasibility must be tested against the criteria of an insurable hazard.
LARGE NUMBER OF HOMOGENEOUS RISKS
A soi.rnd insurance program must encompass a large number of
risks in order that losses may be predicted in accordance. with the
laws of probability, with reasonable assurance that actual results will
conform closely to predicted experience. The risks should be homo-
geneous in order that an equitable rat.ing structure can be developed.
In other words, the losses should be spread among the participating
risks in an equitable manner. If the total body of risks is not suffi-
ciently homogeneous to permit the charging of a. uniform premium
rate, it should be possible to classify the risks into subgroups suffi-
PAGENO="0215"
OLD AGE INCOME ASSURANCE-PART V 207
ciently large and homogeneous to permit the development of sound
rating procedures. A corollary to this concept is that the insurance
program must attract a representative cross section of the risks ex-
posed to the ha.zard, since otherwise the premium rate will be so high
as to discourage voluntary participation by all except the worst risks.
There is obviously a sufficiently large number of pension plans
subject to the hazard of termination to satisfy this first criterion. The
risks are not homogeneous, however, in the sense that they are sub-
ject to approximately the same probability of termination. Unless the
premium rate is adjusted to the probability of termination-or is so
small as to be inconsequential-the better-than-average risks could
be expected to shun the arrangement. Even with an equitable and
realistic rating structure, pension plan administrators might find the
scheme unattractive, since it would involve an added element of cost.
It would appear that participation would have to be compulsory to
obtain a random selection of risks.
OBJECTIVE DETERMINATION OF OCCURRENCE AND AMOUNT OF LOSS
For an insurance arrangement to be feasible, it must be possible
to determine beyond reasonable doubt that the event insured against
has occurred, and the amount of loss sustained should be susceptible
to fairly precise determination.
This aspect of an insurable risk could prove to be very troublesome
for a pension guaranty fund. The basic difficulty would be to define the
event insured against. It is an open question as to whether all plan
terminations should be covered or only those originating in certain
causes. Some would argue, for example, that the guaranty should be
restricted to those plan terminations that occurred because the em-
ployer went out of business, while others would regard it as immaterial
that the employer continues to operate in one form or the other. Vary-
ing attitudes are taken toward mergers, cessation of operations in one
plant or locality, and so forth. Special problems exist with respect to
multiemployer plans. Questions would arise as to whether a discon-
tinuance of employer contributions is merely a suspension, discon-
tinuance, or termination, as those terms are defined in IRS regulations.
Presumably, a pension guaranty program could adopt its own defini-
tion (or definitions) of the risk insured against, as contrasted with
the views of the internal Revenue Service, but unless the insured event
were carefully delineated, complex problems of interpretation would
be involved in cleterminmg whether a articular event or transaction
fell within the prescribed limits.
The determination of the amount of loss would involve potential
difficulties. Plans that provide for a specific unit of benefit for each
year of credited service would present no difficulties if the annual
accrual were a fiat amount or based on current earnings. There would
be complications with respect to i~ians that base the benefit on the
employee's compensation during the years immediately preceding
retirement or provide a basic benefit geared to career average earnings,
subject to a rnin~rnum benefit related to final average salary. The mini-
mum benefit-which may be financed through an auxiliary (or side)
fund, a terminal funding arrangement, or on a curreimt disbursement
PAGENO="0216"
208 OLD AGE INCOME ASSTJRANCE-PART V
basis-may vest at a. different rate than the basic benefit and may in
fact be subject to its own set of eligibility requirements. Money pur-
chase plans, especially those involving split funding arrangements,
would require special consideration; as would plans that provide a
composite benefit, subject only to a minimum period of service. Ancil-
lary benefits, such as those payable for the employee's death or dis-
ability prior to retirement, would complicate matters, along with spe-
cial (nonactuaria.l) early retirement allowances, a.imuity options of
unequal actuarial value, and social security offsets. These problems,
formidable as they may appear, could be overcome by a precise defini-
tion of the benefits covered, excluding those that would unduly com-
plicate the administration of the program.
Once the aggregate value of covered benefits was determined at
point of plan termination, there would still be a question as to what
portion of the liability should be transferred to the guaranty fund and
when. This question will be considered in detail at a. later point.
RANDOMNESS OF LOSS
The occurrence of loss among the risks exposed to it should be ran-
dom and beyond the control of the person receiving reimbursement for
(or benefiting in any other wa.y from) the loss and any person or firm
who would otherwise be liable for the pa.3~ment. Since this condition is
seldom met. in practice, the objective is sometimes restated as follows:
where the occurrence of the. loss may be influenced by the actions of the
individuals purchasing the insurance or receiving the claim payment,
the benefits payable should be such that the occurrence of the loss is
always less advantageous financially than the nonoccurrence of the
loss.
In the absence of proper safeguards, this prerequisite would not be
satisfied by a pension guaranty fund. There are so many ways that
the fimd could be. abused that many persons believe the whole guaranty
idea. is unrealistic. If not. prohibited, an employer could increase bene-
fits retrospectively just before terminating his plan and let the guar-
anty fund make up the deficiency in plan assets. Or he might dis-
continue contributions to the plan in anticipation of formal termina-
tion. If an employer could terminate his plan at any time and for any
reason, with no one having any recourse against him for the unfunded
accrued liability, his willingness to continue his plan through periods
of adverse economic conditions could be seriously undermined. Most
of the opportunities for abuse could be eliminated or minimized by
appropriate limitations in the undertaking, but some of the restric-
tions might produce an arrangement quite different from that envis-
ioned by sponsors of the idea.
LOW PROBABILITY OF LOSS
The probability that any particular exposure-unit will incur a loss
during any given year should be relatively low. If losses occur with
high frequency, it will generally be more economical for the per-
son or firm exposed to the risk to budget for the losses outside of an
insurance mechanism which, of necessity, involves sonic adminis-
trative expense.
PAGENO="0217"
OLD AGE INCOME ASSURANCE-PART V 209
Available data would suggest that the probability of plan termina-
tion is fairly small, at least in periods of economic prosperity and
among plans that have been in operation for several years. Since
most plans that terminate are small, the potential losses to benefit
claimants is an insignificant percentage of the total exposure.
The latest and most comprehensive study of plan terminations was
carried out by the Bureau of Labor Statistics, in cooperation with
the Internal Revenue Service.'0 The study encompassed the years
1955-65. During that period, there were 4,259 pension plan termina-
tions, 30 percent of which were due to mergers. Almost half of the
terminations involved plans in effect for 5 years or less, and two-
thirds of the terminating plans covered fewer than 25 employees. An
average of 20,000 employees per year were affected by terminations,
about one-tenth of 1 percent of the number exposed to loss." In many
cases, no accrued benefits were lost and in other cases the losses were
less than total.
The rate of plan termination, and resulting benefit forfeitures,
would undoubtedly increase sharply during a period of depressed
business activity. Also, the very existence of a guaranty scheme might
increase the risk of plan termination. On balance, however, it would
appear that a pension guaranty fund would stand up fairly well
against this criterion.
SIGNIFICANCE OF LOSS
The loss suffered by the insured from the occurrence of the event
against which insurance is arranged, should be large enough to con-
stitute an economic burden. Otherwise, the expense of maintaining the
insurance mechanism and settling claims might exceed the value of the
loss payment. This principle is given effect in private insurance opera-
tions through the use of so-called deductibles, which cause the insured
to bear the first portion or layer of any loss.
The aggregate loss of accrued benefits occasioned by the termination
of a pension plan is likely to be sizable enough to satisfy this crite-
rion. The loss to some of the participants, however, may not be. This
suggests that the protection of a pension guaranty fund might prop-
erly be limited to individuals having some minimum amount of bene-
fit accruals or having participated in the plan for a specified period
of years.
ABSENCE OF CATASTROPHE HAZARD
tinder an ideal insurance arrangement, the hazard insured against
should not be capable of producing a catastrophic loss to the insuring
agency out of one event or occurrence. In reality, many hazards that
are regarded as readily insurable can, under certain circumstances,
produce losses in the catastrophic area. The problem has been overcome
through the use of exclusion clauses and reinsurance facilities.
It is conceivable that in a period of extended economic distress, pen-
sion plan terminations could confront a guaranty fund with claims
of such magnitude that the solvency of the program would be threat-
ened. At such a time, the financial condition of many firms might be
10 Emerson H. Beler, "Terminations of Pension Plans: 11 Years' Experience," Monthly
Labor Review, June 1967, pp. 26-30.
11 Ibid., p. 26.
PAGENO="0218"
210 OLD AGE INCOME ASSURANCE-PART V
so precarious that they could not absorb the additional burden of
support that would be required. The problems would be alleviated to
some extent by the fac.t that most of the claims against the guaranty
fund would represent deferred obligations, which would not have to
be fully offset by assets at any point in time. In an extreme emergency,
funds could be made available by the Federal Government. either in
the form of a loan, a direct subsidy, or indemnity payments under a
formal reinsurance scheme.
RELEVANT PRECEDENTS
Additional insights into the feasibility of a pension guaranty fund
can be gained by examining the essential elements of existing insur-
ance arrangements that fail in one or more important respects to
satisfy the conventional concepts of a sound insurance program. Some
of these programs are operated under governmental auspices, some
under private auspices, some under joint governmental-private aus-
pices. and one-a pension guaranty fund-under the aegis of a foreign
government.
GOVERNMENTAL PROGRAMS
The governmental program that is most frequently cited as a
precedent for a pension guaranty fund is the Federal Deposit Insur-
ance Corporation. This agency was established in 1933 to restore con-
fidence in the commercial banking system and to prevent another
debacle such as that which occurred between 1929-33 and wiped out
the savings of millions of American families. Participation in the
program is compulsory for banks that belong to t.he Federal Reserve
System but is optional for all other banks. The original act provided
only $2,500 of insurance per covered deposit account, but the maximum
protection was later increased to $5,000, $10,000, and t.hen $15,000,
where it now stands. The program is supported by annual assessments
on member banks at the statutory rate of one-twelfth of 1 percent of
aggregate deposit liabilities (not just the amounts protected by insur-
ance) but because of credits for favorable loss and expense experience,
the effective assessment rate in recent years has been about one thirty-
first of 1 percent of average deposit liabilities. The assessment rate is
the same for all banks, irrespective of age, size, or quality of manage-
ment. This feature of the system was severely criticized in the begin-
ning on the grounds t.ha.t strong, conservatively managed banks would
be required to subsidize the weaker, less conservative banks. The argu-
ment is no longer heard today, probably because the assessment rate is
so low. There were also allegations that deposit insurance would en-
courage reckless lending and investment practices on the part of
insured banks, but. these fears proved to be unfounded. While the
system has never been tested by a major depression, there is general
agreement that. deposit insurance has become a~ permanent feature of
this country's financial structure. It is such an essential element of
economic stability that the system would be maintained even if a tem-
porarv governmental subvention should become necessary.
While there are certain similarities between deposit insurance and
the proposed guaranty fund for pension benefits, there are dissimilar-
ities that largely invalidate analogies between the two schemes. From
PAGENO="0219"
OLD AGE INCOME ASSIJRANCE-PART v 211
the standpoint of any one individual, the FDIC insures against loss
of a known number of dollars that are currently available to the deposi-
tor. The pension guaranty fund would insure against loss of future
dollars the number of which might not be presently determinable and
the right to which would be contingent on survival to retirement and
possibly on continuation of the current employment relationship to
retirement. More important, the assets offsetting the deposit liabili-
ties of a bank are in the possession of the bank (or else the bank would
be insolvent), whereas the assets needed to liquidate the accrued bene-
fits of a pension plan usually are not in the hands of the funding
agency, at least in sufficient quantity. Thus, it may be said that the
FDIC insures against loss of assets already in existence, while the
guaranty fund would, in effect, provide protection against the loss of
assets that never materialized (from the standpoint of the funding
agency). The FDIC insures a reality, while the guaranty fund would
underwrite a declaration of intention. There are types of insurance
that provide protection against failure to perform (reference will be
made to them later), but they are not a component of the bank deposit
insurance program. To the extent that a guaranty fund would make
good on benefit defaults arising out of investment losses, there would
be a parallel to the FDIC program.12
The Federal Housing Administration administers 15 trust funds
for the insuring of various types of mortgages. The oldest and largest
of these is the mutual mortgage insurance fund, which insures resi-
dential mortgages. Detailed eligibility rules have been promulgated
by the FHA and each application for insurance is carefully screened
to determine whether it meets the minimum standards of acceptability.
Among the factors taken into account are the applicant's income,
assets, character, and motivation. All approved applicants pay the
same premium, one-half of 1 percent of the mortgage (with the re-
maining balance being recomputed each year), but there is a pro-
vision for a refund of excess premiums upon final liquidation of the
mortgage, the dividends varying with the risk classification in which
the mortgage was originally placed. Defaults have been low and net
claim payments well within the premium income, but, as with the
FDIC, the system has not been subjected to the exigencies of a severe
and prolonged depression.
There is a basic similarity between mortgage insurance and a
pension guaranty fund in that both involve the ability and willing-
ness of an individual or firm to make future payments. A major
difference is that with mortgage insurance the obligation to pay is
secured by an asset (the home, for example), while the employer's
obligation to make payments to a pension plan is neither secured nor
legally enforceable (except pursuant to the terms of a collective
bargaining agreement). Sensitivity to economic conditions is another
common attribute, but conclusive evidence on this point lies in the
future.
Many States operate guaranty funds to insure payment of work-
men's compensation benefits when the insurer or the employer, as a
self-insurer, becomes insOlvent. New York has a guaranty fund for
12 The foregoing discussion is equally applicable to the Federal Savings and Loan Insur-
ance Corporation.
PAGENO="0220"
212 OLD AGE INCOME ASSURANCE-PART V
claims under automobile insurance policies and another for life in-
surance policies. In all these cases, the State fund is protecting benefit
rights to the limit, of its resources against the broad economic hazard
of insolvency-which would obviously be involved in a program to
insure pension benefits-and even more specifically, the insolvency
of insurance companies. Legislation to establish a Federal guarantee
fund to ensure payment of claims against. insolvent, automobile in-
surers, especially the so-called high-risk insurers, is pending in
Congress.
Somewhat farther afield, but still relevant, are the unsatisfied judg-
n-ient funds in Maryland, Michigan, North Dakota, New Jersey, and
New York (and all the Canadian provinces except Saskatchewan),
established for the purpose of enabling the innocent victims of auto-
mobile accidents to collect on their adjudicated claims or judgment.s
against financially irresponsible motorists.
PRIVATE PROGRAMS
There are two types of commercially written insurance that throw
some light on the feasibility of a pension guaranty fund. The first
is credit insurance, which is written by a number of nonlife insurance
companies. TJncler this type of insurance, the insurer promises to
indemnify a business firm for bad debt losses in excess of those con-
sidered (and agreed upon) to be normal for the firm. The level of
losses deemed to be normal for the business, and hence not idemnifiable,
is determined by the experience of the firm over a period of past years.
The premium charged for the coverage reflects the Dun & Bradstreet
credit ratings of the insured's customers. The risk of loss to the insurer
is greatly influenced by the economic climate, a characteristic shared
with the proPosed pension guaranty fund. Another common charac-
teristic is that both the ability and willingness of business firms to
meet obligations are involved, although there are legal sanctions asso-
ciated with credit insurance that might not be operative with a pension
guaranty.
The other type of commercial insurance that. should be considered
is that kind of coverage that. goes under the name of su.retyshAp. There
are various forms of suretyship coverage, including losses against in-
fidelity of employees, but the most rele\~ant. one for presei~t 1~~r1oses is
that associa.t.ed with contract or performance bonds. The basic purpose
of a. performance bond is to indenmify one party for economic loss sus-
tamed by failure of another party to carry out an undertaking in ac-
cordance with the terms of an agreement between the two parties. A
common example of such an undertaking would be the construction of
a building in accordance with certain specifications. Another would
be an agreement by a parent company to guarantee. payment of in-
terest and principal of a bond issue of a subsidiary. The bond is usually
purchased by the party that must perform. If the responsible part.y
does not perfonn as required and the insurer (frequently called the
surety) has to indenmify the aggrieved party, the insurer is then en-
titled to move aga.inst t.he defaulting party in an attempt to recover
the amount of the loss payment. In other words, this is a. type of insur-
PAGENO="0221"
OLD AGE INCOME ASSURANCE-PART V 213
ance under which the party that procures the insurance and pays the
premium is ultimately responsible for the payment of any claims that
may arise. It has been suggested by some that this principle should be
incorporated into any program that might be established for the guar-
anty of pension benefits. This would make the employer ultimately
responsible (to the extent of his corporate asset) for any benefits paid
in respect of his plan by the guaranty fund.
Another precedent for giving the guaranty fund a right of recovery
from the employer is the doctrine of subrogation that is applicable to
all forms of property and liability insurance. Under this doctrine,
which is supported by common law as well as contract language, if an
insured loss is caused by the tortfeasance or wrongful action of a third
party, the insurance company has a right to seek recovery from the
tortfeasor. The citing of this legal principle is not intended to imply
that an employer who fails to fund or otherwise meet his pension obli-
gations is a tortfeasor in the legal sense, but. in granting more guar-
anteed benefits than he has funded an employer would cause loss to the
insuring agency and under certain circumstances it might be argued
that lie has wilfully and irresponsibly caused loss to the guaranty fund.
PLANS JOINTLY ADMINISTERED BY GOVERNMENTAL AND PRIVATE
AGENCIES
There are a number of insurance programs that involve a partner-
ship of some type between a Federal agency and private insurance
agencies. In some of these programs, for example, Federal Employees
Group Life Insurance and the insurance provided under the Federal
Employees Health Benefits Act, the private agencies are the sole risk-
bearers, the Government playing a strictly administrative role. In
other programs, for example, Medicare and the provision of health
insurance benefits for servicemen's dependents, the private agencies
furnish only fiscal and claims services, with the Government assuming
the entire risk. In other cases, the Federal Government and private
insurance agencies have entered into a joint underwriting venture under
which the Government assumes that portion of the total risk con-
sidered to be uninsurable by private agencies. One case in point is
export credit insurance in connection with which private insurers
assume the normal business risks and the Federal Government,
through a reinsurance arrangement, assumes the political risks. An-
other example is the Servicemen's Group Life Insurance program
under which the participating life insurers assume the normal mortal-
ity risks and the Federal Government absorbs the risks associated with
military service. Still another arrangement that could be used would
be for the private insurers to assume the first or primary portion of the
risk, with the Government serving as the residual riskbearer through
reinsurance or some other device. This is the approach embodied in
the proposed flood insurance program now being considered by Con-
gress and representatives of the insurance industry. This is an attempt
to deal with a hazard that, because of the threat of catastrophic losses,
has heretofore been considered uninsurable for all practicable purposes.
Under the proposed program, which initially would be operative
only in certain designated areas and would cover only one- to four-
PAGENO="0222"
214 OLD AGE INCOME ASSURANCE-PART V
family residential properties (and their contents), insurance against
flood damage would be made available through private insurers in
amounts up to $15,000 per dwelling unit, subject to an aggregate lia-
bility of $30,000 on structures with more than one unit. Contents could
be insured for an additional $5,000 per dwelling unit, subject to an
appropriate deductible. The program would be administered by the
Department of Housing and Urban Development which, after analysis
of pertinent data and consultation with appropriate agencies, would
promulgate two scales of premium rates. The first scale, called "risk
premiums" and computed on the basis of actuarial estimates of the risk
and expenses involved, would be pitched at a level believed to be ade-
quate to cover all valid claims plus allowances for expenses, contingen-
cies, and profits. The second scale would set forth the rates that would
actually be charged for the coverage, and these "chargeable" premium
rates would reflect considera.tion of many factors, including land use
controls, flood proofing, and flood forecasting. The Secretary of HUD
would be authorized to prescribe chargeable rates lower than those
computed on the above bases in order to encourage the purchase of
flood insurance. In low-risk areas, the chargeable premium rates for
existing properties would be the same as, or only slightly less than, the
full-cost rates, but in high-risk areas the chargeable premiums would
always be lower, the disparity growing greater with the increase in
risk.'3 The chargeable premium for insurance on a structure that was
started or substantially improved within 60 days after flood insurance
became available in the area would be identical with the risk premium
(full cost) rate.
The protection would be provided through a pool of private in-
surers, to which all but the smallest companies could belong. The pool
would in all cases receive the full-risk premium, with the Federal
Government, through the National Flood Insurance Fund, paying the
difference, if any, between the risk premium and the premium paid
by the policyholder. The National Flood Insurance Fund would enter
into an excess loss type of reinsurance agreement with the insurance
pool, in accordance with which the Fund would bear all losses in excess
of a stipulated amount per year.'4 The pooi would pay a reinsurance
premium on a basis as yet undetermined.'5
The aspects of this proposed program which might be considered by
those studying the feasibility of a peusion guaranty fund are (1) close
cooperation between the Government and the insurance industry in
the planning stages of the program; (2) the concept of a premium rate
lower than the actuarial value of the risk, with the difference being
borne by the Federal Government in the form of premium equalization
payments; (3) assumption by the Government of all claims in excess of
a stipulated amount per year; and (4) emphasis on loss prevention.
~ Under the tentative scale of vet risk premiums currently being considered, the rates
would range from 2 cents to S5 per $100 of coverage. The maximum rate that would be
charged any policyholder (with respect to existing properties) would be 50 cents per $100,
plus an allowance for expenses, contingencies, and profits.
`~ It is proposed that the Government pay all claims in any year that exceed $50 million
or 35 percent of the risk premiums earned by the insurance pool, whichever is the lesser.
~ A proposal under consideration would link the reinsurance premium to the gross profit
of the insurance pool on the flood Insurance business. In other words, the pool would pay
to the National Flood Insurance Fund, as a reinsurance premium, the excess of premiums
earned over expenses and claims incurred, less a specified percentage for profit.
PAGENO="0223"
OLD AGE INCOME ASSURANCE-PART v 215
SWEDISH PENSION GUARANTY FUND `~
A rather complex but apparently workable system for assuring the
payment of pension benefits has been in operation in Sweden, smce
1.960. While the institutional environment in which this system func-
tions is quite different from that in the United States, lessons can un-
doubtedly be learned from the Swedish experience.
The system was an outgrowth of a series of collective bargaining
agreements between employer and salaried employee associations,
reaching into every branch of industry (and certain trade, service, and
agricultural undertakings), pursuant to which certain pension benefits
were to be provided as a supplement to the national old-age insurance
program. The agreements gave the employers the option of purchasing
the benefits from the Swedish insurance company (the Swedish Staff
Pensioning Society) established by employers about 40 years ago for
the sole purpose of underwriting pension benefits or of assuming legal
responsibility for the direct payment of the benefits. If the employer
elects the first course of action, he fully discharges his legal responsi-
bility by the payment of the necessary premiums and he does not get
involved in the pension guaranty mechanism. If, on the other hand,
the employer prefers to keep the equivalent of the premium payments
in his own firm, he must set up an internal pension fund and register it
with a central registration agency, called the Pension Registration
Institute (PRI), which has other functions and indeed plays a very
active role in the whole pension process by recording benefit accruals,
informing employees of their rights, performing actuarial valuations,
and paying benefits underwritten by the aforementioned pension in-
surance company (SPP). The employer's internal fund, referred to as
the "PRI fund," must at all times hold assets equal to the actuarial
value of all accrued benefits-as calculated and certified annually by
the PRI-but the only assets placed in the funds are unsecured prom-
issory notes of the employer. These notes would be enforceable against
the general assets of the employer in the event of insolvency or bank-
ruptcy but without any special creditors' preference, except as noted
below.
Since the alternative methods of providing the bargained benefits
are supposed to be endowed with equal degrees of security-and to be
equal in all other respects-and since the worth of the employer's
promissory notes is completely dependent upon the financial well-being
of the firm, it was necessary to create another agency to underwrite the
credit of the employer. This organization, a mutual credit insurance
company formed by the employer associations and known as FPG,
assumes the pension obligation of an insured employer who defaults on
his pension payments and then discharges its responsibilities by the
purchase of annuities in the appropriate forms and amounts from
SPP, the pension insurance company. It seeks reimbursement for its
premium outlays by taking over the promissory notes in the employer's
internal fund and competing with other creditors in the resulting
bankruptcy proceedings, with no special preference other than with
16 For a more detailed description of this system see G. M. Ericsson, "PRI-A Self-
Insured Pension Scheme In Sweden", Transactions of the 17th International Congress of
Actuaries, London-Edinburgh, 1964, pp. 554-567 and Sven Hyddn, "A New Approach to
Financing Private Pension Schemes in Sweden", International Review of Actuarial and
Statistical Problems of Social Security, No. 9.
PAGENO="0224"
216 OLD AGE INCOME ASSURANCE-PART V
respect to benefit rights that accrued during a short period preceding
the adjudication of bankruptcy.
FPG makes credit insurance available only to joint-stock companies
and economic associations (roughly equivalent to corporations) and
only to such enterprises that have been in business for at least 3 years
and employ a minimum of five salaried workers. Since there is an
average of three and one-half manual workers for each salaried em-
ployee, the general effect of this stipulation is to limit the coverage to
firms employing a minimum of 20 persons. Applications are carefully
screened since the insurance is written for a 5-year term and if not
renewed remains in effect for a systematically declining amount of
pension liability for an additional 15 years (the time given the em-
ployer to convert his pay-as-you-go scheme to a fully insured arrange-
ment). The financial position, economic prospects, and quality of
management of the firm are carefully considered, the risk appraisal
being at least as rigorous as the standards applied by banks in making
long-term loans. Firms that cannot qualify for credit insurance, either
because they are too small or too unstable, must purchase the required
benefits from the pension insurance company (SPP).
FPG charges a. risk premium of three-tenths of 1 percent of the
accrued pension liability, plus a. nominal loading for administrative
expenses. It has the authority, not invoked as yet, to levy cumulative
assessments of up to 3 percent of the accrued pension liability over
successive 5-year periods. This annual premium rate was adopted
after a study of bankruptcies occurring among Swedish corporate
enterprises during the period 1929_58.17 The investigation proceeded
on the assumption that the present scheme of supplementary pension
benefits had been inaugurated on January 1, 1929. The aim was to
discover the amount of moneys that would have had to be set aside
annually to cover the accrued pension obligations of the firms that
went into bankruptcy, on the severe and unrealistic assumption that
in no case would there have been any corporate assets available for
the satisfactiomi of pension claims.
The study revealed that a. premium slightly less than two-tenths
of 1 percent of the pension debt outstanding would ha.ve covered the
aggregate losses occurring during the period. Only in the years 1933
and 1934 would the credit insurance company have been unable to
meet all claims out of accumulated reserves, the deficit being erased
by 1937. With the exception of the year 1957, the insured firms would
have paid larger premiums in each of the years from 1935 to 1958, in-
clusive, than would have been required to pay the claims arising dur-
ing those years. The premium rate was set at three-tenths of 1 percent
in order to permit the building up of reserves that were considered
to be a substitute for a. large initial guaranty capital fund.
As of the end of 1966. about. 1,900 Swedish corporate employers,
comprising about 15 percent of the total, had elected to utilize internal
funding for the salaried eni~loyees' supplemental benefits, with the
concomitant use of the credit insurance mechanism. Approximately
200,000 employees and 9,000 pensioners were covered by this arrange-
11 The Investigation was limited to enterprises that belonged at some time during the
period to the Swedish Employers' Confederation (SAP), which includes within its mem-
bership practically all firms in the manufacturing industry and the majority of those in
the building and road transport Industries. Bankruptcy was construed to be the equivalent
of failure to pay the required contributions to SAP. In other words, any company that
failed to pay its dues to the SAP was considered to be bankrupt.
PAGENO="0225"
OLD AGE INCOME ASSURANCE-PART V 217
ment, accounting for two-thirds of the salaried employees of firms
engaged in manufacturing and commerce. The internal `ftmds of these
employers were holding $452 million in employer promissory notes,
with the volume expected to increase by $100 million annually over
the next; few years. FPG- was holding reserves of $3,450,000, which
constituted 0.76 percent of its liabilities, as measured by the employer
promissory notes. The company's objective is to accumulate a reserve
of 2 pe.rcent of its liabilities by the year 1980. Through June 30, 1967,
the company has ha.d to make good on its guaranty in only six cases
(three arising during the first 6 months of this year), involving
slightly more than $200,000 ($180,000 in 1967) in "claim" payments.
The company has terminated the insurance coverage of nine firms,
advancing $320,000 in loans to these employers for the purchase of
insurance from .SPP.'8
This credit insurance arrangement bears a surface similarity to
the proposed guaranty of pension obligations in this country in that
the basic risk is the insolvency of the employer. There is a funda-
mental difference, howeve.r, in that under the Swedish arrangement
all accrued benefits are fully offset at all times by legally enforceable
promissory notes of the employer,'9 and the only risk involved is the
probability that because of the emp]oyer's bankruptcy the notes would
not be paid off in full. The Swedish scheme limits the coverage to
those employers who can qualify, forcing the remaining employers
to purchase benefits from the pension insurance company. Pension
plans in this country are too diverse to be adapted to the Swedish
scheme, but certain elements of the scheme such as prescribed levels
of funding, assumption by employers of legal liability for the pay-
ment of pension benefits, and limiting the credit guaranty to those
that earn it, deserve serious study.
ISSUES
Many issues would have to be resolved if a pension guaranty fund
were to be established in the United States.20 This section analyzes
the issues and points to the various courses of action, without attempt-
ing to reach final solutions.
18 The foregoing operating statistics were made available by Mr. Sven 0. Hyddn, manag-
ing director, Pension Guarantee Mutual Insurance (FPG) in a letter to the author dated
July 21, 1967.
~ In Sweden all pension benefits are funded on a level cost basis arid the concept of an
unfunded pension liability (i.e., a supplemental cost) is not recognized.
10 This presupposes an earlier decision to use a guaranty fund mechanism rather than a
suretyship arrangement such as that employed in Sweden. In theory the objectives of the
current proposals for a pension guaranty could be achieved by requiring employers to
undergird their pension promises by the purchase of a performance bond from a surety
company. In the event of a default on pension obligations, the surety company would be
expected to make good on the benefit promises through the purchase of annuities from life
insurance companies, with recourse against the employer for its net outlays. Among other
advantages, this approach would permit the charging of a premium rate commensurate
with the individual risk involved. Some supervisory agency would have to determine that
the mandatory performance bond was purchased and maintained in force and that em-
ployee rights were protected in the event of plan termination. There would also have to be
some mechanism, similar to assigned risks pools in automobile and workmen's compensa-
tion insurance, to provide coverage to employers who could not purchase the required
insurance through regular channels. Since purchase of insurance would be mandatory,
the rates charged by the surety companies would have to come under public surveillance.
Despite the flexibility of this approach, its feasibility would be completely dependent upon
the response of a group of private insurers to a new and hazardous type of coverage. Of
course. a new surety company could be established for the sole purpose of writing this
coverage, but it is doubtful that such a mechanism would offer any real advantages over
the type of guaranty fund described hereafter in this paper.
83-200-67-pt. 5-i5
PAGENO="0226"
218 OLD AGE INCOME ASSURANCE-PART V
ADMINISTERING AGENCY
A pension guaranty fund, hereinafter referred to as the PGF or the
guarantor, could be established and operated under the auspices of
a Federal agency, a private agency, or a. combination G-overnment-nri-
vate instrumentality. The choice would depend in part on political
philosophies and in part on the financial mechanism envisaged. Leg-
islation proposed thus far has contemplated administration by a Fed-
eral agency which would be feasible under any set of circumstances and
would be especially appropriate if the financing scheme should embody
the assessment principle, with a minimum accumulation of assets,
and if the Govermnent were prepared to assume the greater part of
the risk involved.
Administration by a central private agency, specially created for
the purpose with representation from employers, labor unions, banks,
insurance companies, and other interested parties, would appear to
be equally feasible. Private control over t.he guaranty mechanism
might make the whole idea more acceptable to employers (who would
be expected to bear all or a substantial part of the cost) and it would
be particularly desirable if a major investment function should be
envisioned.
Another possible approach would be to utilize a central agency,
either governmental or private in nature, for the collection of pre-
miums, adjudication of plan terminations, and other ministerial func-
tions, with the guaranteed benefits being underwritten, for a con-
sideration, by life insurance companies on their own account or as
members of a pool. The risks could be assigned to individual companies
on a basis similar to that employed with Federal Employees Group
Life Insurance, the Federal Employees Health Benefits Act, or Serv-
icemen's Group Life Insurance. The pool arrangement could be pat-
terned after that being considered for the flood insurance program.
The primary difference between this approach and the second one men-
tioned above would be that in one case the central agency would re-
tain the risks taken over from terminated pla.ns, acting as an insurer
in the process, whereas in the other case the risks would be transferred
to existing insurance companies on some equitable basis.
Any approach that would place upon private agencies the basic
risks inherent in plan terminations would probably have to embody
some mechanism for governmental reinsurance.
EVENT INS~ED AGAINST
The most difficult problem that would have to be confronted in the
establishment of a so~md and equitable system for assuring the pay-
ment of accrued pension benefits would be defining or articulating the
circumstances under which the protection of the system could be in-
yoked. One would naturally assume that the guaranty would be ap-
plicable only when the pension plan `has terminated under certain
prescribed conditions, but all bills on this subject now pending before
Congress clearly contemplate that the guaranty could be invoked by
certain groups of employees even though the plan continues in ex-
istence. For example, the Javits bill defines the insured event as "sub-
stantial cessation of one or more of the operations carried on by the
PAGENO="0227"
OLD AGE INCOME ASSIJRANCE-PART v 219
contributing employer in one or more facilities of such employer be-
fore funding of the * * * liabilities * * *"
This type of thinking appears to be a throwback to the concept of a
partial termination expressed in IRS regulations in language similar
to the above-quoted passage. The IRS was concerned that an employer,
m anticipation of a complete termination of his pension plan, might
engage in a massive discharge of employees-through individual lay-
offs or closing down of specific operations-thus increasing the share
of the plan assets that would be available for the officers, supervisors,
and highly compensated employees. To forestall such action, the IRS
introduced the concept of the partial plan termination, which has the
effect of vesting all accrued benefits of the employees involved. In a
real sense, the concept was introduced as a substitute for reasonable
vesting provisions in the plan at a time when vesting through plan
provisions was not as common as it is today.
The most baffling difficulties conjured up in connection with a pen-
sion guaranty scheme are concerned with the rights of individuals
who lose their jobs while the pension plan to which they belonged con-
tinues in operation. The job loss could result from the elimination of
specific positions or the closing down of an entire plant, division, or
subsidiary. The problem, if there is one, is lack of adequate vesting.
If the pension rights of the displaced employees were vested, they
would retain a claim to benefits enforceable in due course against the
assets of the plan. If the plan itself were eventually to terminate with
insufficient assets, unsatisfied claims would be met by the guaranty
fund. Thus, the defintion of the insured event should not be compli-
cated-perhaps to the point of utter uselessness-by an effort to meet
another problem that is not an inherent part of the guaranty scheme.
The vesting problem would be largely solved if legislative proposals
now being readied by the administration were to be adopted. The re-
maining discussion assumes that any pension guaranty would apply
only to benefit claims arising out of a plan that has terminated in its
entirety.
The definitional problem would be further simplified if it could be
assumed that a plan termination was always an incident to the final
dissolution of the sponsoring firm. This, of course, is not the case.
Most firms (more than 80 percent according to the BLS study of termi-
nations) continue in full operation, usually in their original form,
after terminating their pension plans. In some of these cases (about
one-fourth according to the BLS study), the discontinuance stems
from financial difficulties. In the other cases a variety of causes may be
at work. Some of the more common reasons, other than financial diffi-
culties, why a pension plan might be discontinued are (1) sale or
merger of the sponsoring firm, (2) transfer of the employees to an-
other pension plan, (3) substitution of a profit-sharing plan or other
form of deferred compensation for the pension plan, (4) closing of a
plant or other subdivision that had its own pension plan, and (5) de-
sire of `the sponsoring firm to protect or improve its profit position, and
(6) mutual agreement `by the parties to a collective-bargaining con-
tract. There are undoubtedly cases where the firm ostensibly goes out
of business only to reopen under another name or in another form.
There is serious doubt concerning the propriety-and feasibility-
of invoking a pension guaranty when the firm that created the pension
PAGENO="0228"
220 OLD AGE INCOME ASSURANCE-PART V
obligations continues to operate in one form or the other. It would
clearly be inappropriate to have a. guaranty ftrnd assume the un-
funded obligations of a plan terminated in order to transfer the em-
ployees into another pension plan or a profit-sharing plan, or pursuant
to an agreement between the parties to a collective-bargaining con-
tract, presumably to protect or increase cash wages. It would also
seem improper to permit a prosperous corporation to slough off unto
a guaranty fund the unfunded pension claims of persons who had been
employed in a plant or division (with its own pension plan) shut down
for presumably sound business reasons. This would have the effect of
increasing the labor costs chargeable to the remaining units of the
business, an unfortunate but unavoidable consequence. A firm that
purchases another should be expected to assume the pension obliga-
tions of the acquired firm, the purchase price reflecting the unfunded
liabilities. It is recognized that this requirement might force complete
dissolution of a firm or one of its subdivisions because of the unwill-
ingness of the prospective purchaser to assume the unfunded accrued
pension obligation of the firm to be acquired. Accrued pension benefits
should also be protected when two or more firms merge despite the
fact that the merger negotiations might be complicated thereby. The
guaranty would seem to be justified when the termination was moti-
vated by ffna.ncial difficulties verified by the administering agency.
Yet this would create an unconscionable distinction between partici-
pants in pension plans terminated because of financial exigences and
those in plans terminated for other reasons, unless the sponsoring firms
undertake to honor out of corporate assets the unfunded benefits ac-
crued to date of termination.
The whole matter would be greatly simplified-with some reduction
in the attractiveness of qualifying a plan under Treasury regula-
tions-if the guaranty scheme were established on the basis that the
sponsoring firm, or its successor, would have the primary legal respon-
sibility of meeting the cost of the benefits covered by the guaranty, the
PGF having only the residual liability. Then it would be possible to
define the insured event as the complete termination of a pension plan,
without reference to the circumstances surrounding the event. If the
sponsoring firm had gone out of business, or was in process of doing
so, its obligation to the plan (or the PGF) would be discharged to the
extent of available resources in a lump-sum payment. If the firm con-
tinued in operation, its obligation could be discharged over a number
of years in accordance with the pattern prescribed for the funding of
the supplemental liability. (See next section.) In the meantime the
PGF would stand ready to assume responsibility for any benefits not
ultimately funded by the employer. Unless such a feature can be made
part of the guaranty arrangement, it would probably be necessary to
define the insured event in terms that would limit the guaranty to plan
termination arising out of the final dissolution of the sponsoring firm
(or its successor), whether by ban1~uptcy, insolvency, or voluntary
wmchng up.
Special rules would have to be developed for multiemployer plans,
since among other distinguishing characteristics, they have an exis-
-tence apart from that of any particular employer belonging to the plan.
un some industries in which these plans are found, such as the building
PAGENO="0229"
OLD AGE INCOME ASSURANCE-PART V 221.
trades and the maritime industry, the employment relationship is very
tenuous, frequently being limited to one construction job, voyage, or
other undertaking. In other industries the employment relationship is
more conventional. In either of these settings, it would be desirable to
provide some measure of protection to the accrued pension benefits of
the employee participants irrespective of whether the current employer
or past employers continue to participate in the plan or continue, in
operation. The preferred arrangement would be for the plan to assume
the primary responsibility for the payment of the benefits that would
come under the guaranty, so long as the plan continues in operation,
with the PG-F providing the ultimate guaranty. There would have to
be safeguards to protect the plan against abuse, and the contribution
rate (or benefit levels) would have to reflect the anticipated cost of the
internal guaranty.
A precedent for this type of approach is found in the national indus-
trial group pension plan developed by the Industrial Union Depart-
inent of the AFL-CIO and jointly underwritten by a number of life
insurance companies. If an employer has participated in this plan for
more than 3 years and terminates his participation for reasons beyond
his control, the accrued benefits of his employees are assumed by the
plan on a scale determined by designated priority classes and the num-
ber of years during which the employer participated in the plan. After
only 3 years of employer participation, the benefits of all em~)lOyee5
aged 60 or over with the equivalent of 10 years of credited service (10
"service units") are assumed in full by the plan. The benefits of em-
ployees under age 60 but with a minimum of 10 years of service are
assumed by the plan on a scale graded upward from 30 percent with 3
years of employer participation in the plan to 100 percent after 10
years of employer participation. Employees aged 72 or over at the time
of the employer's withdrawal from the plan who are at least 5 years
beyond normal retirement age have their benefits assumed in full by the
plan even though the employer participated for less than 3 years. (Ap-
parently, the plan is prepared to assume ultimate responsibility for
this class of employees the moment the employer enters into a "par-
ticipation agreement" with the plan and fulfills its contributions
commitment.)
In the absence of an internal guaranty arrangement, the PG-F
could assume ultimate responsibility for the payn~eiit *of the covered
benefits of employees whose employer withdraws from a. multi-
employer plan after a minimum period of participation and for rea-
sons beyond his control. The employees would retain claims aainst. the
plan to the extent of the funds allocable to them, but the PG-F would
make good on any insufficiency of assets-despite the. fact that the
overall plan continues in operation. In other words, the employer's
withdrawa.l from the plan be treated as a. plan termination with
respect to his employees. The guaranty should not be invoked if the
employees' rights are preserved through membership in another plan
or in the same plan through subsequent employment with another firm
belonging to the plan. This approach would not be applicable to multi-
employer plans in industries characterized by temporary employment
relationships. In these cases, termination of the master plan would
have to be the contingency that invokes the guaranty.
PAGENO="0230"
222 OLD AGE INCOME ASSLRANCE-PART V
OBLIGATION OF THE GUARANTY FUN])
It goes without saying that a pension guaranty mechanism should
relate oniy to benefits accrued to date of plan termination, however that
event may be defined. Thus, the PGF should not undertake obligations
with respect to benefits that would have been earned had the plan re-
mained in operation and the employee continued in the service of the
employer until normal retirement age. Moreover, it does not follow
that all accrued benefits would be entitled to the protection of the pro-
gram. Various limitations, discussed later, might be necessary or de-
sirable. Thus, the following discussion should be understood to apply
only to the benefits that would be entitled to protection.
Broadly speakmg, the obligation of the guaranty fund could be ex-
pressed in one of two ways: (1) assurance of ultimate payment of the
benefits covered by the program or (2) completion of the employer's
funding commitment.
Assurance of Ultimate Pa~jment of Guarcnteed Beneflts.-The most
comprehensive approach would be for the PGF to assure ultimate pay-
ment of all covered benefits, irrespective of the amount, source, or
cause of the asset deficiency. The deficiency could be caused by in-
accurate estimates of cost, failure of the employer to undertake an
adequate funding program, lack of time for the completion of a real-
istic funding objective, or loss of asset values through realized or un-
realized capital losses. No minimum standards of funding would be
imposed a.nd, in fact, pay-as-you-go plans would be eligible to partic-
ipate. Risk classifications would be established on the basis of prob-
ability of plan termination, and the premium paid in respect of a
particular plan would be derived by multiplying the probability of
termination times the actuarial value of unfunded benefits. In some
cases, this combination of factors might produce a premium charge as
large as, or larger than, the ammal contributions under a realistic fund-
ing program. Losses incurred on liquidation of assets in connection
with plan termination, the subject of special provisions in the early
legislative proposals, would be blanketed in under this approach inso-
far as they affect the payment of guaranteed benefits. Benefits not sub-
ject to the guaranty would not be protected against liquidating losses,
but, as will be noted later, it is possible to devise a claim settlement pro-
cedure that would eliminate or minimize the risk of such losses.
The advantages that would be associated with the approach are
(1) eligibility of all types of plans, including those financed on a pay-
as-you-go basis where the need for protection is greatest; (2) coverage
of all sources of asset deficiency, including liquidating losses as they
relate to covered benefits; (3) equitable allocation of costs through
use of many risk classifications; and (4) absence of mandated stand-
ards of funding. All of these advantages would have offsetting dis-
advantages from other points of view. The approach would lend itself
to great abuse because of its sweeping coverage; it might encourage
hinhlv speculative investment policies; it would be extremely difficult,
if not~impossible, to predict the probability of termination among vari-
ous classifications of plans-which would be especially important in
connection with unfunded plans; the need for a governmental subsidy
would be greater than under any other approach-because of the fore-
PAGENO="0231"
OLD AGE INCOME ASSURANCE-PART V 223
going factors plus absence of minimum funding standards; and finally
participation by pay-as-you-go plans (which might be encouraged at
the expense of funded plans) would have to be on a voluntary basis
unless they are to be brought under some degree of regulatory control,
as they have been in the United Kingdom.2'
Most of the major disadvantages of this approach could be over-
come by requiring the covered benefits to be funded in accordance with
minimum standards stipulated by law or administrative agency. These
standards would concern themselves with actuarial assumptions, ac-
tuarial cost methods, and the period of time allowed for the full fund-
ing of all covered benefits. If the guaranty were to apply to all accrued
benefits, the employer might be expected to fund his normal cost cur-
rently and his supplemental cost, if any, over a maximum period of 25
or 30 years. If the guaranty were to apply only to a segment of the
accrued benefits, such as those that have vested, the standard might
speak only of the maximum period over which the covered benefits
were to be funded in full, possibly on the basis of specified mortality
and interest assumptions. Alternatively~ the standard might require
the current funding of the normal cost of the vested benefits and the
amortization of the supplemental cost of the vested benefits over a
specified maximum period. Another logical standard might prescribe
the funding of the normal costs of all accrued benefits for the plan
as a whole plus whatever additional sums are needed to assure that
all vested benefits are fully funded by the end of 20 years. Whichever
approach is used, full funding of all vested benefits should probably
be required in 20 years, corresponding to the principle adopted by the
AICPA for the accruing of the costs of vested benefits. A funding
standard promulgated for only a segment of the total benefits would
have to be reconciled with the broader minimum standard articulated
by the Internal Revenue Service.
If the plan should terminate before the required level of funding
has been achieved, the unfunded portion of the accrued liability would
become the obligation of the PGF. The `funding of supplemental costs
arising after inception of the plan from `such causes as retrospective
benefit increases, actuarial losses (unfavorable devi'ation of actual from
assumed experience), and revision of actuarial assumptions should also
come under the "completion" guaranty, but subject to adequate safe-
guard's `against abuse. Since the obligation of the PGF would be
stated in terms of benefit payments, the strengthening of actuarial
assumptions would not add to the existing obligation of the guaranty
fund and might in fact diminish it through stepping up the rate of
funding. Recurring `actuarial losses would be significant in that they
would suggest the strong possibility that the employer's `funding policy
was based on an understatement of the ultimate or "true" cost of the
covered benefits.
The major advantage of this modification would be the reduction
of the risk assumed by the PG-F through the systematic funding of
covered benefits on the basis of presumably realistic cost estimates.
`~ In the United Kingdom, the penalty for nonapproval of a plan by the Inland Revenue
Department is the current taxation of employee participants In respect of employer contri-
butions. If the plan is not funded, the contributions are imputed to the employees.
PAGENO="0232"
224 OLD AGE INCOME ASSURANCE-PART V
The major disadvantage would be the exclusion of plans operating on
a pay-as-you-go basis, unless such plans were forced to convert to a
funded basis. There would be other disadvantages, such as the employ-
er's authority to create additional liabilities for the PGF, without any
recourse against his own corporate assets, and the risks associated
with speculative investment policies.
The viability of the modified approach is clearly dependent upon
the enforcement of minimum standards of funding. There should be
sanctions other than suspension or cancellation of coverage since these
actions would penalize only the plan participants for whose protec-
tion the program was established in the first place. One possible sanc-
tion that would be effective, assuming availability of assets, would be
to make the employer financially responsible for any defaults on
scheduled funding payments. The sanction might take the form of
a tax penalty, possibly equal to the funding deficit, that could be
diverted to the PGF to offset its increased liability. The tax payment
could be refunded, at least in part, if the employer later made restitu-
tion to the plan.
A more comprehensive remedy, that could be applied with or with-
out a funding requirement, would be to make the employer primarily
responsible for any deficiency in plan assets, with the guaranty fund
being only contingently liable. Procedurally, the PGF would assume
full and direct responsibility for the fulfillment of benefit expecta-
tions, but would have right of action against the employer to recover
any asset deficiency. The PGF would be regarded at law as a creditor
of the employer and could be given an appropriate preference in an
insolvency or bankruptcy proceeding. Making the employer legally
responsible for the payment of accrued benefits would encourage con-
servatism in the granting of benefits and in the financing of the bene-
fits. It would eliminate most of the possibilities of abuse (or selection
against the fund) and make it possible for the PG-F rules to be far
less restrictive as to coverage, benefits, and ftmding. On the other hand,
it could discourage the voluntary establishment of qualified plans. For
constitutional reasons, it might be necessary to limit the employer's
legal liability to benefits accruing after enactment of the pertinent
legislation.
Coimpietio'n o~ the Er~io?,'er's Funding Co?n-mitm-ent.-The second
basic approach would be to limit the liability of the PG-F to the
completion of the employer's funding program for covered bene-
fits, without regard to the sufficiency of the projected contributions.
In other words, the guaranty would attach to the funding commit-
ment rather than to the benej~t commitment. In theory, this approach
could he followed without any standards of ftmding other than those
imposed by the IRS as a condition for continued qualification but for
all practical purposes it would have to be grounded on minimum
standards comparable in nature and scope to those suggested in the
preceding section. The initial accrued liability for covered benefits
could be increased by benefit liberalizations, subject to necessary safe-
guards, but a. recomputation on the basis of revised actuarial assump-
tions could be permitted only with the approval of the PG-F. Under
this procedure, the dollar amount of the PG-F's potential aggregate
liability could be definitely determined at any time. It would be the
PAGENO="0233"
OLD AGE INCOME ASSTJRANCE-PART V 225
unfunded portion of the explicitly recorded accrued liability for all
covered benefits of all plans encompassed by the program. If the
covered benefits of a terminated plan could not be satisfied in full
when account is taken of the PG-F's obligation, benefits would have
to be scaled down to manageable proportions.
As with the benefit guaranty approach, there would have to be
sanctions to enforce compliance with the mandated standards of
funding.
This arrangement would insulate the guaranty fund against the
effects of unrealistically low cost estimates on the part of the em-
ployers, an obvious advantage to the PG-F and to employers who
fund on the basis of adequate cost estimates. It would also protect
the PG-F against the undesirable consequences of unduly venture-
some investment policies, unless the fund assumed responsibility for
increases in unfunded liability arising out of realized and unrealized
capital losses-as it might well do. The primary disadvantage of
this approach is that the risk of inadequate asset accumulations would
be shifted to the plan participants. It also fails to deal with the prob-
lem of the pay-as-you-go plan.
The obligation assumed by the PG-F under this approach would
be tantamount to assuring the fulfillment of benefit expectations if
(1) the PG-F prescribed the actuarial cost method and actuarial as-
sumptions to be used and the period of time allowed for the full
funding of all covered benefits, (2) the funding standards were en-
forced, (3) the PG-F assumed responsibility for unfunded actuarial
losses, and (4) the PG-F or an insurer assumed the actuarial risks
associated with benefits that survive a plan termination.
PLANS COVERED
A number of questions are involved with respect to the plans that
would be brought under a pension guaranty program. The most
fundamental question is whether all plans eligible for coverage would
be required to participate in the program. The answer is clearly in
the affirmative. There would be too much selection against the PG-F
if employers were permitted to elect coverage. There would be an
understandable tendency for financially stable firms to stay out of
the system and for the less stable ones to elect coverage. Worse yet,
the latter firms, where the need of a guaranty is greatest, might also
elect to remain outside the program. There might be a universal re-
luctance to participate in any undertaking that would add to the cost
of doing business and, if not properly structured, might be grossly
abused. Compulsion seems to be the only answer, despite the fact
that this feature might narrow the range of plans that could be
brought into the program.
If the program were to be made compulsory, it would have to be
restricted to plans "qualified" under the Internal Revenue Code and
implementing regulations, unless a new control mechanism is. devel-
oped. At the present time the only effective club that the Federal Gov-
ernment has over pension plans is denial of the tax treatment that is
accorded plans which meet certain specifications, designed to prevent
discrimination in favor of stockholders, officers, supervisors, and highly
paid employees. Conceivably, Congress could enact a law making it
PAGENO="0234"
226 OLD AGE INCOME ASSURANCE-PART V
unlawful to hold out the promise of pension benefits, however the
promise might be hedged, unless the anticipated benefits were funded
in a prescribed manner, but the prospects of such legislation in the
near future seems remote. There is the possibility, of course, that a
pension guaranty mechanism might prove to be so attractive that em-
ployers would convert their nonqualified plans into qualified plans in
order to take advantage of the coverage.
Some have questioned whether multiemployer plans should be com-
pelled to participate in the guaranty scheme. The probability of ter-
mination is probably lower among multiemployer plans as a group
than among single employer plans, which would lessen the need for the
guaranty. On the other hand, multiernployer plans are probably fund-
ed at a lower level than single employer plans, as a class, which would
suggest the need for a benefit guaranty. Many of these plans would
have difficulty meeting the minimum standard of funding mentioned
earlier, since a substantial percentage only pay interest on the supple-
mental liability rather than amortizing it. Considerable opposition
to compulsory participation could be expected by multiemployer plan
administrators. On balance, however, it would seem that the guaranty
scheme should be applicable to all qualified plans, whether they be
smgle employer or mult.iemployer plans.
In order to protect the system, a plan should not be eligible for
coverage for the first few years of its existence. Otherwise, an
employer, in contemplation of an event that would invoke the protec-
tion of the system. could establish a plan with liberal past service bene-
fits and let the PGF assume most of the financial burden. The need for
protecting the system against this potentiality would depend on other
provisions of the program. including the definition of the event insured
against and the benefits that would be entitled to the guaranty. The
required length of the probationary period is strictly a matter of judg-
ment but it should perhaps be no shorter than 3 nor longer than 10
years. It is of some significance in this regard that the BLS study
of pension plan terminations cited earlier 22 showed that half of the
plans examined, terminated within 5 years after establishment.
A question might be raised as to the desirability of excluding from
participation in the program plans covering fewer than some specified
member of employees, such as 25. There is no doubt that the probability
of termination is the highest among the smaller plans. The BLS study
revealed that two-thirds of the terminations were accounted for by
plans covering fewer than 25 employees. The issue here is whether the
program should be structured in such a manner as to cover the area
of greatest nee.d or to minimize the financial burden on continuing
plans. Clearly, a guaranty program could be surrounded by so many
safeguards that it would cover only the most unlikely occurrences,
with commensurately low cost to the participating plans. In the light
of the other safeguards recommended in this paper, it would seem
unnecessary-and socially undesirable-to exclude plans purely on the
basis of size.
Another question of considerable moment is whether all plans falling
within fhe eligible group would automatically be embraced in the pro-
grain or whether the administering agency would have the authority
See footnote Hi, p. 209.
PAGENO="0235"
OLD AGE INCOME ASSURANCE-PART v 227
to exclude those plans not measuring up to minimum underwriting
standards. This is primarily a question of priorities. Is the para-
mount objectvie the protecting of benefit expectations or the integrity
of the guaranty fund? If the participants in the less secure plans are
to be denied the protection of the system, much of the latter's value
would be sacrificed. On the other hand, employers who are meeting their
pension obligations in a realistic and forthright maimer can properly
object to assuming a portion of the pension obligations of other em-
ployers, possibly their competitors. The issue would be blunted if a
number of risk classifications could be established, with the premium
being roughly commensurate with the risk of plan termination. The
problem would also be diminished if the projected cost of the program
turned out to be a modest or even negligible proportions.
A related question is whether a plan that falls outside the category
of plans that must participate, could be brought within the scope of
the program by election of the plan administrator. This should be per-
mitted and encouraged if the aim is to maximize the coverage of the
program. On the other hand, the privilege would open the door to
selection against the program. Voluntary election of coverage would
probably be feasible if a sound underwriting procedure could be
devised and enforced.
BENEFITS COVERED
As was stated earlier, the protection of the guaranty fund would
extend only to benefits accrued as of the date of plan termination or,
more precisely, the occurrence of the event insured against. However,
the protection need not extend to all accrued benefits. A distinction
might logically be made between benefits credited for service prior to
inception of the plan and those that accrued thereafter. If only the
latter benefits were to be "insured," the major costs of the program
would be eliminated but by the same token much of the raison d'etre of
the program would be destroyed. The principal problem lies with
past service benefits and others granted retrospectively after the plan
was established.
Another distinction that might be made is that between vested and
nonvested benefits. While the Hartke bills have envisioned the guar-
anteeing of all accrued benefits and the Javits bill all future service
benefits, the current thinking in administration circles seems to incline
toward the guaranteeing of only vested benefits. This thinking, how-
ever, is apparently based on the assumption that Congress will enact
legislation requiring that all benefits vest after 10 years of continuous
service irrespective of the participant's age (except that service prior
to age 25 could be ignored). The contemplated legislation would pro-
vide an orderly transition by requiring only the vesting of benefits
accruing after the date of the legislation but with recognition of the
years ot service prior to that date.23 For example, an employee with 9
~ In his speech before the American Pension Conference on May 11, 1967, AssIstant
Secretary of the Treasury Stanley S. Surrey Indicated that other transitional approaches
were under consideration. One approach would be to require the vesting of ~ll benefits.
whether accruing before or after enactment of the relevant legislation, after 20 years of
continuous service, with the minimum required period being systematically reduced so that
after 10 years the target standard of 10 years would be operative. Another approach
would make the 10-year standard effective immediately but only with respect to one-tenth
of the participant's accrued benefits, with the percentage grading upward in uniform steps
each year and reaching 100 percent at the end of 10 years. Under neither of these alterna-
tive approaches would there be any distinction between past and future service benefits.
PAGENO="0236"
228 OLD AGE INCOME ASSURANCE~PART V
years of credited service at the time of the legislation would achieve
a vested status after 1 more year of service but only the benefits
that accrued during the 10th year would be vested. Benefits for all
subsequent years would, of course, be fully vested. All employees with
10 or more years of credited service at the time mandatory vesting
became effective would be vested in all benefits accruing thereafter. A
participant with 5 years of service would achieve a vested status after
5 more years but only with respect to the benefits accruing during the
last 5 years and thereafter.
Under this concept the question arises as to whether the guaranty
would attach only to the benefits that would be vested under the pro-
posed minimum standard or to those that have vested on a more liberal
standard pursua.nt to the terms of the pension plan. WThile relatively
few plans other than those funded through individual insurance or
annuity contracts vest benefits with less than 10 years of credited serv-
ice many vest past service as well as future service benefits. Moreover,
all benefits of retired persons are considered to be vested and all those
of employees eligible to retire may be so regarded. Any plan that has
had a vesting provision for a number of years is likely to have a greater
volume of vested benefits that would be generated by the proposed
minimum standard. High governmental sources indicate that the Ad-
ministration will take the position that benefits vested by plan provi
sions more liberal than the mandatory minimum should be sub]ect to
the guaranty.
Other distinctions among accrued benefits could be justified. In a
privately circulated memorandum, an official of a. maj or automobile
company proposed that the guarantee attach to all the accrued benefits
of persons in a retired status and those within 10 years of normal re-
timement. ~Tit,h respect to all other en~loyees, the guaranty would at-
tach in annual increments, reaching 100 percent only after the plan
had participated in the guaranty program for 10 years.
There would have to be a determination as to the types of benefits to
which the guaranty would attach. It is obvious that it would apply
to retirement benefits but would it cover death, disability, special early
retirement, and other ancillary benefits? It. would seem that the guar-
anty ought to cover only such ancillary benefits that have matured
and a.re in an active-payment status.
It would be necessary to deal specifically with plan liberalizations
that increase the unfunded liability of the plan, especially increases in
the scale of benefits. The threat to the solvency of the fund is apparent.
The guaranty should not attach to newly created benefits for a period
of years. It would make sense to impose the same probationary
period as that applicable to newly established plans.
It would probably be desirable to place a dollar limit on the bene-
fits that would be guaranteed for any one participant, since there is
an element of social insurance in the whole undertakin'a and some em-
plovers would inevitably subsidize the pension plans of other employ-
ers to some extent. The limit should he stated in terms of the monthly
income provided at a. retirement age specified in the law.
PAGENO="0237"
OLD AGE INCOME ASSIJRANCE-PART v 229
IMPLEMENTATION OF THE GUARANTY
The implementation of the guaranty would involve two basic is-
sues: (1) determination of the amount of the PGF's liability and (2)
manner in which the guaranty would function.24
Determination of the Guarantor's Liabiiity.-The determination of
the guarantor's liability would be divided into three steps. The first
step would be to ascertain in terms of prospective monthly income the
dollar amount of covered benefits accrued as of the date of plan ter-
mination. This could present difficulties if the benefits covered by the
guaranty are not carefully defined. However, it is assumed for the
purpose at hand that the guaranty would be limited to vested ben-
efits, including those vested on a vohmtary basis. It is further assumed
that plaiis will be required by law to state precisely what benefits are
vested, the conditions under which they vest, and how the amount of
vested benefits are computed. There should be no distinction between
the benefits vested upon termination of employment and those vested
upon plan termination except as may be necessary to meet the Treas-
ury requirement that all accrued benefits vest on plan termination to
the extent that they are funded. Otherwise inequities `may be created
as between employees who are laid off in anticipation of the winding
up of a business and those who are still employed at the point of tech-
nical termination of the plan.
The second step would be to derive the actuarial value of the guar-
anteed benefits. Procedurally, this would be simple if only `basic retire-
ment benefits are vested. Their valuation would involve only mortality
and interest (and possibly expense) assumptions. If any ancillary
benefits are vested, additional factors might `have to `be considered.
The basic question here would be whether to use the assumptions that
had guided the employer's funding policy or the assumptions that
would be applied by the agency that would have final Pesponsibility
for payment of the beneflts-the residual risk `bearer. T'he decision
would be influenced or perhaps controlled by the payment mechanism
adopted, to be discussed in the next section.
The final step would be to compare the actuarial liability derived in
step 2 with the value of the assets considered to be available for the
satisfaction of the guaranteed `benefits. The law might state that all
unallocated assets are to be applied on a first-priority basis to the pay-
ment of guaranteed benefits. Presumably it would not direct the re-
capture of assets used to purchase nonvested benefits from an insurance
company or reallocate assets needed to honor benefit commitments to
persons in a retired status even when the benefits do not qualify for
the guaranty. It might recognize the priorities established in the plan
document for the allocation of uncommitted assets in t'he event of plan
termination. Conceivably it might recognize no priorities and thus
assume that guaranteed benefits are entitled to only their pro rata
24 It is assumed for purposes of this discussion that the PGF's obligation would be to
assure payment of the guaranteed benefits. If Its obligation were limited to the completion
of a projected funding program, the amount of Its liability would be determined on a basis
different from that described herein, but Its obligation could be carried out under any of
the approaches outlined.
PAGENO="0238"
230 OLD AGE INCOME ASSURANCE-PART V
share of the assets. Whatever the concept applied, the theoretical
liability of the guarantor would be measured by the difference between
the actuarial value of the guaranteed benefits and the value, at book
or market, of the assets applicable to those benefits. The actual liability
would depend upon the manner in which the guarantor discharges its
obligation.
Manner in Which the Guaranty Would Function.-The guarantor
could discharge its responsibility in a number of ways, the choice
depending to some extent on the nature of the guarantor and to some
extent on the manner in which the funded benefits are to be paid out.
It would seem logical that responsibility for the payment of benefits
purchased from an insurer prior to termination of a covered plan
would be retained by the insurer, with future dividends or experience
refunds going to the PGF. Legal complications would ensue if any
other course of action were to be attempted. Thus, the problem reduces
to the procedure to be followed when the plan assets are held in trust
or in an unallocated fund with a life insurance company.
One approach would be for the funding agency to retain the assets,
paying benefits pursuant to the terms of the plan until the assets are
exhausted, with the guarantor then assuming responsibilty for pay-
ment of the remaining guaranteed benefits. If the guaranty program
were to give guaranteed benefits first claim to available assets, dis-
tributions would have to be limited to those persons whose benefit
rights were protected by the program, unless the assets appea.red to
be more than sufficient to meet the priority claims. In the latter event,
it would be necessary to divide the assets into two accounts, one being
held for guaranteed benefits and the other for nonguaranteed benefits.
This approach would have a number of virtues. The guarantor's
liability would be determined on the ba.sis of actual, emerging experi-
ence; competitive relationships between and among banks and insur-
ance companies would not be distorted (as might happen under other
approaches) ; it would not be necessary to transfer funds to the guar-
antor, thus avoiding the risk of liquidation losses and minimizing the
asset accumulation of the guarantor-an especially desirable objective
if the latter is a governmental agency; and (a minor point) the pen-
sioner would receive only one check per month rather than one from
the funding agency and another from the guarantor.
On the negative side, this approach would not lend itself as well
as some others to an employee guaranty of an asset deficiency, how-
ever it might be formulated. The true deficiency would not be known
for many years, by which time the employer might have gone out
of business. It would be possible, of course, for the guarantor to levy
an assessment on the employer at the time of plan termination in an
amount equal to the actuarially estimated deficit. Secondly, potential
complications would be involved if it became necessary to divide the
accumulated assets between guaranteed and nonguaranteed benefits.
The present value of the guaranteed benefits would have to he actuari-
ally estimated and only by sheer coincidence would the estimated and
ultimately realized costs be the same. Theoretically, the guarantor
should incur no liability but if the cost estimate proves to be too low,
it would have to assume the deficit. On the other hand, if the estimate
were too high, individuals with nonguaranteed claims would have
PAGENO="0239"
OLD AGE INCOME ASSURANCE-PART v 231
been unnecessarily deprived of some of their benefits. A third com-
plication would arise if the trust agreement between the bank and
the employer were to call for dissolution of the trust upon termina-
tion of the plan, with the assets to be applied to the purchase of an-
nuities. There is no reason why such an agreement could not be en-
forced if the asset allocation formula were not inconsistent with the
guaranty program, since this would immediately fix the amount of
the guarantor's obligation. Finally, in the latter stages of liquidation
of a trust, liquidity problems could arise, possibly resulting in some
otherwise unnecessary capital losses.
A second approach would be for the funding agency to pay that
portion of each employee's total guaranteed benefit that could be pro-
vided by the assets in its possession, with the guarantor concurrently
paying the remaining portion. This would necessitate an actuarial
estimate of the benefits that could be paid by the funding agency, with
possible discrimination against either the guaranty fund or the in-
dividuals with nonguaranteed benefits if the estimates should prove
wrong, as would be virtually certain. This method would avoid trans-
fer of funds at time of plan termination and would leave undisturbed
existing competitive relationships.
A third approach that would involve approximately the same ad-
vantages and disadvantages as the first two would be for the guarantor
to transfer to the funding agency the additional sums actuarially esti-
mated to be needed to pay the guaranteed benefits. The benefits would
be charged to the pension fund as they were paid, even if the funding
agency should be an insurer. In other words, the insurer would not
underwrite the benefits, offering only investment and disbursement
services. If the sums transferred proved to be inadequate, the guar-
antor would advance additional funds as the needs manifested them-
selves. If the sums turned out to be excessive, the funding agencies
would be expected to return the unused funds to the guarantor. A deli-
cate question that would be involved in this arrangement would be the
extent to which the funding agency could invade the corpus (or fund)
for its expenses and possibly a profit. Another-equally sensitive-
question would be the extent to which the PGF could influence or direct
the policy to be followed by the funding agency in the investment of
the moneys entrusted to it.
A fourth approach would be for the funding agency to transfer to
the guarantor a sum equal to the assets assumed to stand behind the
guaranteed benefits. Unless the assets were transferred in kind, the
funding agency should be permitted to spread the liquidation over a
period of years to avoid capital losses or other forms of adverse fin-
ancial consequences. If the employer were to be held responsible for
the asset deficiency, the amount of his liability should be fixed at point
of plan termination but he should be permitted to spread his payments
to the guarantor over a period of years, possibly equal to the remain-
ing years in the original period over which he was to have funded the
benefits. In other words, if the employer were supposed to have the
guaranteed benefits completely funded within 20 years from plan in-
ception or a later event, and the plan should terminate within 10 years,
he would be given 10 years in which to make up the deficiency-as
under the original schedule. If he should go out of business before corn-
PAGENO="0240"
232 OLD AGE INCOME ASSURANCE-PART V
pleting the payments, the unpaid amounts would become a claim
against his business assets with whatever preference the law might
assign to it.25
The guarantor might discharge its obligation in one of two ways.
The first would be for the guarantor to act as the risk bearer and pay
the guaranteed benefits directly to the claimants as they become due.
Any benefit amounts not taken care of by the assets transferred to it
from the funding agency would be borne out of the guarantor's gen-
eral assets. This would be the logical course of action if the guarantor
were to be a private agency jointly administered by the private organi-
zations participating in the private pension movement. If the guarantor
were a public agency, it might be preferable for it to discharge its ob-
ligation immediately and cleanly by the purchase of nonparticipating
annuities from individual insurers or a pool of insurers formed for
that purpose. Under this procedure there would be no uncertainty
concerning the guarantor's ultimate liability, and the amount of the
employer's liability, if any, would be promptly and definitely deter-
mined. Furthermore, the assets securing the benefit rights would be
kept in the hands of private agencies. If the decision were made to deal
with individual insurers, appropriate rules would have to be promul-
gated with respect to the qualifications of the insurers and the proc-
ess by which the insurers would be chosen. The latter would presum-
ably be on the basis of competitive bids. A different set of procedures
would be needed if a pooi of insurers were to be used, with safeguards
to assure that the rates are not excessive.
Under all these approaches benefits not subject to the guaranty
would be paid, to the extent that assets are available, by the funding
agency holding the assets at the time of plan termination, unless the
trust agreement (or other legal document) specified some other
arrangement.
FINANCING THE GUARANTY
The final issue, and a most crucial one, concerns the principles that
might be followed in the financing of a guaranty fund. As with the
other issues, there are a number of facets to be considered.
The first matter, about which there seems to be general agreement,
is the determination of the base or bases to which premium or assess-
ment rates would be applied. If the program were to include specific
protection against liquidation losses at time of plan termination, the
premiums for that component should be based on the amount of assets
in the plan. The premium rate, of course, would have to be derived
from some estimate of the volume of liquidation losses to be expected
and under what circumstances. The risk would appear to be limited
primarily to trust fund plans and separate accounts of insurers. Since
it is improbable that this particular risk would be separately dealt
with in a guaranty scheme, no further attention will be given to it.
The premium base for the fundamental risk that would be involved
in the program, namely, inadequate accumulation of assets, should
be the unfunded liability for the accrued benefits eligible for the guar-
~ The concept of requiring the employer to fulfill his funding commitment would be
equally applicable to the other approaches described herein, but the implementation of the
requirement would differ somewhat.
PAGENO="0241"
OLD AGE INCOME ASSURANCE-PART V 233
anty. This, of course, requires a valuation of both the liabilities and
ass3ts subject to the guaranty. In the interest of uniformity, the bene-
fits would have to be valued on the basis of the accrued benefit cost
method which concerns itself only with the benefits assumed to have
accrued, explicitly or implicitly, as of the date of valuation. The pro-
jected benefit cost methods take both accrued and future benefits into
account and produce results that can be quite different from the values
derived by the accrued benefit cost method. More important, certain
of the projected benefit cost methods are structured in such a mamier
that they never produce an unf'i~nded accrued liability. Under those
methods, the assets on hand plus the present value of future anticipated
contributions always equal the present value of total prospective bene-
fits. Thus, for purposes of the guaranty fund, the liabilities would
have to be computed on the basis of the accrued benefit cost method,
irrespective of the cost method used for funding and other purposes.
Under the accrued benefit cost method, assumptions need be made
only with respect to mortality, interest, and possibly expenses. Since
the values derived for a particular set of benefits are greatly influenced
by the mortality and interest assumptions, especially the latter, it
would undoubtedly be necessary, in the interest of equity, for the ad-
ministering agency to specify the assumptions to be used. The standard
might well be based on the nonparticipating annuity rates currently
being charged by the leading life insurance companies. The assets
would have to be valued on a realistic basis and if liabilities are valued
at "market" annuity rates it would seem appropriate to use the market
value of assets. The values would be certified by independent or public
accountants. The difference between the actuarial value of the accrued
guaranteed benefits and the value of the assets allocated by law or plan
provisions to the benefits coming under the guaranty, would constitute
the premium base.
Another matter, with respect to which there is no clear-cut answer, is
whether the system should be financed on the basis of advance pre-
miums, assessments, or a combination of the two. It is essentially a
question of whether premiums would be paid before or after the fact.
In general, advance premium arrangements are considered preferable
to assessment schemes. In this particular case, however, the circum-
stances would suggest careful consideration of the assessment ap-
proach. On the basis of available data, it would be extremely difficult
to predict the claims that would arise under a guaranty program. Any
scale of premium rates that might be set would likely prove to be
excessive or inadequate until claims experience can be accumulated.
Moreover, since claims would undoubtedly be clustered in periods of
economic adversity, it would be necessary to accumulate substantial
reserves. In the minds of some, this would be an undesirable develop-
ment if the guarantor should be a Government agency.
The assessment approach would avoid the difficulties of estimating
claims in advance. The assessments would be based on realized experi-
ence, and the claims and expenses would be apportioned over the par-
ticipating plans in proportion to their respective unfunded liabili-
ties-as under the advance premium method. There could be an initial
assessment to provide "working capital" with annual assessments being
levied thereafter. The basic objection to the assessment approach would
83-200-67--pt. 5-16
PAGENO="0242"
234 OLD AGE INCOME ASSURANCE-PART V
be that it would not accumulate sufficient funds to meet the claims
generated by a severe and prolonged depression. The assessment rate
would have to be increased at the very time that the covered firms,
as a group, would have the least capacity to respond to the assessment.
Some firms might default on their assessment, causing a heavier
burden to be placed on the other firms. The financial crisis might be so
acute that in order to shift a. greater proportion of the claims to the
stronger firms, the administering agency might levy assessments on a
basis other than unfunded liabilities (for guaranteed benefits). There
would also be a question as to whether the amount of the assessments
would be determined by the timing of plan terminations, the retire-
ment of the individuals affected, or the actual payment of benefits. If
either of the latter two events should serve as the basis for assessments,
the wrong generation of plans would bear the burden of past plan
failures. Finally, it is argued by some that the assessment approach is
wrong in principle, in tha.t the plans that create the losses for the guar-
anty fund (and by inference were the poorest risks from the begin-
ning) bear the smallest share of the aggregate burden. The plans that
terminate the first year, for example, would pay only the initial assess-
ment. It would appear, however, that the same objection could be made
to an advance premium plan of insurance to the extent that accumu-
lated reserves do not meet all claims.
A compromise a.pproach would be to charge annual premiums at a
minimum level, using the assessment authority to make up any deficits
that might arise.
If an advance premium were to he charged, it should reflect the
best estimate as to the rate a.t which claims will occur over a future
period encompassing all phases of the business cycle. A study of past
business failures should be helpful in this regard. If the analysis
indicates that the guaranty fund could be sustained by premiums
that would be regarded by the business community as inconsequential,
a uniform premium rate should be developed. If the burden would
appear to be consequential, it would be desirable to consider a rating
structure that would differentiate as to (1) age of firm, (2) size of
firm as measured by the most significant indices, and (3) period during
which the pension plan has been in operation. The purpose would be
to predict, and to reflect in the premium rate classifications, the proba.-
bilit.y of plan termination among the various business firms covered
by the program. It would probably be necessary to place an upper limit
on the premium rate for the most ha.zardous classifications in order not
to place an intolerable burden on financially insecure firms.
Finally, it would be necessary to consider a reinsurance arrangement
if the guaranty fund were to operate under private auspices. The
Govermuent would be the logical reinsurer, and it would be appropriate
for the fund to pay a premium for this service. The reinsurance
premium might be absorbed out of t.he regular premium paid by the
participating firms or it could be the basis of a separate levy. The pro-
gram should be self-supporting in the long run even if the fund oper-
ates under governmental auspices, but the borrowing power of the
Government should be used to meet a short-run excess of claims over
accumulated resources.
PAGENO="0243"
OLD AGE INCOME ASSURANCE-PART V 235
A MINIMUM PROGRAM
Thus far this paper has concerned itself with the basic question of
whether some type of pension guaranty arrangement would be techni-
cally feasible and the issues that would have to be resolved if such a
program were to be established. It should be clear at this point that
a guaranty scheme would be feasible from a technical standpoint if
certain conditions were satisfied and adequate safeguards were built
into the system. Some of the conditions and safeguards would involve
regulatory controls that employers, unions, and other elements of the
pension establishment have in general opposed as being potentially
detrimental to the continued sound growth of the private pension
movement. They would also limit the scope of the `arrangement to
such narrow bounds that the social objectives underlying the proposal
might `be frustrated in large part.
Resolution of the fundamental question of whether a properly
structured and delimited guaranty scheme should be established is
beyond the purview of this paper. That is a political decision that will
have to be made by Congress in its wisdom. The contribution of such
an institution to the public weal must be balanced against whatever
harmful consequences might flow from it. Without taking a position
for or against the proposition, the remainder of this paper suggests
*the characteristics or features that should `be associated with any
guaranty scheme that might be brought into existence. The proposals
envision a minimum program, with the thought t'hat extensions and
liberalizations could be introduced as experience `with the system indi-
cates the wisdom of such action.
ADMINISTERING AGENCY
The program should be administered `by a Federal agency. It would
have to be brought into existence `by Federal legislation, and it would
seem appropriate to enforce the law through `a public agency. If there
are to be effective remedies for noncompliance with the requirements
of the law, the Government should apply them directly rather than
through a private intermediary. Moreover, it would be simpler to
make the Government the residual risk bearer, as it probably must be,
if it `acts as the fiscal agent for the program. T'his recommendation
contemplates th'at private insurers w'ould underwrite the guaranteed
benefits of terminated plans, which would minimize the accumulation
of assets in the administering agency. In any event, the Federal Gov-
ernment is holding billions of dollars in various trust funds at the
present time with no `apparent harm to the economy; so the accumula-
tion of a few billion more in a trust fund for private pension bene-
ficiaries would cause no difficulties other than increasing the cost of
the program to employers because of the relatively low yields on the
Government securities held by the trust fund. Moneys in the trust
fund not needed for current operations should be inveSted in obliga-
gations of t'he Federal Government not private securities.
EVENT INSURED AGAINST
The guaranty should extend only to benefit claims arising out of
complete plan terminations. Coverage of partial terminations would
PAGENO="0244"
236 OLD AGE INCOME ASSURANCE-PART V
not only create complex problems of claim adjudication but would
open the door to countless forms of abuse, possibly to the point of ren-
dering the system inoperable. It would be almost. impossible to define
the insured event if various kinds of partial terminations were to be
brought within the contemplated coverage. Many of the problems cited
by critics of the guaranty fund proposal are centered in the concept
of partial terminations. Exclusion of partial terminations would lessen
to some degree the social utility of the system but if a reasonable
level of vesting is brought about, whether by mandate or voluntary
action, the employees in the greatest need of, and with the strongest
claim to, the benefit guaranty will enjoy the protection of the system.
Not oni should the guaranty be limited to complete plan termina-
tions, it should be invoked only when the firm goes out of business. It
would be grossly unfair to other employers, some of them competitors,
if a firm could terminate its pension plan, transfer to the guaranty
fund the responsibility for making good on the unfunded guaranteed
benefits, and then continue in business, its competitive position im-
proved by reduction in its labor costs. This would be comparable to
having the Federa.l Deposit. Insurance Corporation assume responsi-
bility for losses to bank depositors while permitting the bank to con-
tmlle in uninterrupted operation with no loss to itself or its stock-
holders. If the firm were. sold to, or merged with, another company,.
the surviving company should be required to assume the accrued p~~t-
sion obligations of the acquired firm, at least to the extent they come
lrnder the aegis of the guaranty fund. The requirement would be
deemed satisfied if the surviving company were to provide benefits
under its plan to the former einpolyees of the acquired company in an
amount at least equal to the unfunded benefits of the plan of the hqui-
dated company.
The lack of protection for benefit rights in terminated plans of
employers who continue in business should be rectified by requiring
the employer to continue funding contributions in respect of the bene-
fits that would become the obligation of the guaranty fund in the
event that the employer should go out of business. Tile funding would
normally continue at tile rate Prescribed for gOillg plans, but tile ad-
ministering agency should be given tile authority to relax (or spread
out) the funding contributions in tile ligllt of tile financial situation
of tile employer. If tile employer should go out of business before
completing the fmlding schedule tile unpaid amounts would not. be-
come a claim against his assets alld the guaranty fund would assume
full responsibility for the paylneilt of tile Ullfullded benefits elltitledi
under the law to the guaranty. There would have to be provisions
in the law designed to prevent tile employer from avoiding his obliga-
tiOlls by ostensibly going out of business and then reopening under
another name or in another form. If the business were sold or merged,
the continuing company would have to assume tile funding commit-
ment of the acquired firm. Likewise, if a. plan is terminated in order
to transfer the participants to another plan, new or existing, the con-
tinuing plan sllould assume the obligations of the old.
The foregoing principles would have to be modified in the case of
multiemployer plans. Wilere more or less permanent employment. re-
latioflsllipS exist, tile guaranty Sllould become operative with respect
PAGENO="0245"
OLD AGE INCOME ASSURANCE-PART V 237
*to the employees of a particular firm upon withdrawal of that firm
from the plan for reasons beyond its control and subject to a minimum
period of participation. The guaranty of the PGF would be residual
in character if the multiemployer plan had an internal guaranty
mechanism. When the plans deals with strictly transitory employment
relationships, it would appear that the guaranty could become effec-
;tive only upon termination of the plan itself. In the meantime, the
benefits entitled to the protection of the guaranty should find their
fulfillment in the accumulated assets of the plan. `While the minimum
funding standards outlined in the following sections should be fully
applicable to multiemployer plans, it would probably be impracticable
to try to enforce the completion of projected funding schedules in the
event of plan termination or the withdrawal of a participating firm
from a continuing plan, especially in the face of transitory employ-
ment relationships.
OBLIGATION OF THE GUARANTY FUND
A pension guaranty fund is feasible only if superimposed on mini-
mum standards of funding. Technically, it would .be. sufficient if these
standards related only to the benefits subject to the guaranty. How-
ever, in order to preserve the protection now afforded nonguaranteed
benefits through IRS minimum funding requirements and to harmo-
nize the funding requirements of the guaranty system with the cost.
accrual position of the accounting profession, it is recommended that
the law require annual contributions to a pension plan equal to the
normal cost of currently accruing benefits-whether or not guaran-
teed-plus whatever additional sums are necessary to have all guaran-
teed benefits fully funded within 20 years after the effective date of the
coverage. Any additional guaranteed benefits that might be granted
retrospectively by plan amendment would have to be funded in full
within 20 years after such amendment. Evidence that the minimum
level of funding is being maintained would be furnished annually or
triennially through certification from a member of the American
Academy of Actuaries who would be free to choose his own actuarial
assumptions and cost method in respect of the nonguaranteed benefits.
The administering agency would specify the actuarial assumptions
and possibly the actuarial cost methods to be used for guaranteed bene-
fits. This is based upon the assumption that only vested benefits would
be guaranteed, the valuation of which would require only mortality
interest, and expense assumptions. The administering agency should
be given the authority to collect delinquent funding contributions,
extending into insolvency or bankruptcy proceedings.
As indicated above, the guaranty fund would incur obligations only
when a pension plan termination is accompanied or followed by the
liquidation of the sponsoring firm except for multiemployer plans.
At that point, its obligation would be to assure the ultimate payment
of the guaranteed benefits of the plan. That is, its obligation should be
stated in terms of benefit payments rather than the completion of a
funding objective per se. The theoretical measure of its obligation
would be the difference between the assumed present value of the guar-
anteed benefits and the value of the assets available for payment of the
PAGENO="0246"
238 OLD AGE INCOME ASSURANCE-PART V
benefits. Any benefits already purchased from an insurer ivould be
subtracted from both sides of the equation.26 The true measure of the
obligation would be what it would cost the guarantor to purchase the
guaranteed benefits from an insurer or to pay the benefits directly to
the eligible recipients.
Stating the guarantor's obligation in terms of benefit fulfillment
would suggest that the system should underwrite the entire asset de-
ficiency, whatever the cause. It would surely be appropriate to absorb
any deficit arising out of actuarial losses since the guarantor would be
dictating the assumptions. The underwriting of capital losses would
be a little more debatable if there were no restrictions on investment
policy. On balance, however, and in the interest of simplicity, it would
seem desirable for the system to cover capital losses also. A deficiency
arising out of a retrospective benefit increase or other type of plan
liberalization would also be covered so long as the benefits involved
come under the guaranty.
In accordance with an earlier recommendation, the guaranty fund
would have no recourse against the assets of a liquidating firm, except
for delinquent funding payments.
PLANS COVERED
Participation in the guaranty scheme should be compulsory for all
eligible plans. Compulsion would be necessary to get adequate partici-
pation and to protect the guaranty fund against adverse selection.
Eligibility should be limited to plans that "qualify" under IRS regu-
lations, which unfortunately would rule out pay-as-you-go plans whose
participants would have the most to gain from a benefit guaranty.
There would be little danger to the system in admitting any plan that
would voluntarily subject itself to the funding requirements and other
features of the system.
Multiemployer plans should be required to participate with what-
eve~' modifications might be necessary to fit their particular circum-
stances. The basic modifications that might be appropriate have been
indicated above. Many of these plans could be expected to object to
the proposed minimum standards of funding, as well as minimum
vesting provisions, but it is highly desirable that these plans meet t.he
same funding and vesting standards as single employer plans.
Plans should be eligible for coverage only after they have been in
operation for a minimum of 5 years and presumably should not be
forced into membership until they have benefits subject to the guar-
anty, which could involve a period as long as 10 years. This would
greatly reduce the cost of the system and discourage the establishment
of plans for the sole purpose of enjoying the benefit. guaranty. Almost
half of the terminations studied by BLS occurred among plans that
had been in operation for 5 years or less. Only a fourth of the plans
had been in existence for more than 10 ye.ars.
Conventional insurance theory would suggest that all eligible plans
should be expected to make application for coverage and demonstrate
~ It would be reasonable to require the Insurer to amend its contract to provide that
future dividends or experience refunds in respect of guaranteed benefits would be payable
to the guaranty fund.
PAGENO="0247"
OLD AGE INCOME ASSURANCE-PART V 239
to the satisfaction of the administering agency that its financial con-
dition and economic prospects are such as to justify membership in
the system. Extension of the guaranty to liberalized benefits would
also be subject to underwriting. In practice, such screening would im-
pose a heavy-perhaps intolerable-administrative burden and, more
important, would conflict unduly with the social goals of the program.
Thus, all qualified plans in operation for more than 5 years at the time
the guarantee fund is established should be automatically covered
irrespective of the financial condition of their sponsors. Other eligible
plans should likewise be automatically covered as soon as they satisfy
the 5-year probationary period and have benefits entitled to the
guaranty.
BENEFITS COVERED
The guaranty should be limited to benefits vested under the terms
of the plan or by operation of law. However, it should encompass not
only benefits earned after inception of the plan but also those credited
for service prior to that date. It should also extend to all benefits of
those employees who have retired or are eligible to retire with normal
or reduced benefits. It should not be applicable to those benefits that
vest only by virtue of discontinuance of the plan.
In order to assure a minimum level of protection under the program
and to prevent complete avoidance of the guaranty by employers
inclined in that direction, it would be necessary for the law to require
a minimum degree of vesting, applicable to both single employer and
multiemployer plans. As was pointed out earlier, there is a strong
probability that the administration will recommend legislation that
would provide for vesting of future service benefits after 10 years of
service, with recognition being given to prior years of service in deter-
mining whether the minimum period has been satisfied. Since it would
be many years before this legislative mandate would produce a level
of funding equivalent to that found in many plans today, it would be
desirable for the guaranty to include benefits voluntarily vested under
the terms of the plan.
Vested benefits created through a retrospective liberalization of the
plan should not be eligible for the guaranty until 5 years after the
liberalization. This restriction would be necessary to protect the fund
against those who would otherwise grant benefit increases just prior
to winding up their business.
The monthly benefits of any particular employee should be guar-
anteed only to the extent that they do not exceed the lesser of 50 per-
cent of his monthly compensation at the time of plan termination or
$500. Ancillary benefits should be guaranteed only if they are in a
payment status at the time of plan termination. The amount of such
benefits entitled to the guaranty should be reasonably related to the
amount of monthly retirement benefit guaranteed under the program.
IMPLEMENTATION OF THE GUARANTY
Upon occurrence of a plan termination comin~r within the scone
of the guaranty, as determined by the administering agency, the
guaranty fund would assume full responsibility for the payment of
all guaranteed benefits. It would take title to, or assert in some other
PAGENO="0248"
240 OLD AGE INCOME ASSURANCE-PART V
appropriate manner its jurisdiction over, the assets in possession of
the funding agency assumed to be available for the satisfaction of
guaranteed benefits. Its jurisdiction should extend only to unallocated
funds, thus excluding insurance or annuity contracts already pur-
chased for specific individuals. This would discriminate somewhat
in favor of allocated funding instruments, which may provide for the
purchase of nonvested benefits. but the difference in treatment appears
unavoidable. It would not seem equitable or to be good public policy
to cancel benefits already purchased. On the other hand, it would seem
appropriate for the guaranty fund law to specify that vested (that
is, guaranteed) benefits, including those payable to retired employees,
will have the first claim to all unallocated funds. If all of the funds
are not needed to provide for the guaranteed benefits, the excess would
remain with the funding agency for application to nonguaranteed
benefits pursuant to terms of the plan.
The guarantor should discharge its obligation by purchase of in-
surance or annuity contracts from a pool of life insurance companies
for the full amount of guaranteed benefits. This would fix immediately
and irrevocably the amount of funds needed to underwrite the guar-
anty and, hei~ce, the amount of assets that should be transferred
from t:he funding agency. If there were delinquent funding obligations
outstanding against the employer, the guarantor would be author-
ized and directed to seek collection of these sums from the assets of
the liquidating firm, with whatever creditor's preference Congress
might see fit to provide. Any sums collected in excess of the deficit
originally assumed by the guarantor would he turned over to the
original funding agency for application to nonguaranteed benefits.
In order to minimize liquidation losses, the funding agency should
be permitted to spread the transfer of assets over a period of time,
perhaps up to 5 years. The assets should be transferred in a systematic
manner (in installments), with the funding agency having the option
at all times of transferring the remaining assets in one sum. Invest-
ment earnings on the moneys still held by the funding agency would
inure to the benefit of the guarantor.
The pool of insurers from which the guaranteed benefits would
be purchased would have to be set up and administered in a manner
to assure right of participation by all qualified insurers and to pro-
tect the interests of the guaranty fund. Arrangements similar to
those established for Federal Employees Group Life Insurance and
Servicemen's Group Life Insurance would appear to be suitable. The
benefits should be purchased on a nonparticipating basis in order to
determine definitely and immediately the magnitude of the guarantor's
obligation, thus contributing to an equitable allocation of plan assets
between guaranteed a.nd nonguaranteed benefits.
FINANCING TIlE GUARANTY
The guaranty system should be supported by contributions from
employers whose pension plans fall within the scope of the program,
the objective being to make the program self-supnorting as to both
benefit obligations and administrative expenses. The primary source
of support should be annual premiums levied on the basis of the un-
funded accrued liability for guaranteed benefits. For the purpose of
PAGENO="0249"
OLD AGE INCOME ASSURANCE-PART V 241
determining the premium base, the actuarial liability of the accrued
benefits would be computed on the basis of annuity rates (reflecting
mortality, interest, and expense assumptions) provided by the guar-
antee fund. These rates should bear a reasonable relationship to the
nonparticipating rates for deferred and immediate annuities being
quoted on a plan closeout basis by the principal group annuity com-
panies. The assets would be valued at market, the certification being
made by a public or independent accountant. Account would be taken
of only those assets allocable to guaranteed benefits.
The premium rate should be based upon the best statistical evidence
as to the probable rate of termination among the plans covered by the
guaranty and the magnitude of the losses that would be sustained by
the guaranty mechanism. Technically, there should be rate dif-
ferentials based upon the age and financial strength of the sponsoring
firm but for all practical purposes it would seem appropriate to charge
a uniform rate. It might be necessary to have a different rate (or rates)
for rnultiemployer plans if the modifications suggested earlier are
made applicable to them. As a general proposition the rate, or rates,
should be set at the lowest justifiable level, with the understanding
that assessments would be levied to make up any deficits. There should
be a limit on the amount of assessments that could be levied in any one
year, such as five times the annual premium. The premium rate should
be subject to upward or downward adjustment as experience with the
program develops.
The guaranty fund should have borrowing authority sufficient: to
absorb any deficits that might arise in the short run. Deficits of con-
siderable magnitude could develop in the course of a severe depression.
If the claims against the funds should reach catastrophic proportions-
out of reach of even the assessment authority of the administering
agency-the Government should assume an appropriate share of the
total burden in recognition of the fundamental nature of the risk.
SUMMARY
Within the last few years, strong interest has developed within cer-
tain quarters in some type of cooperative arrangement that would
assure the fulfillment of legitimate benefit expectations under private
pension plans, irrespective of the financial status of the plans or their
sponsors. The concept has found its way into various legislative pro-
posals, some of which are currently pending before Congress.
The Setting
The need for a guaranty arrangement must be evaluated against
the background of the limitations on the employer's undertaking in
respect of a pension plan. The employer may undertake, unilaterally
or pursuant to the terms of a collective-bargaining agreement, to set
aside funds on a specified basis, such as an amount per man-hour or
man-day of work, without formal reference to the scale of benefits
that can be provided by such contributions. The employer's obliga-
tion to the plan is completely fulfilled when he pays over the appro-
priate sums to a funding agency, even though the assets of the plan
eventually prove insufficient to provide the level of benefits projected
on the basis of the anticipated contributions. On the other hand, the
PAGENO="0250"
242 OLD AGE INCOME ASSuRANCE-PART V
employer may undertake, voluntarily or in response to union demands,
to contribute whatever sums are necessary to provide a fixed scale of
benefits set forth in the plan. The benefit formula of such a plan
usually recognizes, and gives credit for, some or all of an employee's
service performed for the employer in question prior to the inception
of the plan, and subsequent benefit liberalizations are frequently given
retrospective effect, both practices giving rise to an unfunded accrued
liability that would be the primary source of loss to any guaranty
arrangement. Except for collectively bargained plans, the employer
reserves the right to alter, modify, or terminate the plan at any time
and to suspend, reduce, or discontinue contributions whether or not
previous contributions have been sufficient to provide all benefits
credited to date. It is also customary for the plan to state that the em-
ployer's obligation, in the event of plan termination, shall be limited
to contributions already made to the plan. In other words, the partici-
pants and pensioners must look to the accumulated assets of the plan
for the satisfaction of their claims.
In order to meet the benefit commitments, explicit or implicit, gen-
erated under a pension plan, the employer generally sets aside funds
with a bank or insurance company in amounts and at times roughly
commensurate with the rate at which the pension costs accrue, a prac-
tice known as fvriuling. Under a modification of this practice called
teiimina2 fwnding only the benefits of retired employees are funded.
In a relatively few cases, the employer pays the benefits directly to
retired employees, a method of financing known as current disburse-
ment or pay-as-you-go. Under existing law, an employer is under no
legal obligation to fund his accruing pension costs, but if the plan is
to enjoy the tax treatment accorded a "qualified" status under IRS
regulations, he must as a minimum fund the normal cost of the plan
plus interest on the initial supplemental liability. Moreover, under a
rule recently adopted by the public accounting profession, the em-
ployer must charge to expense his annual pension cost accrual and to
the extent that he does not hereafter fund the expense charges, he
must reflect in his balance sheet the cumulative excess of charges over
funding contributions.
The pattern of accounting charges and funding payments is based
upon estimates of future costs prepared by actuaries who make assump-
tions as to mortality, investment earnings, disability, nonvested with-
drawals, salary scales, and retirement ages. It is assumed that normal
costs, as determined by so-called actuarial cost methods, will be funded
currently and that supplemental costs, if any, will be funded-if at
all-over an extended period of time, usually ranging from 12 to 40
years. As of any given time, the assets of a pension plan may be less
than the actuarial value of the accrued benefits because of inaccurate
estimates of cost, failure of the employer to undertake a funding pro-
gram that would ultimately meet all costs, lack of time for the com-
pletion of a realistic funding objective, or loss of asset values through
realized or unrealized capital losses. A pension guaranty fund would
be designed to deal with an insufficiency of assets, as respects covered
benefits, at time of plan termination or under other specified circum-
stances.
PAGENO="0251"
OLD AGE INCOME ASSURANCE-PART v 243
Applicabz?ity of Insurance Concepts
Such an arrangement would be based upon insurance principles,
and its feasibility should be tested against the criteria of an insurable
hazard. There are (1) large number of homogeneous risks; (2)
objective determination of the occurrence and amount of loss; (3)
randomness of loss; (4) low probability of loss; (5) significance of
loss; and (6) absence of catastrophe hazard. The first criterion would
be met if all eligible plans were compelled to parLicipate. The second
would be satisfied only if the contingency insured against were
clearly-and perhaps narrowly-defined and the benefits to be insured
were precisely articulated. Losses would not occur in random fashion
unless many safeguards were built into the system. The fourth and
fifth criteria would be fulfilled to a reasonable degree, as would the
sixth. Losses of catastrophic dimensions could occur during depressed
economic conditions but the problem would be minimized by the fact
that most of the claims would represent deferred obligations and would
not have to be fully offset by assets in the guaranty fund at any pornt
in time. In any event, a temporary shortage of assets could be met by
a governmental subvention or loan.
Additional insights into the feasibility of a pension guaranty fund
can be gained by examining the essential elements of existing insur-
ance arrangements that fail in one or more important respects to
satisfy the conventional concepts of a sound insurance program. Les-
sons can be learned from the Federal Deposit Insurance Corporation;
the various Federal mortgage insurance funds; State guaranty funds
to insure payment of automobile, workmen's compensation, and life
insurance claims; and State unsatisfied judgment funds to protect
against financially irresponsible motorists. In the private sector, credit
insurance and performance bonds provide protection against the un-
willingness or financial inability of business organizations to meet
their obligations, a risk greatly influenced by the economic climate.
Then there are a number of insurance programs that involve a part-
nership of some type between the Federal Government and private
insurance agencies. In some of these programs, the private agencies are
the sole risk bearers, the Government playing a strictly administra-
tive role. In others, the private agencies furnish only fiscal and claims
services, the Government assuming the entire risk. In still other cases,
the Federal Government and private insurance agencies have entered
into a jornt underwrzting venture under which the Government as-
sumes that portion of the total risk considered to be uninsurable by
private agencies. Finally, the Swedish pension guaranty fund, which
has been in operation since 1960, provides actual experience with a
pension guaranty undertaking.
Issues
Many issues would have to be resolved if a pension guaranty fund
were to be established in the United States. The first would be whether
the fund, hereinafter referred to as the PGF or the guarantor, would
be established and operated under the auspices of a Federal agency,
a private agency, or a combination Government-private instrumental-
ity. Any of these approaches would seem to be feasible, the choice
depending in part on political philosophies and in part on the finan-
cial mechanism envisaged.
PAGENO="0252"
244 OLD AGE INCOME ASSURANCE-PART V
The most difficult problem that would have to be confronted would
be defining or articulating the circumstances under which the protec-
tion of the system could be invoked. The most basic question is whether
t.he guaranty would become operative only upon termination of the
entire plan or also upon other occurrences that would adversely affect
the benefit expectations of a substantial percentage of the covered
employees. Another fundamental question is whether the pension
guaranty should be invoked when the firm that created the pension
obligations continues to operate in one form or the other, even though
the plan has been completely terminated. A plan may be terminated
under any number of circumstances that would raise doubts concern-
ing the propriety of transferring to the. PG-F the responsibility of
meeting benefit expectations. The whole matter would be greatly
simplified if the guaranty scheme were established on the basis that.
the sponsoring firm, or it.s successor, would have the primary legal
responsibility of meeting the cost of t.he benefits covered by the guar-
anty, the PG-F having only the residual liability. Special rules would
have to be developed for multiernployer plans, since among other
distinguishing characteristics, they have an existence apart from that
of any particular employer belonging to the plan.
Another crucial issue would be the nature of the obligation that
t.he PG-F should assume in respect of the benefits covered by the guar-
anty. One concept would call for the PG-F to assure ultimate. ayment
of all guaranteed benefits, irrespective of the amount, source, or cause
of any asset deficiency that might exist upon occurrence of the con-
tingency insured against. In theory, this concept could be applied
without any ma.ndated standa.rds of funding, but it would be far more
practicable if it were bulwarked by an enforceable requirement that
the covered benefits be funded in accordance with minimum standards
concerned with actuarial assumptions, actuarial cost. methods, and the
period of time allowed for the attainment of a fully funded status.
The approach would be even more feasible-but even less palatable
to employers-if the sponsor of a terminated plan were made
primarily responsible for any insufficiency of assets, with the PG-F
being only contingently liable. Another concept would limit, the PG-F's
obligation to the completion of the employer's funding program for
covered benefits. without regard to the adequacy of the projected con-
tributions. In ot.her words, the guaranty would attach to the fvnding
commitment rather than the b e'ne fit commitment.
A number of questions are involved with respect. to the plans that
would be brought under a pension guaranty program. The first ques-
tion is whether particination in the program would be compulsory or
optional. If participation is to be compulsory, one must confront the
problem of what categories of plans can be forced to come under the
system. Other questions would relate. to the advisability of excluding
from coverage plans that (1') have been in operation less than a speci-
fied period of time. (2) have fewer than a stipulated number of partici-
pa.nts~ (3) can not meet reasonable. underwriting standards. and (4)
vohmtarily seek covera~e. Finally, there is the question whether multi-
employer plans should be. required t.o part-icina.te.
It would be necessary to define the classfication of accrued benefits
to be guaranteed. Various distinctions could be made. The program
PAGENO="0253"
OLD AGE INCOME ASSURANCE-PART v 245
might differentiate as to (1) future service versus past service bene-
fits, (2) vested versus nonvested benefits, (3) mandatorily vested bene-
fits versus voluntarily vested benefits, and (4) retirement versus an-
cillary benefits. Special rules would be needed to protect the PGF
against benefit increases and other plan changes that would enlarge
the unfunded liability. Moreover, it would be desirable to place a
dollar limit on the monthly benefits that would be guaranteed for any
one participant.
The implementation of the guaranty would involve: (1) determina-
tion of the dollar dimensions of the PGF's obligation, and (2) a deci-
sion as to the manner in which the guaranty would be carried out. If
the quarantor's obligation were to assure payment of all guaranteed
benefits, its obligation would be measured by the difference between
the actuarially computed value of the covered benefits less the value,
at book or market, of the assets considered to be available for the satis-
faction of such claims. It would be necessary to prescribe or recognize
rules for the allocation of assets as between guaranteed and nonguar-
anteed benefits. If, on the other hand, the guarantor's obligation were
to complete the ftmding program of the terminated plans, its liability
would be equivalent to the present value of the remaining payments.
The guarantor's obligation as to benej~ts could be discharged in a
number of ways each with its own advantages and disadvantages.
The funding agency could retain the assets allocable to the covered
benefits, meeting benefit claims as they come due until the assets are
exhausted, with the guarantor then assuming responsibility for pay-
ment of the remaining guaranteed benefits. Secondly, the funding
agency could pay that portion of each employee's total guaranteed
benefit that could be provided by the assets in its possession, with the
guarantor concurrently paying the remaining portion. Thirdly, the
guarantor might transfer to the funding agency the additional sums
actuarially estimated to be needed to pay guaranteed benefits, the fund-
ing agencies providing only investment and disbursement services.
Finally, the funding agency might transfer to the guarantor a sum
equal to the assets deemed to stand behind the guaranteed benefits,
with the guarantor assuming responsibility for the payment in full of
all covered benefits. This it could do by paying the benefits directly
to the claimants as they come due or by purchasing nonparticipating
annuities in the proper amount and form from individual life insurers
or a pool of insurers formed for that purpose. Any of the foregoing
approaches could be used, with modifications, to discharge a guaranty
expressed in terms of a funding objective.
The basic issue in the financing realm is whether the guaranty
fund would be supported by advance premiums, assessments, or a
combination of the two. The use of the advance premium approach
would necessitate estimates of future claims and the accumulation of
substantial reserves. The assessment method would avoid these com-
plications but would have offsetting disadvantages. Under both ap-
proaches, it would be necessary to establish a base against which to
levy premiums or assessments and to decide whether to create a number
of risk classifications. The need for reinsurance facilities would also
have to be considered under either approach.
PAGENO="0254"
246 OLD AGE INCOME ASSURANCE-PART V
A Minimum Program
A pension guaranty arrangement would be technically feasible if
certain conditions were satisfied and adequate safeguards were built
into the system. Some of the conditions and safeguards would involve
_regulathry controls that employers, unions, and other elements of the
pension establishment have in general opposed as being potentially
detrimental to the continued sound growth of the private pension
movement. They would also limit the scope of the arrangement to such
narrow bounds that the social objectives underlying the proposal
might be frustrated in large part.
Resolution of the fundamental question of whether a properly struc-
tured and delimited guaranty scheme should be established is beyond
the purview of this paper. If such a program should be deemed to be
in the public interest, it is suggested that it be structured initially
along the lines set forth hereafter, with the thought that extensions
and liberalizations could be introduced as experience with the system
indicates the wisdom of such action.
The program should be administered by a Federal agency with the
necessary enforcement powers and the authority to serve as residual
risk-bearer if circumstances demand it.
The guaranty should extend only to benefit claims arising out of
complete plan terminations, being further limited to those situations
in which the sponsoring firm goes out of business. The lack of pro-
tection for benefit rights in terminated plans of employers who con-
tinue in business should be rectified by requiring the employer to con-
tinue funding contributions in respect of the benefits that would be-
come the obligation of the guaranty fund in the event that the employer
should go out of business.
The fund should undertake to assure payment of all guaranteed
benefits, irrespective of the source of the asset deficiency. However, this
obligation should be protected by a legal requirement that all covered
plans be funded at a rate sufficient to meet the currently accruing cost
of all benefits (whether or not guaranteed) and to have all guaranteed
benefits fully funded within 20 years after the effective date of the
coverage. Firms that terminate their plans before completing this
funding objective would be expected to continue their funding pay-
ments until their funding commitment is fulfilled.
Participation in the program would be limited to "qualified" plans,
which would be compelled to come under the program as a condition
for qualification. Plans should be eligible for coverage only after they
have `been in operation for a minimum of 5 years, but there should
be no other underwriting requirements. Specifically, there should be
no minimum size requirement. Multiernployer plans should be ex-
pected to participate, subject to appropriate modifications in the
definition of the insured event and possibly the premium rate.
The guaranty should be limited to benefits that have vested under
the terms of the plan but the law should require both single-employer
and multiemployer plans to provide a minimum degree of vesting.
Vested benefits created through a retrospective liberalization of the
plan should not be eligible for the guaranty until 5 years after the
guaranty. There should be a limit on the amount of monthly income
that would be guaranteed in respect of any one individual, the amount
being defined in terms of payment at an age specified in the law.
PAGENO="0255"
OLD AGE INCOME ASSURANCE-PART V 247
Upon termination of a covered plan, the guarantor should take title
to the assets in possession of the funding agency assumed to be avail-
able for the satisfaction of the guaranteed benefits. It should then dis-
charge its obligation by the purchase of nonparticipating insurance
or annuity contracts from a pool of life insurers for the full amount of
guaranteed benefits. This would fix immediately and irrevocably the
amount of funds needed to underwrite the guaranty and, hence, the
amount of assets that would have to be transferred from the funding
agency. In order to minimize liquidation losses, the funding agency
should be permitted to spread the transfer of assets over a period of
time.
The guaranty system should be supported by contributions from
employers whose pension plans fall within the scope of the program,
with the objective of making the program self-supporting as to both
benefit obligations and administrative expenses. The primary source
of support should be annual premiums levied on the basis of the un-
funded accrued liability for guaranteed benefits. For the purpose of
determining the premium base, the actuarial liability of the accrued
benefits would be computed on the basis of annuity rates (reflecting
mortality, interest, and expense assumptions) provided by the guar-
anty fund. There should be provision for assessments, within stipu-
lated limits, to meet costs not covered by the regular premiums. The
guaranty fund should have borrowing authority sufficient to absorb
shortrun deficits and should be empowered to assume an appropriate
share of the total burden on a continuing basis if claims should reach
a level beyond that which could be supported by reasonable contribu-
tions from the participating firms.
PAGENO="0256"
AN ANALYSIS OF PROPOSALS FOR IMPROVING THE
FUNDING AND FINANCIAL MANAGEMENT OF PRIVATE
PENSION FUNDS
BY PETER 0. DIETz* and H. ROBERT B~n'rEr1I~, Jr.**
The recent report, "Old Age Income Assurance: An Outline of
Issues and Alternatives," prepared for the Joint Economic Committee
puts forth several suggestions which, if adopted, would greatly in-
fluence the funding and investment management of private pension
funds. These suggestions would affect the size and certainty of re-
tirement incomes for a large proportion of the population and the
employment costs of most corporations. The particular suggestions
we have reference to are those regarding removal of public incentives
for funding of private plans,' revision of funding requirements,2
government-sponsored reinsurance of plans,3 and regulation of fund
managers and their investment decisions.4 The purpose of this paper
is to summarize our reactions to these proposals recognizing that they
will undoubtedly be covered individually at some length in other
papers prepared for the committee. In concluding, we present what
we judge to be reasonable proposals for altering the present arrange-
ments for funding and fund management.
The report suggests that private pension plans be discouraged
through the elimination of tax deductibility for contributions and tax
exemption for fund income. This would certainly reduce the incen-
tive for establishing new private plans and expanding present ones.
The outcome would ultimately be to generate additional pressure for
expanded benefits and coverage under OASDI. Eventually we could
envision the withering away of the private pension system if tax
incentives were eliminated and the social security system enlarged.
The question of whether the Nation would be served best by a pre-
dominantly public retirement system or a mixed, public-private sys-
tem, as we have at present, is one too complex to be argued in depth
here. Our own preference is for the present system with some modi-
fications. This preference is influenced by (1) the high value we place
on the freedom to choose various levels and combinations of bene-
fits under private plans, (2) the belief that pension fund savings add
to total saving and thus enlarge the fund of capital necessary for
economic and social advancement, and (3) the conviction that pri-
mary reliance on decentralized investment decisions will lead to more
effective use of such capital. Assuming this preference for a vigorous
Associate professor of finance, Northwestern University.
** Associate professor of managerial economics, Northwestern University.
1 U.S. Congress, Joint Economic Committee, Subcommittee on Fiscal Policy, Old Age
laconic Assurance: Am Outline of Issues and Alternatives, U.S.G.P.O., 1906, pp. 5-7, 31.
Ibid., pp. 17-18.
`Ibid., pp. 29-30.
`Ibid., pp. 22, 30.
248
PAGENO="0257"
OLD AGE INCOME ASSURANCE-PART V 249
private component in a mixed system is held by the majority of those
concerned, there are still important considerations as to how the pres-
ent system might be improved through public policy.
I. FUNDING
A maj or question raised is whether or not present funding arrange-
ments for private plans should be changed. On the one hand some
observers propose an increase in funding requirements, while others
question the necessity for the current level of funding in the majority
of plans. Which of these views should national policy encourage? It
is true that there is little need to fund a tax supported plan such as
OASDI. The same thing might be said for private plans taken as a.
whole. Theoretically it is surely correct that there is little nee4 for
funding beyond a small liquidity reserve for plans sponsored by
growing industries and companies. Under such circumstances pensions
can be paid out of future earnings. On the other hand, declining in-
dustries and firms should have fully funded plans. Since there are
always declining industries and firms, if companies are to be treated
alike, national policy should require full funding of all plans for
employees to be protected properly. If companies are to be held to
different funding requirements, how would one go about deciding
which firms will grow and which decline? The U.S. economy is
dynamic, and, therefore, it is difficult to forecast with certainty which
companies and industries will grow for a 5- to 10-year period
ahead let alone for the period of a 30- or 40-year pension obligation.
Today most financial analysts and economists would undoubtedly
agree that the telephone-communications industry and the air-trans-
port industry are growing; therefore, no funding should be necessary.
Fifty years ago similar agreement would have concluded that fund-
ing was unnecessary in the traction, coal mining, railroad, and tele-
graph industries. Today the private traction industry is virtually
dead, employment in railroads is declining and a major company in
the telegraph industry faces financial problems because of the lack
of a funded pension plan. Funding protects retirement income of
workers several decades away, and it would be unwise to base a fund-
ing policy on the presently anticipated growth of individual firms or
industries. Therefore, all plans should be as fully funded as financial
resources permit unless there is a universal reinsurance program for
all liabilities. Such a reinsurance program, we believe, is undesirable
for reasons indicated below.
It has been argued that full funding leads to excess savings in the
economy and overly conservative investment policies. To suggest
that the economy is subject to oversavings is to take a narrow view.
Worldwide needs for capital are undoubtedly far in excess of savings.
The problem is not one of excess savings but rather one of developing
effective channels of investment. Indeed, it has been argued that the
growth of pension fund assets has provided a major impetus to the
development of U.S. capital markets.5 The need for worldwide in-
vestment is so great that the U.S. has had to put restrictions on the
~ For further development of this argument see: Sidney E. Itolfe, Capital Markets in
Europe, sponsored by Atlantic Community.
83-200-67-pt. 5-i7
PAGENO="0258"
250 OLD AGE INCOME ASSURANCE-PART V
movement of capital to limit balance-of-payments deficits. Although
it may be argued that foreign investment is too risky for pension
fund assets, growth in pension systems can release other assets for
international development.
There is the further evidence of capital shortage in the secular rise
in long-term interest rates over the last two decades. It is hardly neces-
sary to point out the tremendous capital needed in the future for the
profitable development of space, underwater technology, urban re-
newal, and transportation and communication facilities, to mention
just a few capital intensive areas of the American economy.
The question remains as to whether or not the funding of pension
plans will lead to more efficient allocation of capital than would occur
with a pension system financed primarily on a pay-as-you-go basis
with reinsurance. The funding of pension plans places retirement sav-
ings in the hands of financial institutions whereas in nonfunded plans
the savings are invested by the sponsoring corporations.6 Since finan-
cial institutions are free to invest in a full range of alternatives, aggre-
gate productivity of capital should be greater than if funds were
invested solely in the assets of the company sponsoring the plan. Tax-
deductible pension contributions a.nd tax exemption for fund earnings
foster the establishment of funded plans and thereby improve the
capital allocation process.
The argument that national economic goals are fostered through
financial intermediary channels rather than through direct corporate
savings can only be supported if the investment managers do a good
job of allocating capital. Some indication of this is given in table 1
which shows the portfolio composition of various types of pension
funds at the end of 1966. Corporate noninsured funds have been most
aggressive in terms of common stock investments while multi-
employer and union plans have been more aggressive in the mortgage
market. State and municipal funds historically invested primarily in
various Government obligations. Thus their record in terms of aggres-
sive capital investment has been poorest of all. Insured pension funds
are heavily committed to corporate bonds and mortgages due to legal
restrictions, but have been turning more to equities as these restrictions
are liberalized.~ If there is to be an indictment of pension fund invest-
ment, it is that too much emphasis has been placed on fixed-income
obligations whereas investment objectives indicate very little need for
fund liquidity. However, the record has been improving and, as will
be pointed out in the section on investment management, will continue
to improve if properly fostered.
Further evidence concerning the quality of investment manage-
ment and the productivity of pension fund assets is found in a pre-
liminary report of a research study of pension fund investment man-
agement by the Financial Executives Research Foundation.~ This re-
We are not arguing that corporations bear the cost of pensions. Under various circum-
stances, the costs may be shifted to employees, customers, stockholders, or the taxpaying
public. We are concerned here with who holds the assets from which future benefits will
be drawn. It seems unlikely that a company with an unfunded plan would pay higher wages
or charge lower prices than if it had a funded plan. Thus, no matter who ultimately bears
the cost of the benefits, the company with an unfunded plan should end up with more
resources to invest.
~ Insured pension funds are not shown in table since investments are not reported sepa-
rately from other insurance company assets.
8 Dietz, Peter 0., "Measuring Rate of Return on Pension Funds," Financial Ea,ecutive,
July 1966.
PAGENO="0259"
OLD AGE INCOME ASSURANCE-'---?ART V 251
`TABLE 1.-PORTFOLIO COMPOSITION OF NONINSURED PENSION FUNDS, 1966 (PERCENTAGE OF BOOK VALUE)
Dec.
31, 1966
Fiscal year
1966, State
Corporate
~-
Nonprofit and
multiemployer
and local
government
Total assets (billions of dollars)
$58. 7
$5. 8
$35. 2
Cash and deposits
Percent
1.1
Percent
4.0
Percent
U.S. Government securities
State and local governments
Corporate bonds
Corporate stock
Mortgages
Other assets
3. 7
0
38.2
46. 7
4.9
5.4
7. 7
0
37. 1
30. 1
16.3
4. 8
0.9
19. 9
7. 1
50.3
5. 1
11.6
5, 1
Total
100.0
100.0
100.0
Source: U.S. Securities and Exchange Commission, "Private Noninsured Pension Funds, 1966," release No. 2219. Bu-
reau of the Census, U.S. Department of Commerce.
port shows that average ani1ual rates of return for 20 corporate
noninsured funds ranged from 5.4 to 9.6 percent with the median re-
turn 6.7 percent for the 5-year period, 1959-63. If one looks at the an-
nual returns of common stock investments in these same portfolios
the rates range from 8.1 to 15.3 percent with a median of 9.4 percei4.
The same report indicates median annual common stock returns of
22.3 percent for the 5-year period, 1954-58, and a median stock return
of 16.1 percent for the period 1954-63.
II. REINSURANCE
Among the most interesting suggestions now before the Congress of
the United States are various proposals for reinsuring pension plans.
Although a complete analysis of reinsurance is beyond the scope of
this paper, we would like to examine the effect on pension fund in-
vestment of various forms reinsurance might logically be expected to
take.
It is vital to recognize that on a total basis reinsurance premiums
cannot pay pension benefits. Benefits can only be paid out of the assets
of the plans. Therefore, in the long run, funding provides the best
method for covering the costs of pensions from the employers point of
view as well as providing greater security for the employee. The now
well-known Studebaker case is usually cited as an example of the need
for reinsuranc.e. Had the UAW and the employer insisted on more
adequate funding in relation to promised benefits, the Studebaker
difficulties would have been less serious. Undoubtedly, one of the
great dangers of the reinsurance concept is that it may be relied on to
provide security for an ill-conceived or deliberately underfunded pen-
sion plan. For example, some companies may consider the funding of a
given level of benefits as more costly than the premiums for reinsur-
ance of the same benefits: and, therefore, minimize funding. To the ex-
tent that reinsurance is used in place of sound funding, the security
of private pension promises will be weakened.
The practical problems involved in the development of a reinsurance
system are many. Two of the most difficult, the problems of which
risks are to be insured and the question of an equitable premium rate
structure are directly related to investment management.
PAGENO="0260"
252 OLD AGE INCOME ASSURANCE-PART V
To be insurable, a risk must encompass a large number of homo-
geneous risks, the incidence of loss must be spread randomly over
time and probable total losses should be calculable. ~ Although there
are a large number of pension plans it is not clear that the risks as-
sociated with each are homogenous. There are distinct differences in
the degree of probable loss among plans. This will cause different, al-
though not msurrnountable, problems in determining an adequate
and equitable premium structure. A much more serious problem, how-
ever, is that pension losses, whether either benefits are guaranteed or
plan assets are insured, are not likely to be random through time.
That is losses would tend to cluster in times of economic stress.
An adequate insurance program where the risks being covered are
neither homogenous nor random will almost surely have to depend on
government support. Since risks are not homogenous it will be nec-
essary to require all plans to participate in order to avoid the prob-
lem of adverse selection. Even with a premium structure supported
by low risk plans, the Government will have to be prepared to financ~
the plans in case of catastrophic losses.
It is generally agreed that any feasible reinsurance plan would in-
sure liabilities and not assets. However, the question of providing
protection against the contingency that assets in the pension fund will
decline in value has also been raised. To insure either real or paper
assets against value erosion is akin to insuring the value of the assets
of all firms in the economy. To insure the assets of all productive en-
terprises in the economy against dynamic risk is unthinkable since
no one could ever determine the potential losses.
One system of insurance based on assets might be an arrangement
which discounted assets to reflect the probability of decline in value.
For example, assets held in cash or Government securities might
be assumed to have no discount. Corporate bonds might have a 5 per-
cent discount, while corporate stocks or real estate might be dis-
counted at 25 percent. The problem of setting the discount should be
enough to discourage such a system. But the potential effect of such
a system on the capital markets could be unwelcome. The insuring
agency would be placed in the position of de facto regulation of in-
vestment policy for all pension funds. This is the very antithesis of
what a free investment market demands. For this reason, section 402
of the Javits bill, "Pension and Employee Benefit Act of 1967," deal-
ing with foreign securities is too restrictive. Since we find no way of
insuring assets we would conclude that if a plan had assets equal to
the vested liabilities of the fund1 no insurance would be necessary.
We turn now to the question of insuring the liabilities of the plan.
As we. have just indicated, there is no need to reinsure liabilities
which are covered by assets. Thus only unfunded liabilities need be
insured. The category "unfunded liabilities" is often vague and in-
chides liabilities which may never have to he paid. Co~itra~t this with
the deposit liabilities of commercial banks. and gavinp~s and loan
associatioi~e which are always precisely determinable, and, therefore,
more readily insurable. Therefore1 it is su~g~sted that th~~ maximum
insurance coverage would be confined to unfunded vested liabilities.
See Dan M. McGill, "Guaranty Fund for Private Pension Obligations," p. i99 in this
compendium.
PAGENO="0261"
OLD AGE INCOME ASSURANCE-PART V 25~
The term vested liabilities is more readily definable by actuaries than'
unfunded liabilities. However, even when attempting to define un-
funded vested liabilities, a determination must be made of the assets
in the fund. Here, two choices are readily apparent: book value of
assets or market value.
Given this choice, market values are clearly preferable. Book value
may have no relationship to the economic value of the fund. At the"
end of 1966 the market value of the assets of private noninsured pen~
sion funds was $6.4 billion (or 10 percent) greater than book value.
To use book value would generally punish those trusts that had done~
the most effective investment job. Moreover, only admitting boGk:
values would lead to putting a premium on fixed-income rather thaii
equity investment. This would tend to reduce return on investment
over the long run and increase pension costs.
The choice of market values is not without problems. In the first
place, market values tend to fluctuate so that the amount of insurance
to cover unfunded vested liabilities would fluctuate, and generally be
greatest when losses are high. Secondly, valuing assets at market
values could lead to unnecessary investment speculation by unprin-
cipled fund managers. For example, assume a pension plan in which
vested liabilities are far greater than admitted assets. The plan could
be fully insured by the use of reinsurance schemes. Meanwhile risky
investments could be made in the expectation of rapidly increasing
asset values without the normal contributions needed to properly fund
the plan. If the investments worked out, this would be fine, but if not,
the reinsurer is left holding the bag.
Although the concept of reinsurance might be politically attractive,
it introduces unnecessary economic problems in the private pension
field. First, the problem of preventing potential abuses has been
pointed out. Secondly, there is the problem of defining liabilities and
assets. Third is the potential adverse effect such a system would have
on the investment practices of pension trust. Finally, there is the
question of setting equitable premiums. More than likely, the best
financed plans will need little or no insurance. New and poorly financed
plans would need the most reinsurance. These are exactly the plans
that need to be insured, so the system will run great risks of adverse
selection and could easily become bankrupt unless soundly financed
plans are inequitably "taxed" to pay for defaults. In the absence of
irrefutable evidence that rein~urance is necessary and practical, na-
tional policy demands that we strengthen funding requirements rather
than adopt a reinsurance system.
III. THE PRESENT FINANCIAL MANAGEMENT SYSTEM
In this discussion of financial management we assume that in the
foreseeable future there will be no major changes in the financing of
private pension plans. That is, like today, each will invest its own
funds with only general rules imposed as to the disclosure of assets.
In this sense then we strongly support the conclusions of the Presi-
dent's Committee on Corporate Pension Funds that, "in view of the
wide legitimate differences regarding the most advantageous balance
of retirement funds investments, the committee does not believe it
PAGENO="0262"
~254 OLD AGE INCOME ASSURANCE-PART V
~would be desirable on the basis of evidence to date to require con-
formity to a prescribed rule with respect to the proportion of stocks
to other investments." 10
The question then which might be legitimately asked is whether
or not the present system of private investment provides sufficient
safeguards for employee-beneficiaries, and whether the invested assets
are producing returns which reflect efficient management.
The great majority of plans are financed by employer and employee
contributions which are invested by one or more third-party fiduci-
aries. This third party is variously an insurance company, bank trustee,
or investment counselor. Thus a dual system has been created. This
arrangement has as its major advantage the fact that the fiduciary's
first responsibility is the preserving the corpus of the fund. That is,
the fiduciary represents the beneficiary in assuring the financial in-
tegrity of the fund. On the other hand, the fund sponsor has the
responsibility for selecting the trustee. This gives the sponsor, whether
it be a corporation or joint union-management. board, the right to
measure investment results and the attendant right to change the
trustee if the investment, results are unsatisfactory. The dual system
thus provides the necessary balance between the sponsor's desire to
lower costs of the plan by increasing return on assets and the need
to preserve the assets in the fund. The record shows that where de-
faults on pension promises have occurred it is clue to either made-
qua~e funding or malfeasance, not poor investment management. Such
a system of dual control puts a premium on high rates of return which
can be used either to reduce contributions or to increase benefits or
both, without incurring excessive risks. The pressure on fiduciaries to
earn adequate rates of return will assure the efficient investment of
the economy's peitsion fund saving.
Another major advantage of the private system as it is presently
structured is that each fund is an entity of its own. This provides
flexibility to tailor each fund's investments to its own needs that no
other system could provide." While it is a well-known fact that most
pension funds need little liquidity because contributors are greater
than benefit payments, this is not so for mature or declining funds.
The amount of liquidity risk which can be taken depends on the spread
between contributions and benefit payments. This clearly differs among
funds.
The question of wha.t portion of the fund can be invested in variable-
income assets is still a. matter on which experts may disagree. How-
ever, one approach which may be used is to invest that portion of the
assets needed to cover earned benefits of retired employees with sec-
ondary reserves invested in fixed income assets. The remainder could
then be invested in riskier assets. Clearly these characteristics would
differ from fund to fund. Since liabilities can cha.nge as plans mature
or alter benefit provisions, a financing mechanism is needed which
can reflect these changes. Such changes can be made much more quickly
10 "Public Policy and Private Pension Programs," President's Committee on Corporate
Pension Funds and Other Private Retirement and Welfare Programs, U.S. Government
Printing Office, January 1965, p. XV.
1~ This is not precisely true for all insured plans~ However, recent developments leading
to segregation of pension fund from other Insurance company assets make this flexibility
more nearly attainable.
PAGENO="0263"
OLD AGE INCOME ASSURANCE-PART v 255
on a single fund basis than could ever be accomplished under a uni-
versal system.12
The system of private investment will work and improve only as
long as techniques for measuring investment performance are ade-
quate. The sponsor must have a fair and accurate method for deter-
mining investment excellence. Progress in the field of performance
measurement has been rapid in recent years.13 Work in this area is
continuing under the offices of such noteworthy financial groups as
the Financial Executives Research Foundation and the. Association
of Bank Audit Control & Operation. These groups are concerned not
only with measures of return but with improving methods for measur-
ing risk as well. As these extensive research efforts are concluded
the ability of sponsors to measure results and for trustees to appraise
their own performance will improve. The result is bound to increase
competition among fiduciaries to improve investment practice and
provide superior investment management.
IV. RECOMMENDATIONS
1. We have pointed out that funding is necessary to protect the
rights of individual employees. In this regard, we fully support the
funding recommendations of the 1965 report of the President's Com-
mittee. However, we would go further and suggest that the maximum
time to amortize unfunded liabilities be reduced to 20 years for plans
over 5 years old and 25 years for plans under 5 years of age. This
will somewhat reduce funding requirements for new plans where costs
may be excessively high. Any new past service costs as a result of
benefit increases should be amortized in not more than 20 years for
all plans. We prefer to see less benefit promises and more assurance
that those that are promised are paid.'4 This is entirely in keeping with
the concept of private pension plans as supplementary to the mini-
mum retirement benefits provided by OASDI. Furthermore, we highly
recommend that the present minimum funding period imposed by the
Internal Revenue Service be dropped. We see no economic justification
for restricting contributions for the funding of past service benefits. To
avoid the capricious use of these contributions to reduce taxes, a com-
pany making payments in any given year in excess of the required
4 or 5 percent reduction of unfunded liabilities must reduce any re-
maining portion of these liabilities in subsequent years at the 4 or S
percent minimum. This would eliminate the practice of making large
contributions in some years followed by no contributions for several
years.
2. We have shown that the development of a dual management
system provides the necessary balance between return on investment
aud safety. It is strongly urged that all plans be managed in this
manner and it is recommended that all new plans be placed under
dual management to be qualified for IRS pension plan treatment.
~-` For a further discussion of differing Investment goals of pension funds, see: Diets,
Peter 0., Pension Funds: Measuring Investment Performance (The Free Press, New York,
19G6), chapter III; and Sieff, John A., "Construction of a Retirement Fund Portfolio,"
Financial Analysts Journal, July-August 196~.
13 For example: Diets, Peter 0., op. cit.
14 We are not arguing that all plans are underfunded, but only that some plans need to
raise their level of funding. To Improve the certainty of benefit promises, higher minimum
funding requirements are an alternative to reinsurance.
PAGENO="0264"
256 OLD AGE INCOME ASSURANCE-PART V
However, it i.s a well-known fact that some of the oldest and best
managed plans are company run. To protect these plans a grand-
father clause to this provision is recommended. In these special cases
moderate investment restrictions may be desirable. For example, only
4 or 5 percent of assets should be invested in any one security;
liquidity reserves for all accrued liabilities for those retired under
the plan may be desirable. The remainder of the fund might be
invested in any assets subject only to the 4- or 5-percent rule. These
moderately restrictive rules would be sufficient to protect employees
in these special cases and at the same time impose no investment
penalty on most of the so-called company or union run plans.
3. As a further safeguard of employee interests, investments in
securities of the sponsor or trustee of a pension plan (or a profit-
sharing plan which is intended primarily to finance pension benefits)
should not be permitted. Although some such investments have been
successful in isolated cases in the past, there is no economic justifica-
tion for them. Such investments are always subject to potential abuse
and the possibilities of self-dealing. This provision would include any
securities and/or real estate and should apply to all plans whether
company or union sponsored. Although the Federal Government
would probably have no jurisdiction over State and municipal plans,
the same principles should apply and they should refrain from pur-
chasing securities of their own taxing districts.
There is one and only one possible exception to this provision.
That is the unusual case of a company that for some reason does not.
have cash to meet funding requirements. In such cases companies
have been known to contribute securities of some form. Although we
do not believe such contributions are in either the company's or the
employee's best interest, it is better than no contribution at all. In
order to maintain qualified status such securities should be liquidated
in not more than 5 years, and if a reinsurance scheme were to be
developed, in no case should such securities be included as admitted
assets.
4. As another measure for improving the effectiveness of the dual
system of sponsor-trustee control, we would urge that each fund he
required to report annually to the Department of Labor a complete
listing of its security holdings. This would represent only minimum
interference with the carrying out of investment programs while pro-
viding the necessary data for monitoring investment performance.
Competition among investment managers would be encouraged there-
by, and this would enhance the operation of the system rather than
detract from it, as some have suggested. Experience with the New
York State Disclosure Act which requires such reporting and the
voluntary disclosures of some funds indicate that this places no undue
burden on the fund or trustees. This information would also serve to
improve the level of debate regarding the potential dangers of con-
centrating control over corporate enterprises in the hands of trustees.
These recommendations have been made with the expectation that
they will strengthen the private pension plan movement by increasing
the security of employees covered under these plans and by encourag-
ing a free and dynamic system of pension fund investment which will
reduce pension costs and be reasonably free of any possibilities of
self-dealing or malfeasance.
PAGENO="0265"
CORPORATE FIDUCIARIES OF EMPLOYEE BENEFIT
FUNDS
BY C. WADSWORTH FARNUM*
AN OUTLINE OF RESPONSIBILITIES AND SUPERVISION
It is generally acknowledged that the banks in the United States
have a very great responsibility for the management and safeguarding
of pension and profit-sharing funds under private, tax-qualified plans..
Banks serve as trustee for more than two-thirds of all accumulated
reserves under such plans amounting to more than $70 billion. We'
respectfully submit that an examination of the legal responsibilities.
and the safeguards to protect the interests of beneficiaries is appropri-
ate to a study of "Old Age Income Assurance." Especially is this so
in the light of the number of proposals befOre Congress concerning
audits, disclosures, financial statements, and standards of fiduciary
conduct relating to trust funds. This statement is intended as a broad,
general outline of how pension and profit sharing plans are handled
by bank trustees for the protection of interested parties.
INTERESTED PARTIES
Banks are considered quasi-public organizations. They are entrusted
with the handling of the funds of the public in a number of Very im-
portant capacities. The savings and checking account are the best
known of their services, and the larger banks have been handling per-
sonal trusts and estates for years. It was not surprising that at the very
outset employee benefit funds gravitated to the banks. We in Bankers
Trust Co. have been serving as trustee of employee benefit funds for
40 years, and other banks have also had a long arid successful experi-
ence as trustees of such funds.
We, and other banks that act as trustee, appreciate the importance of
the trust placed in us and recognize the broad social and economic areas
affected by these funds:
1. The trust fund is the primary source to which an employee
must look for his benefits upon retirement or other termination
of service. This is especially significant to an employee if his com-
pany should go out of business, or, if, because of financial cir-
cumstances or other reasons, the plan is discontinued.
*Senjor vice president, Bankers Trust Co., New York.
AUTHOR'S NOTE. This paper concentrates on an important aspect of "Old Age
Income Assurance System"-the aspect of the procedures and safeguards of
corporate fiduciaries in their handling of employee benefit funds. Because cor-
porate fiduciaries manage more than two-thirds of the accumulated reserves
under private employee benefit funds, we believe that a thorough understanding
of the corporate fiduciary and its function is a necessary part of the study.
257
PAGENO="0266"
258 OLD AGE INCOME ASSURANCE-PART V
2. The company has an interest in the soundness of fund invest-
ments since the results directly affect its employee-employer rela-
tions and an unfavorable experience can result in an increase in
future costs or in a reduction in benefits payable to employees.
3. There is a genuine public concern with what happens to
participants' benefits after retirement or other termination of
employment.
4. While the study is primarily one dealing with "Old Age
Income Assurance," the economic aspect of the large amounts of
savings that are being accumulated in these trusts is very impor-
tant to the country. Aside from the immediate effect, the con-
structive investment of these funds will influence future produc-
t.ivity of industry which is the ultimate source from which benefit
payments will be made to employees under these plans. (Papers
will be presented by other authors which will discuss this subject
fully.)
5. In multiemployer-union plans, a board of trustees (rep-
resentatives of employers and the union) has the direct respon-
sibility for supervising the plan. If a bank is used as an ancillary
trustee for handling funds of the plan, the relationship between
the bank and the board of trustees is similar to that which exists
between the bank and the company in single employer plans.
(In the interest of brevity, the discussion is confined to the bank-
company relationship in single employer plans.)
\\Te believe that the extent of existing governmental, legal and
internal safeguards of pension and profit sharing funds held by bank
trustees for the protection of these varied interests should be seriously
considered in any new study of the need for new legislation.
FIDUCIARY RESPONSIBILITY
In the report to the President on private employee retirement plans,
the Committee stated that "the general standards of conduct for any
trustee have been long established by law and custom. These include
the degree of prudence that must be exercised in investing the funds
of others. Similarly, transactions that unfairly benefit other parties
at the expense of the trust constitute unfaithfulness on the part of
the trustee and may provide a ground for legal proceedings under
State laws by the employer-grantor against the trustee."
The time-tested standard is that a trustee is required to employ
such diligence and such prudence is in the care and management of
trust property as in general prudent men of discretion and intelli-
gence employ in their own affairs. A bank trustee may in some impor-
tant respects be held to an even higher degree of care since it holds
itself out to be an expert and because it is better equipped than the
ordinary man. (Scott on Trusts, sec. 174-1.)
The. trustee has an undivided loyalty to the trust. The typical em-
ployee benefit trust is a dedicated fund to be held and managed for the
benefit of employees and their beneficiaries, and the trustee is required
to measure each action taken in the trust in terms of whether it serves
this dedicated purpose. Section 1.401-1 (a) (3) (iv) of the Treasury
Regulations provides the following in relation to employee benefit
trusts which seek the tax advantages of a qualified trust:
PAGENO="0267"
OLD AGE INCOME ASSURANCE-PART v 259
"It must be impossible under the trust instrument at any time be-
fore the satisfaction of all liabilities with respect to employees and
their beneficiaries under the trust, for any part of the corpus or in-
come to be used for, or diverted to, purposes other than for the exclu-
sive benefit of the employees or their beneficiaries."
A specific provision incorporating this language or similar language
will be found in the trust agreement of all qualified employee benefit
trusts, and the bank on becoming trustee accepts the responsibility for
carrying out this provision in relation to its management of the trust
fund. The consequences for failure to comply with this requirement
can be considerable. It can result in a breach of trust with potentially
severe penalties. It can also result in loss of the tax-exempt status
which may be costly to the company and in some plans may reduce
the benefits accruing to employees. Thus, it is in the self-interest of
the trustee, the company and the employee to make very sure that this
requirement is fulfilled.
ACCOUNTABILITY TO THE COMPANY
In the typical employee benefit trust, the interest of the company
is closely allied to the interest of the trust and the employees. A good
employee-employer relationship is valuable to the well being of the
company, and employee benefit trusts have become an important part
of this relationship. As mentioned previously, the company is con-
cerned with the soundness of fund investments and the good manage-
ment of the trust.
In our experience with employee benefit trusts and in the experience
of other banks, the company is taking an increasing interest in check-
mg, auditing and appraising the work of the trustee:
I. It has become universal practice for the banks to give the
company a statement at each month end of all receipts, disburse-
ments, investment changes, and other transactions in employee
benefit trusts during the month.
2. The bank also renders a formal annual accounting to the
company after each year end covering all of its activities during
the year. This annual accounting is usually the basis of the finan-
cial statements on the trust which are made available to employees
and which are filed annually with the Internal Revenue Service,
the Department of Labor and other governmental agencies.
3. The bank's records of an employee benefit trust are open to
examination by the company and its auditors at all times.
4. As a result of various research projects, accepted methods
are being established to measure the investment performance of
employee benefit trusts.
In the typical employee benefit trust, the company has reserved the
right to remove the trustee and to appoint a successor. Most companies
recognize that a very real responsibility goes along with this right, and
they have internal procedures for following the transactions of the
trustee currently, they require that their independent auditors review
the records of the trust, they meet with their trustee periodically to
be informed on the. investment policies of the trustee, and they require
their management to report to their board of directors on the trust.
PAGENO="0268"
260 OLD AGE INCOME ASSURANCE-PART V
This is a form of supervision which is thorough and at the same time
flexible, and is far superior to anything that could be achieved through
legislative action.
INTERNAL AUDITING
The Association for Bank Audit, Control & Operation, of which
most banks are members, sets forth in a statement of principles of in-
ternal auditing that "a major obligation of a bank's board of directors
is the establishment of proper internal controls to protect its cus-
tomers, stockholders, directors and staff."
Internal auditing has as its basic purpose the prevention and detec-
tion of loss. The auditor's responsibility flows directly to the manage-
ment to whom he makes regular reports. Significant in the audit pro-
grain set forth by the association for pension and profit sharing trust
functions are the following:
(1) Verification of authority for action taken under the trust
instrument.
(2) Compliance with applicable statutes and regulations.
(3) Determination that assets are adequately safeguarded and
properly presented in financial reports.
(4) Determination that liabilities are completely disclosed and
any pending litigation affecting trust accounts reviewed.
(5) Audit of trust income, expenses and acquisitions and dis-~
posal of assets.
(6) Evaluation of insurance coverage of trust assets.
In many banks the bylaws requires that the auditor appear per-
sonally before the board of directors or its examination committee
to report on the major findings of the auditing program. This would
include any violation of trust statutes and regulations.
States may require examination by internal auditors accountable
to the board of directors. As an illustration, in New York, section 122
of the banking law requires an annual examination of the bank by the
board of directors and section 123 requires that a report be made to
the board of directors of the bank after completion of the examina-
tion and a copy filed with the superintendent of banks within 10 days
after such meeting. Monetary penalties are levied against the banks
for each day's delay in ffling reports of examination. The superin-
tendent may also require the banks to employ independent auditors.
As a further protection to trust funds, many banks engage completely
independent outside auditing firms to review and supplement the
work of the internal auditor.
COMPTROLLER OF raE CURRENCY
Regulation 9 issued by the Comptroller of the Currency enumerates
in considerable detail the fiduciary powers of national banks and col-
lective investment funds. The responsibility for the proper exercise
of fiduciary powers is placed in the board of directors of the bank.
All matters relating thereto, including the determination of policies,
the investment and disposition of property held in a fiduciary respon-
sibility, and the direction and review of the actions of all officers and
employees in the exercise of its fiduciary duties are the responsibility
of the board.
PAGENO="0269"
OLD AGE INCOME ASSURANCE-PART V 261
The board must insure that trust accounts are periodically reviewed
to determine the advisability of retaining or disposing of assets. The
Comptroller requires that the trust records be kept separate and dis-
tinct from other records of the bank. A record must be kept of all
pending litigation in connection with the exercise of fiduciary powers.
A committee of directors, excluding officers of the bank, must cause
suitable audits at least once during each year to be made by auditors
responsible only to the board. The purpose is to ascertain whether the
trust department has been administered in accordance with law, the
Comptroller's regulations, and sound fiduciary principles. Among
other things, the regulations set forth powers of banks relating to
investments, self-dealing, custody of investments, and collective
investments.
FEDERAL RESERVE
All member banks are subject to examination by the Federal Reserve
examining staff. The Federal Reserve assists the Comptroller of the
Currency in carrying out examinations in accordance with the pro-
visions of Regulation 9. A concept of the extent of the examination is
set forth in the following statement:
"The scope of the examinations of pension trusts held by banks
includes a survey of overall policies, practices and procedures in con-
nection with new-business solicitation and acceptance of accounts,
legal opinions and other documentation, investments, fees and internal
controls and audits. In addition, individual pension trusts are reviewed
as to investments and administration. Particular attention is given to
unusual investments, investment concentrations, and investments in
stock or obligations of the employer corporation. Investments held
are scrutinized to determine conformity to the provisions of the gov-
erning trust instruments and the provisions of the Internal Revenue
Code and regulations of the Treasury Department. In the case of
investments in securities of the employer, the examiner ascertains that
proper disclosures have been made to the Treasury Department."
STATE BANKING DEPARTMENTS
In addition to internal and Federal audits, State banking depart-
ments examine trust departments of State banks. For instance, in
New York State a special trust examining staff of the State banking
department makes a comprehensive examination each year of the ad-
ministrative and operating policies, procedures and acts of all divi-
sions of trust departments. The usual scope of audit functions involv-
ing principal, income and expenses is covered. Further, the examina-
tion by State examiners includes, among other items, the following:
(1) Investigation of matters involving ineligible investments,
self-dealing, `holdings of stock in close corporations, etc.
(2) Verification that investment reviews are made by the board
of directors and the recording of minutes for each trust fund.
(3) Verification of any objections to filing of trustees' reports.
(4) A check on any threatened litigation against the bank
based on its fiduciary activities.
(5) Verifications of commissions charged to trust.
PAGENO="0270"
262 OLD AGE INCOME ASSURANCE-PART V
There is i~o question that State and Federal examiners are diligent
in their audits; searching questions are asked and written answers
required of responsible officials of the bank. In a large bank, trust
examinations will extend over a period of months. Naturally, any
breach of trust, disloyalty to beneficiaries, self-dealing, improper
delegation of responsibility, failure to keep and render proper ac-
counts, failure to control property and many other irregularities
which the audit might disclose would subject the bank to severe
penalties.
TREASURY REGULATIONS
Section 6033 of the Income Tax Regulations requires the bank
trustee to file an annual return with the District Director of Internal
Revenue. Among the data made available are detailed answers to
questions relating to any transactions concerning a company's possible
self-dealings with the trust assets covering borrowing, compensation,
purchase and sale of company securities, and delivery of assets to the
company. Either the bank trustee or the company must furnish a
statement of receipts and disbursements and a balance sheet of the
trust funds. Copies of the trust agreement and all amendments are
filed with the Treasury Department.
DISCLOSURE ACT
The Welfare and Pension Plans Disclosure Act requires, under part
IV of Annual Form D-2, the submission of financial data for trust
funds. The fund data includes a statement of assets and liabilities,
statement of receipts and disbursements, supporting schedules for
specific receipts and disbursements and party-in-interest transactions.
A copy of the trust agreement is included among other documents
relating to pension and profit sharing plans. As required by the act,
the administrator of the plan must make the information available
for examination on request by a participant or beneficiary. The Secre-
tary of Labor is also required to make the information available for
public examination.
INVESTMENTS
Pension and profit sharing trusts make no distinction between
principal and income. Typically, the modern trust agreement gives
the bank trustee broad powers of investment. Irrespective of these
broad powers, the common law of prudence governs the actions of
the trustee.
Some companies, however, prefer to place investment restrictions
in the trust agreement. For example, the agreement may provide for
investments restricted to legals for fiduciaries or legals for life insur-
ance companies or a prohibition on company securities. Whatever the
restrictions may be, the trust funds deposited with a bank are pro-
tected against a breach of trust through elaborate internal and gov-
ernmental audits and controls. State laws protect the trust beneficiary.
For instance, in New York State, section 100-b of the banking law
provides that all investments by a bank shall be at its sole risk, and
the capital stock, property and effects of the bank shall be liable for
PAGENO="0271"
OLD AGE INCOME ASSURANCE-PART V 263
losses, unless the investments are proper and permitted in the trust
instrument. This section also prohibits any corporate fiduciary from
purchasing securities from itself.
In some instances, a company may choose to assume the responsi-
bility for the investment of the funds. Here the bank is under the
duty of following the instructions of the company. We believe that the
company, in exercising the investment function under a trust instru-
ment, assumes a fiduciary responsibility and its acts must be judged
by the same high standards as a bank. The officers and directors of
the company may find themselves personally liable, and action can
be brought by stockholders as well as employees. Similarly, invest-
ment advisers selected by a company to manage~ trust funds will also
find themselves governed by fiduciary standards. Where the bank does
not manage investments, nevertheless it continues to be responsible for
the proper accounting and safekeeping of the assets.
PAYMENTS OF BENEFITS
In the majority of the employee benefit trusts, the bank as trustee
makes the benefit payments directly to the employees or their benefi-
ciaries. In pension plans, the company usually certifies to the bank that
the employee is eligible for payments and determines the amount of
pension or other benefit to be paid. In profit sharing and savings plans,
the procedure is similar except that the bank frequently maintains the
record of the accounts of employees and makes the determination of
the amount payable. In those cases where the bank does not make pay-
ment of benefits, the bank transfers funds to the retirement committee
or other committee of the company, which makes payments. They con-
tinue to be fiduciary funds in the hands of the committee, and the com-
mittee, we believe, has the same responsibilities as a bank trustee has
in the proper distribution of the funds.
Upon receipt of the certification by the company, the bank assumes
the obligation for proper payment. If a mistake is made in paying (1)
the wrong amount, or (2) the wrong person, the bank is required, pur-
suant to long-standing trust practice, to reimburse the trust for any
unrecoverable loss. Also, the trust is protected against loss due to
forgeries.
SUMMARY
In summary, bank trustees are held to a high degree of fiduciary
conduct. Banks are closely controlled and governed by well-established
statutes and by internal, Federal and State audits, examinations, and
procedures. Significant penalties are imposed upon banks when they
commit a breach of trust or fail to comply with Government require-
ments. Financial statements of each trust fund are rendered to the
company, to the Treasury Department, and to the Labor Department.
In considering the need for additional legislation in the areas of dis-
closure, audits, and financial data, we respectfully suggest that exist-
ing controls of bank trustees be recognized.
(Th
PAGENO="0272"