ȱ
ȱ¢ǰȱǰȱȱȱ¢ȱ
ȱǯȱȱ
ȱȱȱ¢ȱ
¢ȱŞǰȱŘŖŖşȱ
ȱȱȱ
ŝȬśŝŖŖȱ
ǯǯȱ
řŖŝŖŞȱ
ȱȱȱ
Pr
epared for Members and Committees of Congress
ȱ¢ǰȱǰȱȱȱ¢ȱ
ȱ
¢ȱ
One issue that never seems far from the minds of policymakers is Social Security. At the heart of
the issue is the large shortfall of projected revenues needed to meet the mounting costs of the
system. For the moment, the amount of Social Security tax receipts exceeds the amount of
benefits being paid out. The Social Security trustees estimate that, beginning in 2017, the amount
of benefits being paid out will exceed tax collections. According to the trustees’ best estimate, the
trust fund will be exhausted in the year 2041.
Some have argued that because the Social Security trust fund is intended to meet rising future
costs of the program, its surplus should not be counted as contributing to official measures of the
budget surplus. With regard to current saving, however, it makes no difference whether the
surplus is credited to the trust fund or simply seen as financing current federal government
outlays (including Social Security benefits). Off-budget surpluses contribute to national saving in
exactly the same way as on-budget surpluses do. The additional saving they represent adds to the
national saving rate and allows current investment spending to be higher than it would otherwise
be.
With respect to household saving for retirement, how much is “enough” may be a subjective
matter. But, one standard might be whether accumulated wealth is sufficient to avoid a decline in
living standards upon retirement. A number of studies have found, however, that Americans may
not tend to save enough to avoid such a decline in their living standard.
Social Security may affect saving in several ways. It may reduce household saving as participants
pay some of their Social Security contributions by reducing what they otherwise would have set
aside. It reduces the risk associated with retirement planning and so may free participants to cut
precautionary saving. It may, however, encourage additional saving by making it possible to retire
earlier, thus giving participants a longer period of retirement to plan for. To the extent that Social
Security involves a transfer of income from workers to retirees, it tends to reduce household
saving by shifting resources from potentially high savers to those who save less.
Proposed reforms have different effects on saving. Those that would move toward a more fully
funded system would be likely to increase national saving, investment, and the size of the
economy in the future. Reforms that would partially “privatize” using individual accounts, might
tend to reduce national saving, unless contributions to those accounts were mandatory. Those that
invested Social Security funds in private sector assets would be unlikely to have any effect on
national saving. This report will be updated as economic events warrant.
ȱȱȱ
ȱ¢ǰȱǰȱȱȱ¢ȱ
ȱ
ȱ
Introduction ..................................................................................................................................... 1
Saving and the Economy ................................................................................................................. 2
Explaining Household Saving ......................................................................................................... 3
Life Cycle Saving...................................................................................................................... 3
Precautionary Saving ................................................................................................................ 3
Recent Trends in Household Saving ......................................................................................... 4
Social Security and Household Saving............................................................................................ 6
Saving and Social Security Reform................................................................................................. 8
Fully Funded vs. Pay As You Go .............................................................................................. 8
Defined Benefit vs. Defined Contribution ................................................................................ 9
Investing the Trust Fund in Private Securities ........................................................................ 10
ȱ
Figure 1. Personal Saving as a Percentage of Aftertax Personal Income ........................................ 4
ȱ
Author Contact Information ...........................................................................................................11
ȱȱȱ
ȱ¢ǰȱǰȱȱȱ¢ȱ
ȱ
ȱ
One issue of perennial concern to policymakers is Social Security. At the heart of the issue is the
large shortfall of projected revenues needed to meet the mounting costs of the system, beginning
in about 2017. Taxes on those currently in the workforce are credited to the trust fund and
benefits to retirees are debited from the trust fund. The balance of the fund itself consists of
Treasury securities.
At the moment, the amount of tax receipts exceeds the amount of benefits being paid out, and so
the balance in the trust fund is growing. The current “best estimate” of the Social Security
trustees is that beginning in 2017, the amount of benefits being paid out will exceed tax
collections. At that time, the trust fund will have a balance of $4.7 trillion credited to it.
According to the trustees’ best estimate, the trust fund will be exhausted in the year 2041.1
Because benefits are projected to exceed receipts in 2017, there is concern among many
policymakers that changes need to be made now. If steps are not taken now, it is argued, much
more drastic changes will be needed down the road.
The major function of Social Security is to provide a base upon which to secure the income of the
retired population. It does so by transferring income from the working population to those who
are no longer in the labor force because of retirement or disability. This is also known as an
intergenerational compact, by which those currently working support the retired population with
the expectation that future workers will, in turn, provide for their retirement benefits. It also
serves a social welfare function, by paying relatively more to retirees who were low-income
earners, and survivors and dependents.
From a macroeconomic perspective, however, what matters is the effect on national saving. In a
general sense, the economy is blind to the sources of saving. What matters is that saving, whether
from the household, business, or public sector is channeled into investments which increase the
capital stock, raise productivity and add to economic growth.
If individuals set aside substantial amounts during their working lives then the accumulated
wealth and their expected Social Security benefits may be sufficient to provide for their
retirement years. The more individuals save for their retirement, the higher their standard of
living will be when they retire. If individuals are not saving enough to provide for some minimal
standard of living in retirement, then increased saving in the public sector is one way of
increasing national resources out of which to fund retirement benefits in the future. The larger the
economy, the better able the nation will be to ensure a given minimum standard of living to all
retirees.2
1 The trustees publish three estimates using different assumptions about costs. They call the intermediate projection
their “best estimate.” This refers to both Old-Age and Survivors Insurance (OASI) and Disability Insurance (DI), which
collectively are referred to as OASDI. See the 2008 OASDI trustee’s report, at http://www.ssa.gov/OACT/TR/TR08/.
2 A larger economy will make it relatively easier to provide retirement benefits. In dealing with the long-range Social
Security funding problem, for example, a larger economy would ease any actions needed to fully fund benefit
obligations. Increasing federal government saving now reduces outstanding federal debt and reduces outlays now
devoted to financing the existing public sector debt.
ȱȱȱ
ŗȱ
ȱ¢ǰȱǰȱȱȱ¢ȱ
ȱ
This paper examines the determinants of household saving and how household saving may be
affected by Social Security. The potential effects of possible changes in Social Security on
household and national saving are also discussed.3
ȱȱȱ¢ȱ
By definition, saving is that proportion of income that is not consumed. Rather, it is made
available to sustain and increase the stock of productive capital. Combined with labor and
technological advances, it is this capital that contributes to the production of all the goods and
services that make up national output.
Reducing the share of income that is consumed and increasing the share that is saved, will tend to
raise the amount of capital available to the labor force. The more capital available to the
workforce, the greater the production of goods and services will be. In other words, by saving
more now, people will be able to consume more in the future.
From a macroeconomic perspective, it makes little if any difference where the saving comes
from. In terms of the collective resources of the nation, a dollar of saving means a dollar of
investment whether it is household, business, or public sector saving.
The measure that matters with respect to the federal government’s saving is the unified budget.
For some purposes, the budget is divided into two accounts; one is referred to as “on-budget” and
the other is “off-budget.” The so-called off-budget surplus consists almost entirely of Social
Security receipts and outlays.4
Some have argued that Social Security should not be counted as contributing to official measures
of the budget surplus. Keeping Social Security off budget is a procedural device by which
Congress signals that the Social Security program should be insulated from other operations of
the budget. “Lockbox” proposals go further in an attempt to deter the current and future
Congresses from using Social Security surpluses to justify spending increases for other federal
programs or for tax cuts.5
With regard to current saving, however, it makes no difference whether those revenues are
credited to the trust fund or simply seen as financing current federal government outlays
(including Social Security benefits).6 Without the excess revenues from Social Security taxes, the
deficit would be larger, and the federal government’s requirements for a share of the national
savings pool would be larger as well. Off-budget surpluses contribute to national saving in
exactly the same way as on-budget surpluses do. The additional saving they represent adds to the
national saving rate and allows current investment spending to be higher than it would otherwise
be.
3 For a more comprehensive discussion of Social Security reform proposals, see CRS Report RL31498, Social Security
Reform: Economic Issues, by Jane G. Gravelle and Marc Labonte.
4 The Postal Service is also included in the off-budget account.
5 Some proponents of accounting devices that isolate Social Security from the rest of the budget claim that this would
result in greater government saving. This assumes that spending and tax policy decisions depend on what happens to
the trust fund.
6 Neither would it matter with respect to Social Security deficits, currently projected to begin in 2017.
ȱȱȱ
Řȱ
ȱ¢ǰȱǰȱȱȱ¢ȱ
ȱ
¡ȱ
ȱȱ
Any examination of Social Security and its effects on households must begin with an explanation
of household saving behavior. What do households take into account when deciding whether, and
how much, to save?
ȱ¢ȱȱ
Most economists analyze household saving behavior using what is known as the life-cycle model
of consumer behavior. The life-cycle model begins with the basic assumption that most
individuals are not myopic but rather take their expected lifetimes into account when deciding
how much out of current income to save and how much to spend.
The life-cycle model assumes that individuals seek to avoid large fluctuations in their standard of
living over the course of their lifetimes. The model further takes as a given that individuals’
incomes tend to follow a predictable pattern over the course of their lifetimes. Typically, that
would mean relatively low levels of income during the initial years of work, increases in income
up to retirement, and then a drop in income during retirement.
If consumption followed the same pattern as income, it would make for substantial changes in
living standards over the course of a lifetime. Instead, consumers are presumed to vary the rate at
which they save out of income in order to dampen the effect of changes in income on
consumption. Thus, the typical pattern would be that individuals save relatively little in the early
stages of their working lives. Then, during peak earning years, saving rates tend to be higher in
order to accumulate wealth off of which to live during retirement, when saving tends to fall off
considerably.7
¢ȱȱ
There is a second consideration that may motivate household saving in addition to retirement.
While there may be a typical pattern to incomes, on average, over lifetimes, there is also a certain
amount of uncertainty associated with an individual’s income at a given time in the future. For
example, some incomes vary over the course of the business cycle, and from time to time people
may also experience episodes of either voluntary or involuntary unemployment. Thus, in addition
to serving as a buffer against lifetime income fluctuations, some fraction of saving may also act to
insure against the risks of shorter-term fluctuations in income. This kind of saving may also help
to insure against whatever risk might be associated with an individual’s pension plan. It is
generally referred to as precautionary saving.
Although this may not account for all the possible incentives households have to save, it is
enough of a framework to make some general observations about household saving behavior. For
example, a temporary decline in incomes will not necessarily have an effect on consumer
spending, or on an individual’s long-term saving rate, since households have already set funds
7 People with higher incomes tend to save more than those with lower incomes; not necessarily just because their
incomes are higher, but also because they are more likely to be in their peak earning years.
ȱȱȱ
řȱ
ȱ¢ǰȱǰȱȱȱ¢ȱ
ȱ
aside for just such a rainy day. But, any change in prospects that are likely to extend over an
entire lifetime might well affect saving behavior.
ȱȱȱ
ȱȱ
Social Security was created to help secure the economic condition of the retired population by
forcing them to save more now in exchange for expected future benefits. Concern remains,
however, that many are still not setting aside enough on their own in order to provide adequately
for retirement.8
It has often been noted that Americans save less than they used to, and that they save less than
most other industrialized nations. Those concerns were part of the motivation for the policies that
contributed to the elimination of the federal budget deficit briefly in the late 1990s.9 Personal
saving, as measured in the standard economic accounts, has been falling steadily since 1990.
Figure 1 shows personal saving as a percentage of aftertax personal income since 1970. The
personal saving rate has fallen steadily from more than 10% in the early 1980s to near zero since
2005.
Figure 1. Personal Saving as a Percentage of Aftertax Personal Income
16
14
12
10
t
8
6
rcen
e
P
4
2
0
-2
-4
0 72 74 76 8 80 82 84 6 88 90 92 4 96 98 00 2 04 06 08
197 19 19 19 197 19 19 19 198 19 19 19 199 19 19 20 200 20 20 20
Source: Department of Commerce, Bureau of Economic Analysis.
8 Social Security is not the only source of retirement income. In fact, in the national income and product accounts
published by the Department of Commerce, contributions to Social Security are not counted as part of household
saving.
9 Such as the omnibus Budget Reconciliation Act of 1993, which included both tax increases and spending cuts. There
has also been a reduction in defense spending relative to GDP. See Alberto Alesina, “The Political Economy of the
Budget Surplus in the United States,” Journal of Economic Perspectives, vol.14, no. 3, summer 2000, pp. 3-19.
ȱȱȱ
Śȱ
ȱ¢ǰȱǰȱȱȱ¢ȱ
ȱ
However, this measure of saving may give an exaggerated picture of the drop in saving. One of
the most marked characteristics of the economy during much of the 1990s was a dramatic rise in
equity prices. But, capital gains are not included in the measure of saving shown in Figure 1. A
measure of saving based on changes in household net worth would tell a different story. In fact,
much of the decline in measured savings, at least through 1999, may have been due to a large
increase in equity prices. Between 1990 and 1999, total household net worth more than doubled,
while the ratio of household net worth to aftertax income rose from 473% to 632%.10 The
increase in wealth resulting from the rise in equity prices may have led some to feel they did not
need to save as much.11
Between 1999 and 2003, however, equity prices fell by about one-third, and during that period,
the saving rate continued to decline. That household saving continued to be anemic even after the
stock market cooled suggests that there are other factors that need to be considered. One
candidate would seem to have been the boom in the housing market. Between the third quarters
of 1997 and 2007, the house price index published by the Office of Federal Housing Enterprise
Oversight (OFHEO) increased by more than 90%.
A number of studies have suggested that housing price appreciation may have had a significant
effect on household saving. Belsky and Prakken, for example, found that in the long run, the
effects on household saving of house and equity price variations were similar.12 They also found
that house price appreciation had a more immediate effect and that the effect of equity price
appreciation took longer to be fully reflected in the saving rate. The authors suggested that may
be because historically equity prices have been more volatile than house prices, and so
households may be more confident in the durability of house price gains. The authors also
indicated that the strong effect of the post-2000 boom in house prices may have been partly due
to the simultaneous decline in interest rates, which encouraged homeowners to refinance as well
as borrow. They left open the question of whether, in other circumstances, house price
appreciation would have the same effect on household saving.13
Although there are reasons to think that housing price appreciation might not be a substitute for
saving, the empirical studies, not quite amounting to a consensus, found evidence to suggest that
it might. At the moment, both house and equity prices are below recent peaks. That might make
the current low saving rate seem more of a problem.
With respect to household saving, how much is “enough” may be a subjective matter. One
standard might be whether accumulated wealth is sufficient to avoid a decline in living standards
upon retirement. The life cycle model discussed above assumes that individuals seek to avoid
substantial ups and downs in consumption over the course of a lifetime. It might be assumed,
then, that the goal of retirement saving is to avoid a significant drop in living standards after
retirement.
10 Figures are from the Board of Governors of the Federal Reserve System.
11 See CRS Report RL33168, Why is the Household Saving Rate So Low?, by Brian W. Cashell.
12 Eric Belsky and Joel Prakken, “Housing’s Impact on Wealth Accumulation, Wealth Distribution and Consumer
Spending,” National Association of Realtors National Center for Real Estate Research, 2004, 26 pp.
13 Whether or not house price appreciation might substitute for other forms of saving may depend on if there is a strong
bequest motive for saving. Those who save in order to leave a bequest to their children may desire to leave a larger
bequest if their children are expected to face higher house prices.
ȱȱȱ
śȱ
ȱ¢ǰȱǰȱȱȱ¢ȱ
ȱ
A number of studies have found, however, that Americans may not save enough to avoid such a
decline in their living standard. One, by Hamermesh, of consumption patterns over time found
that accumulated wealth, both private and through Social Security, was not sufficient to sustain
consumption in early retirement. This study found that, typically, households gradually reduced
their consumption spending within several years of having retired.14
Bernheim and Scholz, also found that Americans were not saving enough to prepare for
retirement.15 In particular, they found a distinct difference in saving behavior between those with
a college education and the rest of the population. In general, those households with a college
education were found to have saved enough to avoid a substantial cutback in consumption on
retiring, whereas those with less than a college education had not.16
In a separate study of household saving behavior, Bernheim concluded that the typical baby-
boom household was saving at about one-third the rate at which they would need to save in order
to continue to maintain their current standard of living into retirement.17
Engen, Gale, and Uccello make several interesting points regarding the adequacy of household
saving for retirement. First, they point out that the ups and downs of the stock market may have
little effect on many of those households that are not saving enough, since the ownership of
financial assets is heavily concentrated among those households likely to already be saving
enough. Second, with regard to what level of saving is considered adequate, it is important to
consider the large increase in the consumption of leisure when comparing consumption before
and after retirement. Those studies that do not account for the value of leisure time may
overestimate the extent of any decline in post-retirement consumption.18
ȱ¢ȱȱ
ȱȱ
Social Security was intended to provide a base upon which to secure the income of the retired
population. As such, it might have been expected to increase the national rate of saving by adding
to what households were already setting aside.
However, there is an important question as to whether or not Social Security is a close substitute
for personal saving. If it is, Social Security may lead individuals to save less than they would
have in its absence.
14 Daniel S. Hamermesh, Consumption During Retirement: The Missing Link in the Life Cycle, The Review of
Economics and Statistics, vol. LXVI, no. 1, February 1984, pp. 1-7.
15 B. Douglas Bernheim and John Karl Scholz, “Do Americans Save Too Little?,” Federal Reserve Bank of
Philadelphia Business Review, September-October 1993, pp. 3-20.
16 Income and education tend to be correlated, as are saving and income. Nonetheless, Bernheim and Scholz suggest
that one reason some save less than others is that they may not be fully aware of the importance of saving, and that
education might be effective in encouraging them to save more.
17 American Council for Capital Formation, Center for Policy Research, Special Report, “Do Households Appreciate
Their Financial Vulnerabilities? An Analysis of Actions, Perceptions, and Public Policy,” August 1994.
18 Eric M. Engen, William G. Gale, and Cori Uccello, Are Households Saving Adequately for Retirement? A Progress
Report on Three Projects, paper presented at the third annual conference of the Retirement Research Consortium, May
2001, p. 19.
ȱȱȱ
Ŝȱ
ȱ¢ǰȱǰȱȱȱ¢ȱ
ȱ
Consider the stereotypical saver described by the life cycle model discussed above. How would
the introduction of a program which required workers to contribute to a retirement fund for their
own eventual benefit affect their saving behavior? Whether and how much individual savings are
affected by the introduction of such a program would likely depend on the specific features of the
program. Suppose that the contributions yielded the same return as other forms of financial assets
households might otherwise buy with their saving. In that case, it might be reasonable to expect
that individuals would be indifferent between either saving directly or via the contributions to
their retirement fund. The introduction of such a program might simply cause individuals to
reduce other saving to offset the amount of their contribution.
Suppose, however, that when the program is introduced, retirement benefits are immediately
available to everyone who is qualified. This would be closer to a pay-as-you-go system. In this
case, those who were very near retirement would contribute relatively little to the program while
still receiving the full stream of benefits in retirement. Because those who immediately receive
benefits would have contributed little to the program there would be a significant transfer of
income from those still working to those in retirement.
Of these two groups, workers tend to save more than do those who are retired. The transfer of
income from relatively high savers to relatively low savers would tend to bring down the overall
household saving rate. In a strictly pay-as-you-go pension system there would be no offsetting
saving on the part of the public sector; all of the contributions would be distributed. Thus, the
introduction of a pure pay-as-you-go Social Security system would tend to reduce the national
saving rate.
Another way in which Social Security might influence saving was suggested by economist Martin
Feldstein. Feldstein argued that Social Security, or any pension for that matter, might lead people
to retire sooner than they otherwise would have. For one thing, once covered workers become
eligible for an annuity, their pay effectively drops by the amount of the annuity, since they are
only working for the difference between their earnings and what their annuity would be.
In an effort to measure the effect of Social Security on saving, Feldstein examined the effect of
Social Security wealth on personal saving using data from 1930 to 1992. Social Security wealth
was defined as the discounted present value of promised Social Security benefits. Feldstein found
that, for each dollar increase in Social Security wealth, personal savings fell by two or three
cents.19 Other studies on the effect of changes in wealth on household saving have found that for
each dollar increase in household wealth, saving out of current income falls by somewhere
between 1 cent and 7 cents.
If Social Security encourages workers to retire early, then there is a longer period of retirement to
save for. But, workers anticipating an earlier retirement might tend to save more on their own.
That aspect of Social Security might then tend to raise the personal saving rate. There are two
potentially offsetting effects—Social Security substituting for personal saving, and encouraging
longer retirement. The first effect tends to reduce personal saving, the second tends to raise it.
Social Security may also affect the precautionary motive for saving. Social Security, as it now
stands, is a defined benefit program. In other words, retirement benefits, although they are based
19 Martin Feldstein, “Social Security and Saving: New Time Series Evidence,” National Tax Journal, June 1996, vol.
49, no. 2, pp. 151-164.
ȱȱȱ
ŝȱ
ȱ¢ǰȱǰȱȱȱ¢ȱ
ȱ
on career earnings, are fixed in real terms upon retirement. After that, they do not vary and
continue as long as the beneficiary survives. In contrast, were individuals to provide entirely for
their own retirement, there would be several sources of risk. For example, there would be
uncertainty regarding how long a period of retirement would have to provided for, and there
would be some risk associated with those investments which make up individuals’ nest eggs.
Thus, Social Security may affect saving in several ways. It may reduce household saving as
participants pay some of their Social Security contributions by reducing what they otherwise
would have set aside in other investments. It reduces the risk associated with retirement planning
and so may free participants to cut precautionary saving. It may also encourage additional saving
by making it possible to retire earlier thus providing participants a longer period of retirement to
plan for. To the extent that Social Security involves a transfer of income from workers to retirees,
it tends to reduce total saving by shifting resources from high savers to relatively low savers.
ȱȱȱ¢ȱȱ
As it now stands, Social Security is partially funded. In 2017, benefit payments are projected to
exceed tax receipts and, if that happens, will have to be funded out of general revenues. Because
of that, some policymakers urged that changes need to be made. The argument is that whatever
costs there are to assuring future benefit payments and boosting the confidence of participants in
the program will be more easily borne if they are spread out over a long period of time rather than
put off until some inevitable day of reckoning.20
In the long run, from a national perspective, what matters is how much people save now. Whether
it is household saving, business saving, or a federal budget surplus, increased saving means
increased investment, a larger capital stock, and higher future living standards. The more that is
saved now, the larger the economy will be in the future. If Social Security is changed, those
changes could affect saving in one way or another. Using the basic model of life-cycle and
precautionary saving explained above it is possible to make a few relevant observations about the
potential effects of various kinds of Social Security reform proposals on saving.21
There are two broad kinds of reform possibilities that have been most discussed. One is a shift
towards a fully-funded plan. The second is a switch from the defined benefits that currently
characterize Social Security to one that is at least partly a defined contribution plan with variable
benefits. This second kind of proposal includes some of the suggestions that would privatize
some or all of the Social Security program. Another proposal that has been advanced would have
some of the trust fund invested in private securities such as corporate stock.
¢ȱȱǯȱ¢ȱȱȱ ȱ
If Social Security were fully funded, that would mean that each generation contributed enough to
fully provide for their benefits on retirement, and there would be no intergenerational transfers. A
20 In principle, the Social Security trust fund represents the obligation of the federal government to pay future benefits.
In practice, in an economic sense, it is current Social Security receipts and outlays that matter.
21 See Eric M. Engen and William G. Gale, “Effects of Social Security Reform on Private and National Saving,” in
Social Security Reform: Links to Saving, Investment, and Growth, Federal Reserve Bank of Boston, Conference Series
No. 41. June, 1997, pp. 103-142.
ȱȱȱ
Şȱ
ȱ¢ǰȱǰȱȱȱ¢ȱ
ȱ
pure pay-as-you-go system, on the other hand, would have no trust fund at all, and all Social
Security retirement benefits would be paid for by the current contributions of the working
population. In this case there would be a continuing transfer of income between generations.
Because Social Security is only partially funded, at some point in the future all benefits will be
transfers from the working population to retirees if no changes are made.
Switching to a more fully funded program would necessarily involve some combination of
increased contributions and reduced benefits. However, any increase in taxes, or cut in benefits,
might be partially offset by a reduction in other forms of household saving. A reduction in
benefits could also lead households to save less as they seek to maintain a constant level of
consumption given a cut in income. For those still working, a cut in prospective benefits might
encourage additional saving.
A shift towards a fully funded system would also tend to reduce the intergenerational
redistribution of income. A pure pay-as-you-go system takes income from workers, who tend to
be savers, and gives it to retirees who tend to save relatively little. In a fully funded system,
workers would finance their own retirement benefits. Shifting from a pay-as-you-go system to a
fully funded one might reduce the overall bias against saving which is due to the shift of income
from savers to dis-savers.
Shifting Social Security closer to a fully funded program might also increase confidence on the
part of participants that future benefits would be paid. This might serve to diminish the
precautionary incentive to save and tend to reduce household saving.
A shift toward a more fully funded system might lead to a reduction in measured household
saving, but it is unlikely that the reduction in household saving would offset the increase in public
sector saving. The net result is that such a shift would be likely to raise the national saving rate.
ȱȱǯȱȱȱ
Other proposals for Social Security reform involve at least a partial shift from a defined benefit
plan to a defined contribution plan. A defined benefit plan is one where the benefits are set in
advance, and while participants must contribute, their benefits do not depend on the performance
of those assets in which they are invested. Participants in a defined contribution retirement plan
contribute a set amount periodically into an account, and their ultimate retirement benefits depend
on the return on the investments held in those accounts.
Depending on the existing level of confidence in future benefits, a shift toward a defined
contribution plan might involve an increase in the perceived, or actual, risk faced by participants.
To some extent, future benefits would depend on the performance of those assets in which
contributions were invested. An increase in risk might lead households to save more for
precautionary reasons. A switch toward a defined contribution plan might not, however, involve a
great increase in perceived risk given the apparent skepticism among those working now as to
whether or not they will get their full Social Security benefits on retiring. If the assets in which
the defined contributions are invested yield a higher return, participants might have an incentive
to reduce other forms of saving.
Switching to a defined contribution plan, or partially privatizing Social Security, would reduce
the federal government surplus because money that had been collected as taxes would be invested
directly by individuals. The measured household saving rate would go up as households
ȱȱȱ
şȱ
ȱ¢ǰȱǰȱȱȱ¢ȱ
ȱ
themselves invested those funds which previously had been paid into the Social Security trust
fund. If these new contributions were mandatory, there would be no net effect on national saving
as the increase in household saving would offset the decrease in federal government saving.
If the defined contributions were not mandatory, however, it is possible that the household saving
would not rise enough to offset the decline in the federal government saving rate. Many would be
likely to spend at least a portion of the reduction in taxes and some, who might not ever plan to
retire might spend all of it. Unless the contributions were made mandatory, the net effect of a
switch in the direction of a defined contribution plan could reduce the national saving rate.
There could also be some indirect effects on saving. For example, there might be some increase in
the risk associated with retirement savings with a switch to a defined contribution plan which
could encourage additional precautionary saving.
ȱȱȱȱȱȱȱ
Another reform proposal that has been advanced is that some of the Social Security trust fund,
which currently consists exclusively of Treasury securities, be invested in private securities. Such
a change would have no effect on national saving.22
If the trust fund were to invest in private securities, corporate stock for example, it would have to
either sell some of the Treasury securities it now holds or it would buy stocks out of current tax
receipts, and the Treasury would have to find another market for any securities it would otherwise
issue to the trust fund. In any case, the supply of Treasury securities would increase, their prices
would tend to go down, and their yields would tend to go up. At the same time the demand for
corporate stock would increase, and the yield on those stocks would tend to fall. Ultimately the
public sector would own some private assets, and the private sector would hold a larger
proportion of public sector debt. The only change would be in how the public and private sectors
invested their money. There is no evidence indicating that households would increase, or reduce,
their saving simply as a result of a shift in relative rates of return on selected financial assets.
If trust fund assets were invested in private sector securities, which yielded a higher rate of return,
the trust fund might be made better off. But the improvement in the trust fund would come at the
expense of the rest of the federal government budget and those private investors that would
otherwise have purchased those assets. The borrowing costs of the federal government would rise
because of the increased supply of government securities. The income from capital of other
private investors would fall because the increase in demand for private sector securities would be
likely to reduce their rate of return.23
22 Under current rules, if the trust fund purchased private sector securities, those expenditures would be counted as
outlays in the unified budget. The increase in measured outlays would reduce the unified budget surplus. But, federal
government saving would not change. The Congressional Budget Office is considering changing they way they count
purchases of private securities in the budget to eliminate this inconsistency. Whether budget policy depends in any way
on particular accounting practices is an issue beyond the scope of this paper. For a discussion, see Congressional
Budget Office, Cost Estimate H.R. 4844, Railroad Retirement and Survivors’ Improvement Act of 2000.
23 That is, unless the yields on government bonds had to rise so much, and the yields on the private securities purchased
by the trust fund had to fall so much to allow the markets to clear that the shift resulted in no change in capital income
to either the private sector or the trust fund.
ȱȱȱ
ŗŖȱ
ȱ¢ǰȱǰȱȱȱ¢ȱ
ȱ
If individual accounts were created out of trust fund assets and those accounts were invested in
private securities the situation would be similar. If Social Security funds were held in personal
accounts and invested in private securities, their yield would likely be higher than if they were
invested in government securities. But, someone would have to buy the assets that would
otherwise have been purchased by the trust fund. And those who did would experience a drop in
income from capital.
None of this is to say that investing the trust fund in private securities is undesirable, or would
have no effect. Rather, it is to say that such a change would be unlikely to affect the national
saving rate.
Social Security will never function exactly like a collective retirement plan where each individual
sets aside a given amount in order to provide for his own retirement. Even with a fully funded
plan, there will be income transfers among participants. For example, those who live longer than
average will gain at the expense of those who die prematurely. As long as the plan is less than
fully funded, there will be an ongoing transfer from the working population to those who are
already retired.
In the case of individuals, how much they set aside during their working years will determine
how well they live in retirement. Similarly, for the nation as a whole, how much people save now
will play a role in the size of the economy in the future. The collective saving of households,
business, and the public sector will determine how much is invested. The more people invest, the
larger a stock of capital people will have and the more productive the labor force will be. A more
productive labor force means higher standards of living in the future.
By saving more now, the economy in the future will be larger than it otherwise would be. The
larger the economy is and the higher incomes are will make it easier to afford paying retirement
benefits no matter how they are financed.
ȱȱ ȱ
Brian W. Cashell
Specialist in Macroeconomic Policy
bcashell@crs.loc.gov, 7-7816
ȱȱȱ
ŗŗȱ