ȱ
ȱȱ¢ȱȱȱȱǵȱ
ȱǯȱȱ
ȱȱȱ¢ȱ
ȱŘřǰȱŘŖŖŞȱ
ȱȱȱ
ŝȬśŝŖŖȱ
ǯǯȱ
řŘŗŗşȱ
ȱȱȱ
Pr
epared for Members and Committees of Congress
ȱȱ¢ȱȱȱȱǵȱ
ȱ
¢ȱ
The national rate of saving may not be in the spotlight as often as other policy issues, but it
nonetheless has important effects on many issues that are, including the trade deficit and the
future ability of Social Security to secure the well-being of retirees. By saving more (and
consuming less), there are more resources available to expand the capital stock. A growing capital
stock is important because it adds to labor productivity and promotes rising living standards.
By economic definition, saving equals investment. It is not where the saving comes from, but
how much saving is available that determines the level of domestic investment. While it may not
matter for overall investment where saving comes from, other consequences follow from how
much of domestic investment is financed from domestic sources of saving. If saving from
domestic sources is insufficient to finance total investment, then the shortfall is made up by
importing saving. But, being a net importer of saving also means that we are a net importer of
goods and services.
There are three sources of domestic saving: households, business, and government. The targets of
past efforts to raise national saving have been households and the federal budget. Of the two,
households may be less susceptible to policies intended to stimulate saving. The response of
households to saving incentives is still open to question, and if household saving does not rise by
more than any reduction in tax revenue, the net effect on national saving will be minimal, or
could even be negative.
Increased saving by the federal government is much more likely to raise the national rate of
saving. An increase in taxes may result in a decline in household saving, but any reduction in
household saving is likely to be small relative to the increase in government saving.
This report will be updated as economic developments warrant.
ȱȱȱ
ȱȱ¢ȱȱȱȱǵȱ
ȱ
ȱ
Introduction ..................................................................................................................................... 1
A Brief Accounting of National Saving........................................................................................... 1
Why Saving Matters.................................................................................................................. 2
Household Saving............................................................................................................................ 3
Social Security and Household Saving ..................................................................................... 5
Business Saving............................................................................................................................... 6
Public Saving................................................................................................................................... 6
Social Security and Public Saving ............................................................................................ 7
Conclusion....................................................................................................................................... 8
ȱ
Author Contact Information ............................................................................................................ 8
ȱȱȱ
ȱȱ¢ȱȱȱȱǵȱ
ȱ
ȱ
The national rate of saving may not be in the spotlight as often as other policy issues, but it
nonetheless has important effects on many issues that are. It is an accounting identity in
economics that saving must equal investment. By saving more (and consuming less), there are
more resources available to expand the capital stock. A growing capital stock is important
because it adds to labor productivity and rising living standards.
In an economic sense there is no best saving rate. The rate at which we save is a reflection of our
relative preferences for consuming now, or deferring consumption to some future time. But when
the issue of saving arises in a public policy context it is usually because of concern that we may
not be saving enough to promote long-run economic prosperity.1
If raising the national saving rate is a public policy goal, what are the options? This report
discusses the various sources of saving in the United States and factors that determine the rates of
saving. It then examines policies that might affect saving decisions in each sector of the economy
and for the nation as a whole.
ȱȱȱȱȱȱ
Saving, in a nutshell, is income minus consumption. In an economic sense, however,
consumption is not the same as expenditures. Translating the theoretical notion of saving into a
statistical measure is a challenge. This section uses a basic economic accounting framework to
show the importance of saving to the overall economy.
U.S. saving is a part of the national income and product accounts (NIPA) published by the
Department of Commerce, Bureau of Economic Analysis (BEA). The NIPA constitute the
accounting framework used to produce quarterly estimates of gross domestic product (GDP).
GDP is the total value of goods and services produced, and it can be calculated in two different
ways. One way is to add up the value of all the goods and services produced (the product side of
the accounts), and the other way is to add up all of the income earned in the production of those
goods and services (the income side of the account).
On the product side, GDP is the total value of those goods and services produced for personal
consumption (C), investment (I), government (G), as well as for export (X). On the income side,
the income earned in the production of those goods and services can be accounted for as
consumer spending (C), taxes (T), private saving (S), and spending for imported goods and
services (M).
Since both of these approaches measure the same variable (GDP), they can be set equal to each
other, in this way:
C + I + G + X = C + T + S + M
1 For recent saving data, see CRS Report RS21480, Saving Rates in the United States: Calculation and Comparison, by
Brian W. Cashell.
ȱȱȱ
ŗȱ
ȱȱ¢ȱȱȱȱǵȱ
ȱ
Subtracting consumption (C) from both sides simplifies the equality:
I + G + X = T + S + M
This can now be rearranged by subtracting government spending (G) and exports (X) from both
sides, which gives:
I = S + (T − G) + ( M − X )
This equation illustrates the importance of saving. Total investment is equal to the sum of private
saving (S), public saving (T-G), and the net inflow of capital from abroad (M-X).2 To put it more
simply, investment equals saving.
In the broadest sense, saving is income less consumption. Consumption is typically taken to mean
spending on goods and services by households. But consumption can also refer to the wear and
tear (depreciation, or capital consumption) on the capital stock that occurs in the production of
those goods and services. Thus, some saving must be allocated to the replacement of the existing
capital stock as it ages or wears out. Only if saving is more than sufficient to replace the existing
capital stock as it wears out, or becomes obsolete, will the capital stock grow.
For this reason, saving data distinguish between “gross” and “net” saving. The difference between
the two measures is the estimated deterioration in the existing capital stock. In some cases, such
as computers, capital may be completely depreciated in a very short period of time, while in
others, such as buildings or heavy equipment, capital may take a long time to wear out.
¢ȱȱȱ
Given this accounting framework, it is easy to see the importance of saving to the economy. By
definition, saving equals investment. It is not where the saving comes from, but how much saving
is available that determines the level of domestic investment. However, the composition of saving
does affect the ownership of the capital stock and determines who has a claim on the income
generated by that stock. To the extent that domestic investment is financed by importing saving
from abroad, foreigners will have a claim on the interest and dividend income produced by those
investments.
The amount of saving that flows into the country also affects the balance of trade in goods and
services. In order for foreigners to invest their savings in the United States, they must first buy
dollars in foreign exchange markets. An increase in the inflow of foreign savings means an
increase in the demand for dollars, and an increase in the supply of foreign currency in currency
exchange markets. That tends to push up the value of the dollar relative to other currencies.3
2 Just as there is a balance between the income and product sides of the NIPA, there is a balance in international
payments. If Americans buy more goods and services from abroad than they export, then (at least in this simplified
example) the net outflow of dollars will be used by foreigners to buy dollar-denominated assets, which thus helps to
finance domestic investment.
3 This does not necessarily mean that the value of the dollar will rise over time, but rather that it will be higher than
otherwise would have been the case.
ȱȱȱ
Řȱ
ȱȱ¢ȱȱȱȱǵȱ
ȱ
The increase in the exchange value of the dollar, in turn, makes U.S. exports more expensive to
foreigners and imports less expensive to American consumers. That stimulates an increase in
imports while discouraging exports. The result of increased imports and reduced exports is a
larger trade, or current account, deficit. Thus, along with the increased net inflow of foreign
saving, there is also an increase in the net inflow of foreign goods and services.
While it may not matter for overall investment where saving comes from, there are other
consequences that follow from how much of domestic investment is financed from domestic
sources of saving. If saving from domestic sources is insufficient to finance total investment, then
the shortfall is made up by importing saving. But, being a net importer of saving also means that
the United States is a net importer of goods and services.
ȱȱ
The standard model of consumer spending used in economic analysis makes the basic assumption
that consumers seek to avoid large swings in their living standards over the course of their
lifetimes. Thus as incomes rise and fall both in the short and long term, individuals will vary their
saving rate in order to minimize the effect on their consumption.
Typically, over the course of an individual’s lifetime, income has a tendency first to rise over the
course of a career, and then fall in retirement. The life cycle model presumes that, in order to
dampen the effects of this income cycle on consumption, and living standards, individuals vary
the rate at which they save. Saving will thus tend to be relatively higher during their peak earning
years, and less at the beginning of their careers and during retirement.4
Any analysis of household saving must account for those factors which cause incomes to vary
over the course of a lifetime. One is retirement. Ideally, if the assumptions of the life cycle model
are valid, the purpose of saving is to avoid a drop in living standards after retirement. Insofar as
retirement saving is paramount, the individual saving rates will tend to vary with age, and the
overall household saving rate will tend to vary in response to demographic changes.
Another consideration is referred to as precautionary saving. The life cycle pattern of income
described above is only a generalization. Individual incomes may vary substantially over periods
of time in ways unrelated to the life cycle. The more likely and the more extreme those variations
are may affect how much more individuals may save above and beyond what is needed for
retirement. Thus variations in confidence about overall economic conditions and individual career
prospects may influence the saving rate.
Recent history suggests that major shifts in equity values may also affect the overall household
saving rate. An increase in equity values that is perceived to be durable may allow households to
continue accumulating wealth while saving less out of current income.5
4 The life cycle model was developed in order to reconcile two apparently contradictory observations. The first is that at
any point in time, people with relatively more income tend to have relatively higher saving rates. The second was that
as incomes rise over time there is no tendency for saving rates to rise. The life cycle model holds that saving is more a
function of age than income.
5 Saving rates derived from the NIPA are based on current income and expenditures, they do not reflect changing asset
prices. For that reason, some analysts consider changes in net worth a more meaningful measure of saving.
ȱȱȱ
řȱ
ȱȱ¢ȱȱȱȱǵȱ
ȱ
These factors may all be largely beyond the influence of public policy which suggests that
government’s ability to influence household thriftiness may be limited. One approach that has
been tried is to offer tax advantages for those households that save. In 1982, individual retirement
accounts (IRAs) were introduced as a way to stimulate household saving. Whether they did so has
become a subject of some controversy.6
After the introduction of IRAs, substantial amounts of funds flowed into the new accounts. But
that alone was not evidence of any effect on the total amount of household saving. The question is
not just whether household saving rose after the introduction of IRAs, but rather whether
households saved more than they would have in the absence of IRAs.
Comparing an actual outcome with a hypothetical base is a common problem in economic
analysis. Sometimes a theoretical argument can be made as to what the effects of a specific policy
are. That is not the case with respect to the effects of IRAs on household saving. A change in the
rate of return of saving has two different and offsetting effects on the amount an individual
chooses to save. Which of them dominates is an empirical question.
A rise in the rate of return may motivate an individual to save more simply because it makes
saving preferable to consumption. It raises the cost of current consumption relative to saving in
much the same way as a rise in the cost of beef relative to chicken will tend to increase the
consumption of chicken and reduce the consumption of beef. This is referred to in the jargon of
economics as a “substitution effect.”
But there is another way in which a rise in the rate of return may influence saving behavior.
Perhaps the easiest way of explaining this is in the context of what is known as a “target saver.” If
the goal of saving is to accumulate a given level of wealth by retirement age, a higher rate of
return on saving means that goal can be satisfied while saving less. In the jargon of economics,
this is referred to as the “income effect,” because the increase in income as a result of the higher
rate of return allows individuals to meet their goals while saving less.
Which of these two effects dominates and what the net effect of an increase in the rate of return
on saving is have been the subject of some controversy. That the personal saving rate fell
following the introduction of IRAs did little to encourage those who hoped that giving tax
advantages to saving accounts would encourage new saving.
There is another consideration, with respect to IRAs, that needs to be taken into account. In
theory, economic incentives are effective only at the margin. At their introduction, the tax benefits
of IRA contributions were limited to the first $2,000. That limit increased to $5,000 in 2008.7 A
substantial share of total household saving is accounted for by households that save more than
$5,000. It is unclear why a benefit limited to the first $5,000 of saving would motivate those who
would save that much in the absence of IRAs to increase the amount they save. Rather, it seems
more likely that people in those circumstances would simply change where they put their savings
in order to take advantage of the higher after tax rate of return.8
6 See CRS Report RL30255, Individual Retirement Accounts (IRAs): Issues and Proposed Expansion, by Thomas L.
Hungerford and Jane G. Gravelle.
7 See CRS Report RS22019, IRAs and Other Savings Incentives: A Brief Overview, by Jane G. Gravelle.
8 Jane Gravelle, “Do Individual Retirement Accounts Increase Saving?,” Journal of Economic Perspectives (Spring
1991), pp. 133-149.
ȱȱȱ
Śȱ
ȱȱ¢ȱȱȱȱǵȱ
ȱ
Even assuming that tax-favored saving accounts do generate additional household saving, the
provision of the tax advantage results in a reduction in federal tax revenues increasing the budget
deficit and cutting the saving rate of the federal government. For IRAs to raise the national saving
rate any increase in household saving would have to be larger than the reduction in federal
revenues.
Empirical studies of the effects of IRAs on saving have come to different conclusions. Poterba,
Venti and Wise examined the financial holdings of similar households over time after the
introduction of IRAs. They found that the non-IRA financial assets did not decline as would have
been the case if households simply shifted saving from non-IRA investments into IRAs. They
argue that the bulk of contributions to IRAs represented new saving.9 But some of that increase in
household saving was still offset by a reduction in federal government saving.
In contrast, Engen, Gale, and Scholz found the opposite. They begin by pointing out the many
difficulties involved in separating new saving from saving that would have occurred anyway. In
particular, they argue that one reason for the growth of non-IRA assets after the introduction of
IRAs was the boom in the stock market. They found that saving incentives have a strong effect on
the allocation of saving and wealth, but that there was little or no effect on the overall level of
household saving.10
ȱ¢ȱȱ
ȱȱ
As it currently stands, Social Security is partially funded. That means that part of the benefits now
being paid to retirees are funded from their earlier contributions and part are transfers from the
working population. To the extent that Social Security operates on a pay-as-you-go basis, it tends
to reduce aggregate household saving because it transfers income from workers, who tend to be
savers, to retirees who save relatively less.
A shift toward a fully funded retirement system would require some combination of increased
contributions and reduced benefits. But the net effects of those changes on household saving are
uncertain. Households might reduce their saving in order to pay higher contributions. Reducing
benefits might prompt households to save more to make up for the lost future income. It is also
conceivable that some current household saving is motivated by uncertainty regarding the future
of the Social Security program. To the extent that any Social Security reform inspired greater
confidence in future benefits, saving might fall. However, any reduction in household saving that
might result is likely to be outweighed by the increase in public sector saving resulting from
either benefit cuts or increased contributions.11
9 James M. Poterba, Steven F. Venti, and David A. Wise, “How Retirement Saving Programs Increase Saving,” Journal
of Economic Perspectives (Fall 1996), pp. 91-112.
10 Eric M. Engen, William G. Gale, and John Karl Scholz, “The Illusory Effects of Saving Incentives on Saving,”
Journal of Economic Perspectives (Fall 1996), pp 113-138.
11 See CRS Report RL30708, Social Security, Saving, and the Economy, by Brian W. Cashell.
ȱȱȱ
śȱ
ȱȱ¢ȱȱȱȱǵȱ
ȱ
ȱȱ
Business saving, as it appears in the national economic accounts, is the sum of undistributed
profits and capital consumption or depreciation. How much a firm contributes to saving is a
consequence of how it allocates its income. It can either pay it out as dividends, use it to replace
its existing stock of capital as it wears out or becomes obsolete, or can use it to increase its total
net stock of productive capital. How a firm decides to allocate its income will be in large part
determined by whether or not it is in an expanding or shrinking market. If the market for a firm’s
production is increasing, its dividend distribution to stockholders is likely to be smaller than
would be the case for a firm in a stagnant or diminishing market.
The fact that households are the ultimate owners of the business sector implies that business and
household saving rates are linked. For example, a firm may invest in replacing worn out capital
and further expanding its stock of capital rather than making a large dividend distribution. The
increased capital stock of the firm raises its productivity and profits, increasing the wealth of the
stock holding households.
If the goal of household saving is to accumulate wealth, they may have reason to be indifferent
between saving out of dividend distributions or having the value of their stock holdings increase.
Firms in an expanding market with increasing investment and profit opportunities will likely face
less demand for dividend distributions than will firms in shrinking markets.
Whether changes in business saving affect household saving is an empirical question. There is
evidence to suggest that business and household saving are closely interrelated. Bosworth, for
example, found that business and household saving were closely related in a number of countries.
In the United States, he found strong evidence that variations in business saving influenced
household saving decisions.12
If an increase in business saving causes households to save less, then any policy intended to
stimulate business saving might have little effect on total private sector saving.
ȱȱ
The public sector can also be a source of saving. Saving by both the federal and state and local
government is simply the difference between revenues and expenditures. When the public sector
is in deficit, it is absorbing funds that might otherwise be available to the private sector to finance
investment spending. When the public sector is in surplus, the excess of revenues can be used to
reduce outstanding public debt. Debt reduction is equivalent to saving. Instead of buying the debt
issues of other institutions, government can buy back its own debt issues.
Just as household saving may be affected by changes in business saving, changes in the public
sector saving rate can influence household thriftiness. One way this might happen was initially
suggested by Barro.13 Consider the case of an increase in government spending. Barro argued that
12 Barry P. Bosworth, Saving and Investment in a Global Economy (Washington: Brookings Institution Press, 1993).
13 Robert Barro, “Are Government Bonds Net Wealth?,” Journal of Political Economy (November/December 1974),
pp. 1095 - 1117.
ȱȱȱ
Ŝȱ
ȱȱ¢ȱȱȱȱǵȱ
ȱ
if taxpayers regard the well-being of their children equally with their own, it should not matter to
them whether the spending increase is financed by increasing taxes or by borrowing. If the new
spending is financed by borrowing instead of raising taxes, households will expect that at some
time in the future taxes will have to raised to pay the increased interest on the debt or to pay off
the debt.
Knowing that, if they are perfectly rational and if they are farsighted, they will increase their
saving so as to be able to meet that future burden. If this assumption of extreme rationality were
correct then any change in government saving would be completely offset by a change in
household saving. There is little empirical evidence, however, to suggest that households react in
this way to changes in government saving.14
That is not to say that household saving is unaffected by variations in public sector saving. Any
change in household income affects both saving and consumption. Consider a tax increase, which
would raise the public sector saving rate. In order to pay the increase in taxes households must
either reduce consumption or saving.
How much of a tax increase is paid by reducing saving may depend on whether the tax increase is
perceived to be temporary or permanent. The life cycle model, described above, assumes that one
purpose of saving is to help insulate households from substantial variations in consumption over
the course of their lifetimes. If that assumption is correct, an increase in taxes that is perceived to
be permanent is more likely to be paid by a reduction in household consumption. A permanent tax
increase would reduce income over the course of an individual’s entire lifetime, and it is likely
that lifetime consumption would be reduced as well.
An increase in taxes that is perceived to be only temporary is more likely to be paid by
temporarily reducing saving. If it were paid entirely out of saving, then there would be no net
effect on the national saving rate.
One complication of any government action to increase public saving, be it increases in taxes or
spending cuts, is that it may contribute to slower economic growth in the short run. It is a periodic
dilemma for policymakers that the short run goal of promoting full employment conflicts with the
long-term goal of increasing national saving.
ȱ¢ȱȱȱȱ
With respect to federal government saving, the measure that matters is the unified budget surplus
or deficit. For some purposes, the budget is divided into two accounts; one of these is referred to
as “on-budget,” and the other is “off-budget.” The so-called off-budget account consists almost
entirely of Social Security receipts and outlays.
Other things being equal, an increase in tax revenues will raise government saving. It makes no
difference to government saving whether the revenues are in the on- or off-budget accounts.
Revenues that are credited to the Social Security trust fund contribute to government saving in
exactly the same way as revenues that fund current federal government outlays. Without the
14 Edward M. Gramlich, “Budget Deficits and National Saving: Are Politicians Exogenous?,” Journal of Economic
Perspectives, vol. 3, no. 2 (Spring 1989), pp. 23-35.
ȱȱȱ
ŝȱ
ȱȱ¢ȱȱȱȱǵȱ
ȱ
revenues from Social Security taxes, the current budget deficit would be larger, and the federal
government’s borrowing needs would be larger as well.
Any reform of Social Security that shifts away from a pay-as-you-go towards a more fully funded
program would necessarily involve some combination of increased contributions and reduced
benefits. Any such change would raise the public sector saving rate. It might be that households
would reduce their saving in order to meet the pay higher contributions, but it is unlikely that any
reduction in household saving would completely offset the rise in public saving.
ȱ
From a macroeconomic perspective, it makes no difference who saves. But, how much
households, businesses, and government save collectively affects the level of domestic
investment, inflows of foreign saving, and the balance of trade. Economic theory offers no single
standard against which to assess the national saving rate as too high, or too low. However, some
of the consequences of how much, or how little, we save are often seen by policymakers as signs
of economic weakness.
The three sources of domestic saving are households, business, and government. Households and
the federal budget have been the targets of past efforts to raise national saving. Of the two,
economists have greater confidence in the latter. The response of household saving to past efforts
is still open to question, and if household saving does not rise by more than any tax revenue lost,
the net effect on national saving will be minimal, or could even be negative.
Increased saving by the federal government is much more likely to translate into a higher national
rate of saving. An increase in taxes may result in a decline in household saving, but any reduction
in household saving is likely to be small relative to the increase in government saving.
ȱȱ ȱ
Brian W. Cashell
Specialist in Macroeconomic Policy
bcashell@crs.loc.gov, 7-7816
ȱȱȱ
Şȱ