Why is the Household Saving Rate So Low?

^Pr^ep^ar^{ed f}^o^r^M^e^m^b^ers^an^d^Com^mⁱ^t^t^ees^of^C^ong^r^e^ss

Despite mostly favorable economic conditions since the early 1990s, one development—the focus
of this report—may be cause for concern. While the economy was growing, household saving fell
to next to nothing. At the beginning of the 1990s, households saved, on average, about 8% of
their after-tax income. Since then, it has steadily declined, and since 2005, the proportion of
income households set aside has been near zero.
One of the most striking aspects of the economic expansion of the 1990s was the dramatic rise in
equity prices. Between 1991 and 2001, the Standard and Poor’s index of 500 stock prices rose by
217%, and the NASDAQ stock price index increased by 313%. The standard economic model of
saving behavior assumes that households are likely to spend more and save less out of current
income given an increase in their wealth. Given that, it seems reasonable to suppose that those
increases in equity prices had something to do with the decline in the personal saving rate. A
number of studies have found that the decline in the personal saving rate during the 1990s was
due at least in part to the rise in equity prices.
The stock market peaked in the second half of 2000. Between September 2000 and October 2002,
the S&P 500 fell by 46%. Since then, S&P 500 stock prices have recovered, but they remained
slightly below the 2000 peak at the end of 2006. Even though the stock market boom ended, the
household saving rate continued to decline. The fact 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 is the boom in the housing market. Since late 1997, the price of
housing has risen rapidly. Between late 1997 and late 2006, house prices doubled, on average.
Although there are reasons to think that housing price appreciation might not be a good substitute
for saving, a number of empirical studies, while not quite a consensus, found evidence to suggest
that it might.
When changes in personal wealth are taken into account, the decline in the official saving
measure may be less of an issue, although there will always be concerns about the durability of
any gains in asset prices. With respect to the household sector’s contribution to total national
saving, that has clearly fallen. Meanwhile, investment spending has risen relative to gross
domestic product (GDP), in spite of the decline in personal saving, due to the increase in capital
inflows from abroad. Some of the productivity gains likely to result from that increased
investment, however, will benefit foreign rather than domestic investors. If house price
appreciation slows for a long time, households may end up with less wealth than they anticipate.
That would make the current low saving rate seem more of a problem.
This report will be updated as developments warrant.

Introduc tion ..................................................................................................................................... 1
The Decline in Household Saving...................................................................................................1
Measuring Saving............................................................................................................................2
An Alternative Measure............................................................................................................3
Who Saves?...............................................................................................................................4
Household Saving and Retirement..................................................................................................5
Equity Prices and Personal Saving............................................................................................6
House Prices and Saving...........................................................................................................7
Other Considerations.................................................................................................................9
Are Households Saving Enough?..............................................................................................9
Conclusion ..................................................................................................................................... 10
Figure 1. Personal Saving as a Percentage of Disposable Personal Income...................................2
Figure 2. Flow of Funds Household Saving Rate............................................................................4
Table 1. Saving by Income Quintile................................................................................................5
Author Contact Information...........................................................................................................11

With the exception of a contraction in 2001, the economy has experienced healthy economic
growth since the early 1990s, aided in part by a pickup in productivity growth that began in 1995.
At the same time, even in a period of relatively rapid growth in output, inflation remained low.
One development however, and the focus of this report, may be cause for concern. While the
economy was growing, household saving fell to almost nothing. At the beginning of the 1990s,
households saved, on average, about 8% of their after-tax income. Since then, it has steadily
declined, and since 2005, the proportion of income households set aside has been near zero.
A negative saving rate would seem to be cause for worry. Congress has already indicated its
desire to promote household saving by, among other things, creating individual retirement
accounts, and saving is an important consideration in proposals to reform Social Security.
Household saving is important for two distinct reasons. First, most households must accumulate
wealth over the course of their working lives if they are to avoid a decline in their standard of
living after retiring. Saving next to nothing would seem unlikely to afford a comfortable
retirement for many, especially if Social Security benefits are ever reduced as part of any reform
plan, as many workers fear may occur. Second, households are an important source of funds to
finance capital investment, increasing the capital stock and adding to worker productivity.
This paper begins by showing how much the household saving rate has declined in recent years.
Next, it explains how household saving is measured, and provides some detail on how saving
varies across the income distribution. Finally, it discusses factors that may account for the decline
in household saving, as well as how much of a policy concern the decline in household saving ¹
may be.

The personal, or household, saving rate has fallen steadily since 1990. Personal saving as a
percentage of after tax personal income fell from just under 8% in 1990 to near zero since 2005.
Figure 1 shows the monthly personal saving rate since 1970.
Clearly the personal saving rate has dropped. But it is important to understand what, exactly, this
measure of saving takes into account. As is the case with many economic statistics, there are
differences between the theoretical notion of saving and statistical efforts to measure it.

¹^For^a^c^o^m^p^le^te^a^c^c^ount^ing^of^s^a^vⁱ^ng^{in t}^h^e^Uⁿⁱ^t^e^d^Sta^t^e^s^{, s}^e^e^C^R^S^R^e^por^{t R}^S²¹⁴⁸^0,^Savⁱ^ng^Rate^sⁱⁿ^the^Uⁿⁱ^t^e^d
^State^s^:^C^a^lc^ul^atioⁿ^an^{d C}^o^m^par^is^on^,^by^Bria^{n W}^.^Ca^she^ll.

Figure 1. Personal Saving as a Percentage of Disposable Personal Income
^Source:^Depart^{ment of Comme}^{rce, Bureau}^{of Economic Analysis}^.

The simplest definition of saving is income minus consumption. But what counts as income and
how to measure consumption can be complicated. The national income and product accounts
(NIPA) published by the Bureau of Economic Analysis (BEA) of the Department of Commerce ²
are the primary source for statistics on overall U.S. economic activity. The aim of the NIPA is to
account for both income and expenditures that are related to the current production of goods and
services. Changes in asset values are not included since they have nothing to do with current
production. Thus, capital gains and losses are not counted as income in the NIPA.
Even though capital gains are not included in the NIPA measure of personal income, they still
affect the measurement of the personal saving rate. Tax payments on capital gains realizations are
included in personal tax payments. Thus, after-tax personal income is reduced by the amount of
the tax payments. Saving, which is calculated by subtracting consumption from income, is
reduced as well.
A difference between saving in a theoretical sense and the NIPA measure of saving is in the ³
treatment of consumer durable goods. Ideally, the consumption of durable goods would be
measured by the value of the flow of services they provide over their useful lives, similar to an
investment. In the NIPA, however, consumption of durable goods is treated like any other type of

²^T^h^{is is the}^so^urc^e^f^o^{r the}^da^ta^s^h^ow^{n in}^Fⁱ^g^u^{re 1}^.
³^C^ons^um^e^r^dur^a^b^le^g^oods^c^ons^is^t^m^a^inly^o^f^pur^c^h^a^s^e^s^of^a^u^tom^obil^e^s^,^f^u^rⁿ^itur^e^{, a}ⁿ^d^ho^us^e^hol^{d e}^q^uip^m^eⁿ^t.

consumer spending, by simply counting the expenditures on those goods. That portion of income
that is invested in housing is not considered to be consumption in the NIPA and thus contributes ⁴
to total personal saving.
A similar difference occurs with respect to expenditures on education. They are currently treated
in the NIPA as consumption spending, but given that they yield benefits over an individual’s
lifetime they might more appropriately be counted as investment spending. If they were, that
would raise the measured saving rate.
The Board of Governors of the Federal Reserve Board (Fed) publishes an estimate of household
saving that differs somewhat from the one published by BEA. The BEA estimate of household
saving is calculated by subtracting consumption expenditures from income. In contrast, the Fed
estimate is based on a balance sheet for the household sector. When households save, it shows up
in the Fed’s flow of funds accounts as an increase in household net worth. The gross increase in
household net worth is the sum of the net acquisition of financial and tangible assets minus the
net increase in household liabilities.
Using net acquisition of tangible assets to measure saving means that not only is investment in
housing included in the flow of funds measure of saving, but also net investment in consumer
durable goods. Since consumer spending on durable goods is substantial, including it in the flow
of funds measure of saving means that the Fed’s estimate of household saving is usually higher
than BEA’s estimate. Figure 2 shows the household saving rate from the flow of funds accounts
including the net increase in household durable goods. The alternative Fed measure shows the
same steady decline in the household saving rate beginning in the 1980s as the BEA measure. But
instead of falling from about 10% to below zero, the Fed measure shows a decline from about

14% to about zero.

Accounting for net investment in consumer durables as saving is entirely appropriate given that
saving is defined as income less consumption. But when policymakers express concern about the
low saving rate the distinction between the Fed measure and the BEA measure may not be very
significant. The public policy concern over low household saving is one related to retirement
saving, and investments in automobiles and household furnishings are not likely to be a source of
income for retirees. What is significant is that the independent measures both show that
household saving has fallen to very low historical levels.

⁴^M^a^rsh^a^l^l^B.^Reiⁿ^sd^orf^,^“A^l^t^e^rn^atⁱ^v^{e M}^easu^r^{es o}^f^S^a^viⁿ^g^,^”^S^u^r^vey^o^f^Cu^rren^t^B^u^siⁿ^e^ss^{, Se}^pte^m^be^r²⁰⁰⁴^{, p}^p^.^17-²⁷^.

Figure 2. Flow of Funds Household Saving Rate
^Source:^{Board of Governor}^{s of}^{the Federal Reserve System.}
The saving rates depicted in Figure 1 and Figure 2 are averages. But not everyone saves at the
same rate and it might be more informative to have an idea of how saving rates vary across
households, by income in particular. Table 1 presents data on saving by quintile of income. The
lowest quintile represents the bottom 20% of the income distribution, the second quintile
represents the second lowest 20% in the distribution, and so on. The figures in the first row show ⁵
saving rates. They are derived from data from the 2006 Survey of Consumer Expenditures,
published by the Bureau of Labor Statistics. The figures in the second row are taken from the

2004 Survey of Consumer Finances published by the Board of Governors of the Federal Reserve ⁶

System.
Both sets of numbers indicate that those at the upper end of the income distribution account for
most saving. On average, they save at a higher rate than those lower in the distribution, and a
larger proportion of them save. Given that the overall saving rate is close to zero, the saving of
those at the upper end of the distribution is about equal to the dis-saving of those at the bottom of
the distribution.

⁵^T^h^e^s^e^s^a^vⁱ^ng^r^a^te^s^s^houldⁿ^o^t^be^ta^k^e^{n a}^s^pr^e^cⁱ^s^e^e^s^ti^m^a^tes^.^T^h^ey^ar^e^de^rⁱ^v^e^{d by}^s^ubtr^a^c^ting^e^x^pe^ndi^tur^e^s^f^r^om
ⁱⁿ^co^m^e^{as rep}^o^r^t^e^dⁱⁿ^{a su}^rvey^.^I^f^t^h^{e est}ⁱ^m^at^{es o}^fⁱⁿ^co^m^e^an^d^exp^eⁿ^dⁱ^t^u^{res are b}ⁱ^asedⁱⁿ^op^po^si^t^e^di^r^ectⁱ^oⁿ^s^t^h^eⁿ^t^h^ere
^c^ould^be^c^o^nside^r^a^b^le^e^r^r^o^{r in}^the^re^sulting^sa^vⁱ^ng^e^s^tim^a^t^e^.
⁶^T^h^is^s^u^r^v^ey^is^c^onduc^te^d^e^v^e^r^y^thr^e^e^y^e^a^r^s^.

Table 1. Saving by Income Quintile
^{Income Quintile}
^Lowest^Second^Third^Fourth^Fifth
^{Saving as a perc}^{ent of aft}^ertax^ho^{usehold income (20}⁰⁶⁾^-104.7^-14.7^5.4^18.7^33.6
^{Percentage of fa}^{milies that saved}⁽²⁰⁰⁴⁾^34.0^43.5^54.4^69.3^79.2
^Sources:^Depart^{ment of}^Labor,^{Bureau of}^{Labor Statistics; Board of Governors of t}^{he Federal Res}^{erve System.}

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 are expected
to 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 model of saving based on this “life cycle”
presumes that, in order to dampen the effects of the lifetime 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 lower at the beginning of their careers and
during retirement.
From an economic perspective, there is no “ideal” saving rate. The rate at which an individual
saves is, for the most part, simply a reflection of his willingness to consume less now in order to
be able to consume more in the future. Taking the life cycle model as a guide, however, prudent
individuals might be expected to save enough to avoid a substantial decline in their living
standard on retirement.
If consumers do seek to maintain a fairly stable level of consumption over their entire lives, then
the level of consumption at any given point in their lives will depend on their current wealth and ⁷
some expectation about their earnings over the rest of their lives.
There are a number of implications that follow this model of consumer behavior. One is that an
unexpected one-time income windfall is more likely to be saved than spent. A rise in income that
is temporary is less likely to raise consumption than one that is permanent as long as consumers
seek to maintain a relatively constant standard of living. A permanent increase in income is more
likely to induce an increase in consumption, so that the saving rate will be less affected. Similarly,
an increase in overall wealth raises consumption possibilities over the long run. An increase in
wealth that is believed to be permanent will tend to raise consumption. Unless there is a
corresponding increase in income, an increase in wealth is likely to reduce the measured rate of
saving.
Consumption is ultimately constrained by income and wealth. An increase in wealth increases
potential lifetime consumption and diminishes an individual’s incentive to save. The effect of an

⁷^T^o^the^e^x^te^{nt tha}^t^c^ons^um^e^r^{s c}^a^{n borr}^o^w^a^g^aⁱ^{nst f}^u^ture^inc^o^m^e^{, tha}^t^e^nha^nc^e^s^theⁱ^{r a}^b^ility^{to m}^a^inta^{in a}^f^a^irl^y^ste^a^d^y
^le^v^e^{l of}^c^ons^um^{ption ov}^e^r^a^l^ong^pe^rⁱ^od^of^tim^e^.

increase in wealth on consumption and saving may depend on a number of variables. Most
important, if the life cycle model is correct, only a fraction of any increase in wealth would affect
current consumption because any rise in consumption would be spread out over one’s remaining
lifetime. Some increases in wealth may not be viewed as permanent as might be the case with
assets whose prices tend to be highly variable. If there is uncertainty about the permanence of an
increase in wealth then it is less likely to affect consumption and saving. In addition, a strong
desire to leave a bequest may dampen the influence increasing wealth has on consumption.
Numerous studies of wealth and consumer spending have been conducted. Empirical estimates of
the magnitude of the effect of changes in wealth on consumption vary, but they tend to be fairly
small. The estimates suggest that for each additional dollar of wealth, consumption spending ⁸
tends to rise by anywhere from 1 cent to 7 cents.
One of the most striking aspects of the economic expansion of the 1990s was the dramatic rise in
equity prices. Between 1991, the beginning of the expansion, and 2001, the year it ended, the
Standard and Poor’s index of 500 stock prices rose by 217%, the Dow Jones Index rose by 247%,
and the NASDAQ stock price index increased by 313%. Given the model of saving behavior
described above, it seems reasonable to suppose that those increases in equity prices had
something to do with the coincidental decline in the personal saving rate.
A number of studies have found that the decline in the personal saving rate during the 1990s was
due at least in part to the rise in equity prices. Lusardi, Skinner, and Venti found that between
one-half to two-thirds of the decline in the personal saving rate was attributable to the effect of ⁹
wealth gains on consumption spending. They also point out that because of rising equity prices,
firms were able to reduce their contributions to defined benefit retirement plans and still keep
those plans fully funded. Pension benefits could be paid out of the capital gains of the pension
fund. That tended to reduce a component of personal income (employers’ contributions to pension
funds) without reducing consumption and thus contributed to the decline in the saving rate.
Maki and Palumbo analyzed the drop in saving of the 1990s and found that almost all of it could ¹⁰
be attributed to the performance of the stock market. Moreover, they found that those
households with the highest incomes experienced the largest gains in wealth and also the largest
declines in saving rates. In fact, they found that, between 1992 and 2000, almost all of the decline
in the aggregate personal saving rate can be attributed to reduced saving of those households in
the top 20% of the income distribution. They also found that the saving rate of those households
in the bottom 40% of the income distribution actually rose over the same period.

⁸^{See Marth}^a^Starr-McCl^u^{er, “St}^o^ck^Ma^rk^{et Wealth}^an^d^C^oⁿ^s^u^m^{er Sp}^en^dⁱⁿ^g^{,” Fi}ⁿ^aⁿ^{ce an}^d^Ecoⁿ^o^m^ic^{s Discu}^ssioⁿ
^Se^rⁱ^e^s^{, 19}^98-^20,^B^o^a^r^d^of^G^o^v^e^rⁿ^or^s^of^the^Fe^de^r^a^{l R}^e^s^e^r^v^e^Sy^s^t^em^,^A^p^r^{il 19}^98,²¹^p^p^{. A}^l^s^o^{, C}^ong^r^e^s^sⁱ^ona^{l B}^u^dg^e^t
^Offⁱ^c^e^,^The^Ec^o^nom^ic^an^{d Bu}^dge^t^O^u^tlo^ok^:^Aⁿ^U^p^date^,^A^u^gu^st¹⁹⁹⁸^,^p.¹⁹^.
⁹^Aⁿⁿ^a^m^a^ri^a^L^u^sardⁱ^,^Joⁿ^a^t^h^an^S^kⁱⁿⁿ^e^r,^and^S^t^even^V^eⁿ^tⁱ^,^“S^aviⁿ^g^P^u^zzl^{es an}^d^S^a^viⁿ^g^P^o^lⁱ^cⁱ^e^{s i}ⁿ^t^h^{e Un}ⁱ^t^e^d^S^t^at^es,^”
^N^a^tiona^{l B}^u^r^e^a^u^of^Ec^on^om^ic^R^e^s^e^a^r^c^h^{, W}^o^r^k^ing^P^a^pe^r⁸²^37,^A^p^r^{il 2}⁰⁰¹^,⁵²^pp^.
¹⁰^D^e^aⁿ^{M. Ma}^kⁱ^aⁿ^{d Mic}^h^a^e^l^G^.^Pa^lum^{bo, “}^D^is^eⁿ^ta^ng^ling^the^W^e^a^{lth Ef}^f^e^c^t:^A^C^ohor^{t A}ⁿ^a^l^y^s^is^o^f^Hous^e^h^old^Sa^vⁱ^ng
^{in t}^h^e¹⁹⁹⁰^s^,^”^B^o^a^r^d^of^G^o^v^e^r^nor^s^of^the^Fe^de^r^a^{l R}^e^s^e^r^v^e^S^y^s^t^em^,^Fina^nc^e^aⁿ^{d Ec}^o^nom^ic^s^Dⁱ^s^c^us^sⁱ^on^Se^rⁱ^e^s^{, 20}^01-²¹^,
^A^p^r^{il 200}^1,³⁸^p^p^.

It seems both theoretically and empirically clear that the large increase in equity prices of the
1990s likely had something to do with the decline in the measured rate of personal saving. If the
drop in saving was simply a response to the increase in asset values, then households at the end of
the 1990s were not necessarily any less well prepared for the future than if stock prices had not
risen so rapidly and the saving rate had not fallen. In this sense existing measures of personal
saving may not be very good guides to the extent to which individuals are preparing for
retirement or the proverbial rainy day.
Because capital gains can contribute just as much as saving to a household’s retirement nest egg,
it might be appropriate to include them in any measure of retirement saving. After all, the aim of
saving for retirement is to accumulate enough wealth to be able to continue, in retirement, the
lifestyle to which one has become accustomed during one’s working life. Gale and Sabelhaus
have calculated what the saving rate might have been if capital gains were included in the current ¹¹
measure of personal saving. They found that, after adjusting for inflation, household wealth rose
at least as rapidly in the late 1990s as at any time since 1960.
Until the end of 2000, this explanation for the decline in the saving rate was widely accepted. But
subsequent developments suggested that stock price movements alone are insufficient to explain
the downward trend in household saving.
The stock market peaked in the second half of 2000. Between September 2000 and October 2002,
the S&P 500 index of stock prices fell by 46%. Since then, stock prices have recovered, but they
remained slightly below the 2000 peak at the beginning of 2008. Even though the stock market
boom ended, the household saving rate continued to decline.
Much of the wealth gain due to the equity appreciation of the 1990s endures, but it was the
appreciation that allowed households to accumulate wealth without saving out of current income.
If rising equity prices in the 1990s are the explanation for declining household saving, then
household saving might have been expected to rise when the stock market declined during 2001
and 2002. But that did not happen.
The fact 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 is the boom in the
housing market. Since late 1997, the price of housing has risen rapidly relative to prior years.
Between late 1997 and late 2007, the house price index published by the Office of Federal
Housing Enterprise Oversight (OFHEO) doubled.
There are reasons it might seem less likely that an increase in the value of houses would have the
same effect on household saving behavior as equity price appreciation. An increase in house
prices increases the wealth of current homeowners, but it also increases the cost of housing. For
renters, that might result in a reduction in saving in order to pay increased rents. For many
homeowners, an increase in housing wealth may simply translate into an increase in consumption
of housing, rather than spending more for other goods and services. In that case, an increase in
housing prices would be unlikely to show up as a reduction in saving out of current income.

¹¹^W^illia^m^G^.^G^a^l^e^a^{nd Joh}ⁿ^Sa^be^l^h^a^u^{s, “}^T^he^Sa^vⁱ^ng^{s Crisis: In t}^h^e^Ey^e^of^the^Be^holde^r?^”^T^h^{e Mi}^l^ken^In^stⁱ^t^u^t^{e R}^evi^ew^,
^Thⁱ^r^d^Qu^ar^t^e^r¹⁹⁹⁹^,^p^p^.⁴⁶^-⁵⁶^.

Another consideration is that some savers may have a bequest motive for saving. In other words,
they save more than is necessary for their own well-being in retirement in order to be able to pass
some wealth on to their children. A permanent increase in the price of housing now would also
increase its cost for future generations. Presumably, many of those for whom the bequest motive
is important would increase their desired bequest to compensate for the higher cost of housing.
It may also be more difficult for homeowners to consume any of the gains from house price
appreciation. Transaction costs in the housing market are significantly higher than they are in
equity markets, although homeowners may be able to borrow against any gains they have. It has
also been suggested that because of the perceived illiquidity of housing wealth homeowners have
no expectation of being able to consume out of capital gains in housing.
A number of studies have found that the effect on saving of asset price appreciation depends on
the kind of asset. For example, Juster, et al., found that for each $1 increase in wealth, household ¹²
spending rose (saving fell) by about 3 cents. But they also found that the effect of asset price
appreciation on spending depended on the type of asset. They found that household spending was
much more influenced by gains in equity wealth than was the case for any other type of asset.
The authors found a large effect on saving of appreciation in stock prices, as much as 19 cents for
each dollar gain in equity wealth. But the effect of house price appreciation was much smaller.
The effect of a $1 increase in housing wealth was a 3 cent drop in saving, but the estimate was
not statistically significant. While the increase in equity prices during the 1990s may explain the
drop in household saving, this study suggests the rise in house values is an unlikely reason for the
saving rate’s continued decline.
Other studies, however, have suggested that housing price appreciation may have had a
significant effect on household saving. Belsky and Prakken found that in the long run, the effects ¹³
on household saving of house and equity price variations were similar. 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.
Case, et al., did a cross-country and cross-state comparison of changes in both housing and equity ¹⁴
wealth and the effect on household spending. They found that increases in housing wealth had a
greater effect on household spending (and, hence, saving) than increases in equity wealth. They
also found evidence that the correlation between increases in housing wealth and spending

¹²^{F. T}^hom^a^s^J^u^s^t^e^r^{, J}^o^s^e^ph^{P. L}^upt^{on, J}^a^m^e^s^P^.^Sm^{ith, a}^nd^Fr^aⁿ^k^St^a^f^f^o^r^d^,^The^D^e^c^lineⁱⁿ^H^o^us^e^hol^d^Savⁱ^ng^an^{d t}^h^e
^W^e^{alth Effe}^c^t^{, B}^o^a^r^d^of^G^o^v^e^r^nor^s^of^the^Fe^de^r^a^{l R}^e^s^e^r^v^e^S^y^s^t^em^W^o^r^k^ing^P^a^pe^r²⁰^04-^31,^J^une²⁰^04,¹⁷^p^p^.
¹³^Er^ic^B^e^ls^ky^a^{nd J}^o^e^l^P^r^a^k^k^e^{n, “}^H^ous^ing^’^s^I^m^pa^c^t^{on W}^e^a^{lth A}^c^cum^u^la^{tion, W}^e^a^{lth D}ⁱ^s^t^rⁱ^buti^oⁿ^a^{nd C}^ons^um^e^r
^Speⁿ^dⁱ^ng^,”^N^a^tio^na^{l A}^s^s^o^cⁱ^a^tion^of^R^e^a^ltor^s^N^a^tiona^{l C}^e^nte^r^f^o^r^R^e^a^l^Es^ta^te^R^e^s^e^a^r^c^h^{, 20}⁰⁴^{, 2}⁶^pp.
¹⁴^Ka^{rl E. Ca}^se^{, Joh}ⁿ^M^.^Quig^le^y^,^a^{nd R}^obe^{rt J.}^Sh^ille^{r, “}^C^om^pa^ring^W^e^a^{lth Eff}^e^c^t^{s: T}^h^e^Stoc^k^Ma^rk^e^t^v^e^{rsus the}
^H^ous^ing^Ma^r^k^e^t^,”^Iⁿ^s^tit^ute^of^B^u^sⁱ^ne^s^s^a^{nd Ec}^on^om^ic^R^e^s^e^a^r^c^h^{, U}ⁿ^iv^e^r^sⁱ^t^y^of^C^a^li^f^o^rⁿ^ia^{, B}^e^r^k^e^l^ey^,^W^o^r^k^ing^P^a^pe^r
^W^01-⁰⁰^4,²⁰^05,³²^p^p^.

increased after 1986 when tax law changes made it more advantageous for households to borrow
against home equity.
Benjamin, et al., found that changes in real estate wealth had a greater effect on spending and ¹⁵
saving than variations in financial wealth. The authors found that for each $1 increase in real
estate wealth household saving would fall by 8 cents, while for each $1 increase in financial
wealth household saving would fall by just 2 cents. They suggest that the effect of variations in
financial wealth might be smaller because much financial wealth is tied up in pension and
insurance funds and gains are not easy to withdraw. They also point out that financial wealth is
concentrated and that some financial asset holding may represent controlling interests in
businesses and thus not readily liquidated.
Although there are reasons to think that housing price appreciation might not be a substitute for
saving, the empirical studies, while not quite amounting to a consensus, found evidence to
suggest that it might. The recent decline in house prices might begin the pose a problem if it
proves to be more than temporary and there is no offsetting increase in household saving.
In addition to equity and housing price appreciation, the recent sharp rise in energy prices has ¹⁶
also been offered as an explanation for some of the decline in household saving. But if energy
prices do not decline from recent highs that explanation does not alleviate concerns for the long
run.
Accounting issues have also been suggested as possible reasons for the decline in measured
saving. In particular, the fact that stock repurchases are not counted as part of personal income
may have caused both income and saving to be understated. There has been an increase in stock ¹⁷
repurchases in recent years.
Clearly, there is no shortage of potential explanations for the decline in household saving. A
recent study published by the St. Louis Federal Reserve Bank reviewed all of the potential
explanations and came to the conclusion that while each of them may have its merits, none of ¹⁸
them is compelling and that much of the decline in saving remained a puzzle.
Even if the decline in the saving rate could be explained, it does not answer the question of
whether Americans are saving enough, or accumulating enough wealth, to avoid experiencing a
decline in their living standard in retirement. One study attempted to estimate how much

¹⁵^J^o^hⁿ^D^.^B^e^nja^m^in,^P^e^te^r^C^h^inl^o^y^,^a^{nd G}^.^D^o^na^{ld J}^ud,^“^R^e^a^l^Es^ta^te^Ve^r^s^us^Fⁱ^na^nc^ia^{l W}^e^a^{lth i}ⁿ^C^o^ns^um^ption,”
^J^ourⁿ^a^{l o}^f^Re^al^Es^tate^Fin^anc^e^a^{nd Ec}^on^om^ic^s^,^vo^l^.²⁹^,^no^.³^,²⁰⁰⁴^,^pp^.³⁴¹^-³⁵^4.
¹⁶^Ch^arl^e^{s S}^t^ei^nd^el^,^“Ho^w^W^o^rri^so^{me Is a Neg}^a^tⁱ^v^{e S}^a^viⁿ^g^Rat^e^?,^{” F}^e^d^e^ral^{Reserve Ban}^k^o^f^Ne^w^Yo^rk,^Cu^rren^t^Issu^es
^{in Ec}^o^nom^ic^s^an^{d Fi}^na^nc^e^,^Ma^y²⁰⁰^{7, 7 pp.}
¹⁷^Ibid^.
¹⁸^M^a^ssi^m^o^G^uⁱ^d^o^lⁱⁿ^and^E^lⁱ^zab^et^{h A}^.^L^a^Jeuⁿ^e^sse,^“T^h^e^De^clⁱⁿ^{e i}ⁿ^t^h^{e U.}^S^.^P^e^rs^oⁿ^a^l^S^a^viⁿ^g^Rat^e^:^{Is It}^Real^an^d^{Is It}^a
^P^u^zzl^e?,^F^e^d^e^ral^{Reserve Ban}^k^o^f^S^t^.^L^o^ui^s^R^evi^ew^{, v}^o^l.^89,ⁿ^o^.^{6 (}^N^ov^em^be^r^/^D^e^c^e^m^b^e^r²⁰⁰⁷⁾^,^pp.⁴⁹¹^-⁵^14.

households should be setting aside in order to maintain their living standards into retirement.¹⁹
Using a retirement planning financial model and data from a survey of the financial status of
households approaching retirement, it was found that low-income households (those with
incomes below $15,000) needed to save only about 1% of their income. But that assumed that
there would be no cuts in Social Security benefits. When it is assumed that Social Security
benefits are cut by 30% in 15 years, the desired saving rate rises to 6%.
Households that have higher incomes need to save at higher rates if they are to avoid a decline in
their living standards on retirement, because Social Security benefits will account for a much
smaller share of their retirement income. The desired saving rate for those households with
incomes between $15,000 and $100,000 was about 14%, but rose to about 20% when it was
assumed that future Social Security benefits would be cut.

The drop in personal saving that occurred over the past 15 years may not be as troubling as might
first appear, at least with respect to retirement planning, for several reasons. The decline in
personal saving coincided with large increases in equity values and housing wealth. When
changes in personal wealth are taken into account, the decline in the official saving measure may
be less of an issue, although there will always be concerns about the durability of any gains in
asset prices. With respect to the household sector’s contribution to total national saving, that has
clearly fallen. Meanwhile investment spending has risen relative to gross domestic product
(GDP), in spite of the decline in personal saving, due to the increase in capital inflows from
abroad. Some of the productivity gains likely to result from that increased investment, however, ²⁰
will benefit foreign rather than domestic investors.
While the low saving rate may not pose an immediate economic risk, there could be significant
consequences if it persists. Households apparently felt a diminishing need to save out of current
income during the 1990s because of the large increase in stock prices. Stock prices have since
declined, but house price appreciation seems to have encouraged households to save even less.
An additional consideration is that the large federal government surpluses of the late 1990s have
now evaporated. For the immediate future, the federal government will continue spending more
than it is taking in, which means that the public sector saving rate will be much lower than it was
in the late 1990s. Without an increase in household or business saving, either investment spending
must fall, or the United States will have to continue importing substantial amounts of capital from
abroad.
If foreigners were to reduce or stop their investments in the United States, then the decline in
demand for U.S. financial assets would cause their prices to fall and interest rates would rise.
Without an offsetting increase in domestic saving, the rise in interest rates would be likely to
result in a decline in domestic investment.

¹⁹^{B. Do}^ug^la^{s Be}^rnhe^im^{, L}^o^re^nz^{o Forni,}^Ja^g^a^de^e^s^{h G}^o^k^h^a^l^e^,^a^{nd L}^a^ureⁿ^c^e^{J. Kotlik}^of^f^,^“^H^o^w^Muc^h^S^hou^{ld A}^m^e^rⁱ^c^aⁿ^s
^B^e^Sa^vⁱ^ng^f^o^r^R^e^tir^em^eⁿ^t?^,”^Fe^der^a^{l R}^e^s^e^r^v^e^B^a^nk^of^C^l^e^v^e^l^a^{nd W}^o^r^k^ing^P^a^pe^r⁰⁰^-⁰^{2, J}^aⁿ^u^a^r^y¹⁹⁹⁹^{, 13}^p^p^.
²⁰^Se^e^C^R^S^R^e^por^{t R}^L³³¹⁴⁰^,^I^s^t^h^e^U^.^S.^Tr^ade^D^e^fic^it^C^aus^e^d^by^a^G^l^oba^{l S}^a^vⁱⁿ^g^G^l^u^t^?^,^by^Ma^r^c^L^a^bon^te^.

^B^r^ian^W^.^Cash^ell
^Specialis^{t i}ⁿ^Macroecon^o^mⁱ^c^P^o^licy
^bcas^h^e^l^l^@^c^rs^.l^oc.g^ov^{, 7}^-⁷⁸¹⁶