[An updated version of this article can be found at Saving in the 2nd edition.]
[Editor's note: this article was written in 1992. Since then, many of the data have changed.]
Saving means different things to different people. To some it means putting money in the bank. To others it means buying stocks or contributing to a pension plan. But to economists, saving means only one thing—consuming less in the present in order to consume more in the future.
An easy way to understand the economist's view of saving—and its importance for economic growth—is to consider an economy in which there is a single commodity, say, corn. The amount of corn on hand at any point in time can either be consumed (literally gobbled up) or saved. Any corn that is saved is immediately planted (invested), yielding more corn in the future. Hence, saving adds to the stock of corn in the ground, or in economic jargon, the stock of capital. The greater the stock of capital, the greater the amount of future corn, which can, in turn, either be consumed or saved.
Any country that saves little—that eats a high fraction of its seed corn—does so at the price of a lower future standard of living. The United States is an example of this grim law of saving. Since 1980 the annual rate of U.S. net national saving (net national output less private consumption expenditures less government consumption expenditures, all divided by net national product) has averaged only 4.2 percent. This saving rate is 60 percent smaller than the rate observed between 1950 and 1979! In the past few years saving has fallen even more. Since 1986 the United States has saved less than 3 percent of its net output. The saving rate in 1992 was only 2.8 percent.
The current rate of U.S. saving is remarkably low, not only by U.S. standards but also by international standards. Differences in how the statisticians in different countries define income and consumption make it difficult to compare saving rates across nations. But after correcting as well as possible for such data problems, one concludes that Americans are saving at less than one-third the rate of the Japanese and at less than half the rate of the Germans and many other Europeans. And as U.S. saving has declined, America's lead in per capita income (adjusted for purchasing power) has diminished.
Why do countries save at different rates? Economists do not know all the answers. Some of the factors that undoubtedly affect the amount people save are culture, differences in saving motives, economic growth, demographics, how many people in the economy are in the labor force, the insurability of risks, and economic policy. Each of these factors can influence saving at a point in time and produce changes in saving over time.
Motives for Saving
The famous life-cycle model of Nobel Laureate Franco Modigliani asserts that people save—accumulate assets—to finance their retirement, and they dissave—spend their assets—during retirement. The more young savers there are relative to old dissavers, the greater will be a nation's saving rate. Most economists believed for decades that this life-cycle model provided the main explanation of U.S. saving. But in the early eighties Lawrence H. Summers of Harvard and I showed that saving for retirement explains only about one-fifth of total U.S. wealth. Most of U.S. wealth accumulation—the remaining four-fifths—is saving that is ultimately bequeathed or given to younger generations. The motive for much of the substantial flow of bequests and gifts from older to younger Americans is surely altruism. But a large component of the bequests may be involuntary, and simply reflect the fact that many people do not spend all their savings before they die.
In recent years a much larger fraction of the retirement savings of the American elderly has been annuitized. That is, the savings take the form of company pensions or Social Security that pay regular checks until death, with no payments after the person dies. Having one's retirement finances come in the form of an annuity eliminates the risk of living longer than your money lasts. One possible result of the increased annuitization of retirement assets may be that people, especially those who have already retired, have less incentive to save more in case they "live too long."
That precautionary motive is one of the key reasons people save. Besides the risk of living longer than expected, people save against more mundane risks, like losing their jobs or incurring large uninsured medical expenses. Computer simulation studies show that the amount of precautionary saving can be very sensitive to the availability of insurance against these and other kinds of risks. For example, one such study that I did shows that the failure to insure low-risk, but high-cost, health expenditures such as nursing-home care can lead to a 10 percent increase in national saving.
Another issue related to motives and preferences for saving is the role of the rich in generating aggregate saving. Do rich Americans account for most of U.S. saving? Not really. Relative to their incomes, some of the rich save a lot, and some dissave. So, too, for the poor. As Donald Trump will tell you, there is considerable mobility of wealth in the United States, at least over long periods of time (see Human Capital). The fact that the ranks of the rich are continually changing suggests that some of those who are initially rich dissave and dissipate their wealth, while others who are not initially rich save considerable sums and become rich.
Economic Growth and Demographic Change
A country's saving rate and its economic growth are closely connected. In an economy experiencing no growth in either technology or population, one would expect, at least in the long run, saving to be zero, with the exception of the saving needed to replace depreciating capital. That an economy's overall long-run saving rate is zero doesn't mean that no one saves or dissaves. Rather, it means that the positive savings of those accumulating assets exactly balances the negative savings of those decumulating assets. For growing economies, long-run saving is likely to be positive to ensure that the stock of capital assets keeps pace with the number and technical abilities of workers.
In the United States the continuous decline in saving may be traced, in part, to a decline in technical progress. Economists measure technical progress as the increase in output that cannot be traced to increases in inputs—technical progress is a measure of the efficiency with which inputs are transformed into output. Since 1970, according to the U.S. Bureau of Labor Statistics, technology has been improving by only 0.57 percent per year, compared with 1.79 percent per year between 1950 and 1969. U.S. population growth has also declined. In combination, the slow growth in both technology and population spells a very low long-run rate of U.S. saving.
But demographic change should be causing U.S. saving to increase. In the last five years the numerous baby boomers have been reaching their primary saving years, which should have boosted the U.S. saving rate by several percentage points. That makes the fact that U.S. saving has been so low all the more surprising.
National saving is the difference between national output and national consumption. Labor income represents about three-quarters of U.S. output. So changes in labor income, if not accompanied by equivalent changes in consumption, can greatly affect an economy's saving rate. Take, for example, the recent remarkable increase in U.S. female labor force participation. In 1975 half of the women age twenty-five to forty-four participated in the labor force. In 1988 over two-thirds were in the labor force. This increase in female labor supply is the primary reason for the more than 20 percent increase in U.S. per capita income since 1975.
If this additional income had all been saved, the U.S. saving rate post-1980 would have exceeded 20 percent. Because much of the increase in labor supply was for women age eighteen to thirty-five, particularly married women, one would expect them to have saved some portion of that income for their old age. But they did not. This fact is another part of the recent U.S. saving puzzle.
Adding to the puzzle is the ongoing increase in the expected length of retirement. More and more Americans, particularly men, are retiring in their midfifties. At the same time, life expectancies continue to rise. Today's thirty-year-old male can expect to live to age seventy-four, 3.5 years longer than the typical thirty-year-old in 1960. If he retires at age fifty-five, today's thirty-year-old will spend almost half of his remaining life in retirement. Economic models of saving suggest that aggregate saving should depend strongly and positively on the length of retirement. Thus, with the retirement age coming down and life expectancy going up, economists would expect people to save a lot more—not a lot less.
Government policy also can have powerful effects on a nation's saving. To begin with, governments are, themselves, large consumers of goods and services. In the United States the government accounts for over one fifth of all national consumption. More government consumption spending does not, however, necessarily imply less national saving. If the private sector responds to a one-dollar increase in government consumption by reducing its own consumption by one dollar, aggregate saving remains unchanged.
The private sector's consumption response depends critically on who pays for the government's consumption and how the government extracts these payments. If the government assigns most of the tax burden to future generations (with, for example, our current pay-as-you-go Social Security system), current generations will have little reason, other than concern for their offspring, to reduce their consumption expenditures. If current generations are forced to pay for the government's spending, the size of the private sector consumption response will vary according to who among people living today foots the bill. The older the people taxed, the larger will be the reduction in consumption. The reason is that older people consume a higher percentage of their income than younger ones. Thus, taxing dollars away from retirees, say, instead of forty-year-old workers, will bring a larger decrease in private sector consumption.
Finally, different taxes have different incentive effects. For example, the government might raise its funds with taxes on capital rather than taxes on labor income. By lowering the after-tax return to saving, taxes on capital income discourage saving for future consumption.
Alternative Explanations of the Recent Decline in U.S. Saving
What explains the recent decline in U.S. saving? One explanation that can quickly be dismissed is that increased government consumption is to blame. The ratio of government consumption to national output since 1980 has been essentially the same as it was in the previous two decades.
Another explanation is that the U.S. government, by cutting income taxes and running large deficits, shifted the burden of paying for government consumption from current to future generations and induced a spending spree by current generations. This explanation, however, ignores other policies that redistributed away from current generations, like the baby boomers, toward future generations, such as their children. One important example is the 1983 Social Security Amendments, which cut baby boomers' prospective Social Security retirement benefits by roughly 20 percent in order to limit the payroll taxes their children would have to face when they join the work force. Of course, it is possible that people in their thirties and forties do not realize that their future benefits will be lower than was promised before 1983.
The argument that cutting federal income taxes raised disposable incomes and stimulated Americans to consume excessively also ignores increases in state and local taxes, as well as federal payroll taxes. On balance, disposable income as a share of net national output was only slightly higher in the eighties than in the previous decades. For all governments—federal, state, and local—the ratio of taxes, minus transfer payments (for example, welfare and Social Security), to U.S. output averaged .220 between 1980 and 1987. It averaged .226 for the seventies, .239 for the sixties, and .224 for the fifties.
Nor do disincentives to save appear to be responsible for the decline in U.S. saving in the eighties. Marginal personal tax rates on capital income fell through the last decade, with the top marginal rate declining from 70 percent in 1980 to 31 percent today. Individual Retirement Accounts (IRAs), intended to promote saving, probably reduced it. How could this be? Imagine that the government lets someone who invests in an IRA avoid taxes whose present value is $1,000. That person is then $1,000 wealthier. He will save some portion of this wealth and consume a portion. Say he saves $300, a generous estimate, and consumes $700. Then personal saving increases by $300. But assume that the government does not cut spending to finance this tax cut. Then the deficit rises by $1,000. The net result is that although personal saving rises, national saving falls by $700.
Most theories of consumption predict that households will increase their spending after their wealth increases. In the eighties real capital gains accruing to U.S. households on their holdings of equities were about $900 billion (measured in 1988 dollars). But in the eighties the real capital gains on the total portfolio of U.S. households, including all net assets (all assets net of all liabilities), measured in 1988 dollars was only $260 billion. It is true that most U.S. assets gained value in nominal terms—their dollar values at the end of the eighties exceeded their dollar values in 1980. But assets other than stock lost value in real terms. Their nominal appreciation failed to keep pace with inflation. The value of residential structures, most of which were owner-occupied housing, fell by over $600 billion between 1980 and 1989. The $260 billion figure represented only 1.7 percent of total U.S. net wealth as of the end of the 1980s and is too small to account for much of the decline in net national saving.
The eighties witnessed changes in income inequality, demographics, the expected duration of retirement, and female labor force participation, but as already mentioned, these changes should have led to more, not less, saving. One possible explanation for the recent decline in saving is a reduction in saving for bequests, which may tie in with the decline in the birthrate. Only significant additional research can test the validity of this explanation. At the moment, however, there is no "smoking gun" explanation for the critically low level of U.S. saving.
The Implications of Low Saving for Baby Boomers
Americans as a group used to save more than they did in the eighties. And as a consequence the collective stock of U.S. wealth holdings is still quite large—roughly $20 trillion. This is enough to finance all Americans' consumer expenditures for over five years. But 59 percent of this wealth is owned by people who are fifty or older, who appear to be spending a good deal of it on themselves. If the elderly do end up spending rather than bequeathing the bulk of existing U.S. wealth, will younger Americans, particularly baby boomers, accumulate enough savings to maintain the standard of living they currently enjoy in their old age?
Based on current evidence, the answer appears to be no. Compared with their parents, baby boomers can expect to retire earlier, live longer, rely less on inheritances, receive less help from children, experience slower real wage growth, face higher taxes, and replace a smaller fraction of their preretirement earnings with Social Security retirement benefits. Unless baby boomers change their saving habits and change them quickly, they may experience much higher rates of poverty in their old age than those currently observed among U.S. elderly.
Laurence J. Kotlikoff is a professor of economics at Boston University and a research associate with the National Bureau of Economic Research. He was previously a senior economist with the President's Council of Economic Advisers.
Auerbach, Alan J., and Laurence J. Kotlikoff. Dynamic Fiscal Policy. 1987.
Kotlikoff, Laurence J. Generational Accounting: Knowing Who Pays and When for What We Spend. 1992.
Kotlikoff, Laurence J. What Determines Savings? Chap. 6. 1989.
Kotlikoff, Laurence J., and Lawrence H. Summers. "The Adequacy of Saving." American Economic Review 72, no. 5 (1982): 1056-69.
Kotlikoff, Laurence J., and Lawrence H. Summers. "The Role of Intergenerational Transfers in Aggregate Capital Formation." Journal of Political Economy 89, no. 4 (1981): 706-32.