This article is part of a broader study of saving funded by the National Institute on Aging and TIAA-CREF.
The baby boom generation—the roughly 76 million people born between 1946 and 1964—has been reshaping American society for ve decades. From jamming the nation’s schools in the 1950s and 1960s, to crowding labor markets and housing markets in the 1970s and 1980s, to affecting consumption patterns almost continuously, boomers have altered economic patterns and institutions at each stage of their lives. Now that the leading edge of the generation has turned 50, the impending collision between the boomers and the nation’s retirement system is naturally catching the eye of policymakers and the boomers themselves.
Retirement income security in the United States has traditionally been based on the so-called three-legged stool: Social Security, private pensions, and other personal saving. Since World War II the system has served the elderly well: the poverty rate among elderly households fell from 35 percent in 1959 to 11 percent in 1995.
But the future is uncertain. Partly because of the demographic bulge created by the baby boomers, Social Security faces a long-term imbalance. The solution, even if it involves privatization, must in some way cut benefits or raise taxes. The private pension system has changed dramatically in ways that give workers increased discretion over participation, contribution, and investment decisions and easier access to pension funds before retirement—thus raising questions about how well future pensions can help finance retirement. Personal saving, also problematic, has remained anemic for over a decade. Net personal saving other than pensions has virtually disappeared.
These developments would be enough to raise concern about retirement preparations under the best of circumstances. But the prospect of a huge generation edging unprepared toward retirement raises worrisome questions about the living standards of the baby boomers in retirement, the concomitant pressure on government policies, and the stability of the nation’s retirement system.
Are the baby boomers making adequate preparations for retirement? In part, the answer depends on what is meant by “adequate.” One definition is to have enough resources to maintain preretirement living standards in retirement. A rule of thumb often used by financial planners is that retirees should be able to meet this goal by replacing 60-80 percent of preretirement income. Retired households can maintain their preretirement standard of living with less income because they have more leisure time, fewer household members, and lower expenses. Taxes are lower because retirees escape payroll taxes and the income tax is progressive. And mortgages have, for the most part, been paid off. On the other hand, older households may face higher and more uncertain medical expenses, even though they are covered by Medicare.
From a public policy perspective, assuring that retirees maintain 100 percent of preretirement living standards may be overly ambitious. But should policymakers aim to ensure that they maintain 90 percent of their living standards? Or that they stay out of poverty? Or use some other criterion? Retirement planning takes time, and these issues need to be addressed sooner rather than later.
A second big question is how to measure how well baby boomers are preparing for retirement. Studies that focus only on personal saving put aside for retirement yield bleak conclusions. One found that in 1991 the median household headed by a 65-69 year old had financial assets of only $14,000. But expanding the measure to include Social Security, pensions, housing, and other wealth boosts median wealth to about $270,000.
A third issue—crucial but as yet little explored—is which baby boomers are not providing adequately for retirement and how big the gap is between what they have and what they should have. Some boomers are doing extremely well, others quite poorly. Summary averages for an entire generation may not be useful as descriptions of the problem or as suggestions for policy.
The uncertain prospects for the baby boomers in retirement are particularly troubling because, as a society, we as yet understand little about the dynamics of retirement. Only one or two generations of Americans have had lengthy retirements, and the crucial retirement issues—health care, asset markets, Social Security, life span—keep changing rapidly, making long-term predictions even harder.
How Well Are the Boomers Doing?
Interpreting the Evidence
Only a few studies have examined how well the boomers are preparing for retirement. The Congressional Budget Office recently compared households aged 25-44 in 1989 (roughly the boomer cohort) with households the same age in 1962. Boomer households, it turned out, had more real income and a higher ratio of wealth to income than the earlier generation. Though this finding seems promising, in fact the CBO study implies that baby boomers are going to do well in retirement only if (i) the current generation of elderly is thought to be doing well, (ii) the retirement needs of the two generations are the same, (iii) the experience from middle age to retirement is the same for both, and (iv) boomers will be content to do as well in retirement as today’s retirees. None of these is certain. For example, although today’s elderly are generally thought to be doing well, some 18 percent were living below 125 percent of the poverty line in 1995. And the boomers’ longer life expectancy means that they will need greater wealth for retirement.
Whether the boomers and the previous generation will have similar experiences from middle age to retirement is an open—and still evolving—question. The earlier generation benefited from the growth of Social Security and housing values in the 1970s. But the boomers have gained from the dramatic rise in the stock market since the early 1980s, from smaller household size, which reduces living expenses, and from higher employment rates for women, which will raise their pension coverage. In addition, boomers are more likely to be in white-collar work and so should expect earnings to peak later in life and be able to work longer if they wish.
Finally, boomers may not be content with the living standard of today’s retirees. They may aim instead for retirement living standards more comparable to those of their own working years. For all these reasons, how to inter-pret CBO’s finding is unclear, even if the finding itself is unambiguous.
The most comprehensive study of these issues was undertaken by Stanford’s Douglas Bernheim in conjunction with Merrill Lynch. Bernheim developed an elaborate computer model that simulates households’ optimal saving and consumption choices over time, as a function of family size, earnings patterns, age, Social Security, pensions, and other factors. He then compared households’ actual saving with what the simulations indicated they should be saving. His primary finding, summarized in a “baby boomer retirement index,” is that boomers are saving only about a third of what they need to maintain preretirement living standards in retirement.
The index has attracted much attention but is not well understood. It does not measure the adequacy of saving by the ratio of total retirement resources (Social Security, pensions, and other assets) to total retirement needs (the wealth necessary on the eve of retirement to maintain preretirement living standards). Instead, it examines the ratio of “other assets” to the part of total needs not covered by Social Security and pensions.
As a result, the index reveals little about the overall adequacy of retirement preparations (see table 1). In case A, a hypothetical household needs to accumulate 100 units of wealth. It is on course to generate 61 in Social Security, 30 in pensions, and 3 in other assets. Total retirement resources are projected to be 94 percent of what is needed to maintain living standards. But according to the boomer index, the household is saving only 33 percent of what it needs.
|Table 1. Two Ways to Measure Adequacy of Retirement Saving
Units of wealth
|Boomer index (%)
Thus, a baby boomer index standing at one-third does not imply that, absent changes in saving behavior, boomers’ retirement living standards will be one-third their current living standard. It could mean that (as in case B), or it could mean retirement living standards will be 60 percent of current living standards (case C), or 94 percent (case A), or even over 99 percent (if Social Security and pensions were 99 and other saving were 0.33).
A second problem is that changes in the boomer index over time, or differences across groups, do not correspond to changes or differences in the adequacy of overall retirement saving. If, as in case D, the household in A rolls over its pension into an IRA, the boomer index soars, though total retirement resources are unchanged. If, as in case E, household A rolls over half of its pension into other assets and spends the rest on a vacation, the household has a higher boomer index, but less adequate total retirement preparation.
Finally, the boomer index can be extremely sensitive to estimates of retirement needs. In case F, retirement needs are 5 percent lower than in A, and the index rises from 33 percent to 75 percent. In case G, retirement needs are 7 percent lower than in A, and the index rises to 150 percent.
Bernheim points out that his model understates the retirement saving problem. The wealth measure, he notes, includes assets the household has earmarked for retirement as well as half of other (non-housing) wealth. The model also assumes no cuts in future Social Security benefits, no increases in Social Security taxes, and no increase in life span.
But in other ways the model overstates the problem. It assumes that any man not covered by a pension at the time of the survey, when respondents are 35-45 years old, will never be covered, though pension coverage rates tend to rise a good bit as a worker ages. The model also likely understates pension benefits since it uses benefit data from the 1970s. Because the pension system grew rapidly from the 1940s to the 1970s, workers retiring in the 1970s likely had fewer years in the pension system and hence lower benefits than the boomers will upon retiring.
The model excludes all housing wealth and inheritances—no small matter, since, by Bernheim’s calculation, including housing would raise the index to 70 percent, and a fair proportion of boomers is likely to receive substantial inheritances.
The model assumes that people will retire at age 65, though the normal Social Security retirement age will be 66 for most boomers, 67 for the youngest. The model also excludes all earnings after “retirement,” though about 18 percent of the income of the elderly today is from working. And with partial retirement on the increase, retired boomers may work even more regardless of the adequacy of saving.
Finally, the model makes no allowance for retirees’ lower work-related expenses or lower expenses for mortgages or other durable goods—such as furniture, appliances, and cars. Whether all these biases are larger or smaller than those in the opposite direction noted by Bernheim is unclear. Measuring and including these items is an important area for further research.
A New Perspective
Fundamental questions about retirement saving remain not only unanswered, but unasked. What proportion of households is saving adequately for retirement? What are the characteristics of those households? How has the proportion changed over time? Among those not saving enough, how big is the problem?
Table 2 begins to answer such questions by presenting my own estimates of the proportion of married households, with the husband working, who are “on track” toward accumulating enough wealth for retirement. The measure of “on track” is based on calculations in a study by Bernheim and John Karl Scholz, of the University of Wisconsin, that determines how much a household needs to have saved by a given age, given its earnings, prospective Social Security benefits, pension status, family size and other characteristics. (That study uses the Bernheim model described above, so the data suffer from all the biases already mentioned. Another bias here is that the sample includes only married couples where the husband works full time. Other married couples and singles are likely to be faring worse.)
|Table 2. Proportion of Married Households Saving Adequately for Retirement
|Proportion Saving Adequately When:|
|Year||Net assets exclude housing equity||Net assets include half of housing equity||Net assets include all of housing equity|
|Baby boomer householdsb|
|Source: Author’s calculations from the Survey of Consumer Finances.
a Husband is aged 25-64 and works at least 20 hours per week.
b Husband was born between 1946 and 1964 and works at least 20 hours a week.
When housing equity is not counted, slightly less than half of all households—and about the same share of all boomers—were saving adequately in 1992. When half (all) of housing equity is counted, the adequacy rate climbs to 61 percent (70 percent).
Adequacy rates rise with education and income. Within the baby boom generation, adequacy rates generally decline somewhat with age. They are higher for boomers with pensions than for those without, either because pensions raise households’ overall wealth or because people more oriented toward saving and thinking about retirement are also more likely to have jobs with pensions.
High adequacy rates do not necessarily require high levels of saving. For example, suppose annual retirement needs are 75 percent of final earnings. According to the Social Security Administration, Social Security benefits replace about 46 percent of final earnings for the average worker earning $50,000 at retirement. (Note that in this example Social Security replaces 61 percent—46/75—of total retirement needs, as in case A in table 1. The percentage would be higher for workers with lower earnings.) With pensions typically replacing 25-30 percent of final earnings, a household with Social Security and a pension would not need much more saving to maintain adequate living standards, especially if the household can work for a time in retirement or expects to receive bequests.
As table 3 shows, the wealth shortfall among households that are not saving adequately (ignoring all housing equity) is relatively small for many. The median inadequate saver has a shortfall of $22,000, or about six months of earnings—a problem that could be solved either by postponing retirement for six months or by lowering retirement living standards a little. Even among 60-64 year-olds, the median inadequate saver could completely resolve his or her saving shortfall by working for two more years past age 65.
|Table 3. Median Shortfall in Retirement Wealtha|
|Age||In dollars||In terms of annual earnings|
|Source: Author’s calculations based on the 1992 Survey of Consumer Finances.
a The sample is households not saving adequately for retirement when housing equity is not included.
Thus, the glass can be viewed as half full or half empty. When housing equity is ignored, the typical household seems to be barely saving adequately or just missing. When housing is included, over two-thirds of households appear to have more than the minimum needed, given their age and other factors. Roughly speaking, a third of the sample is doing well by any measure, a third is doing poorly by any measure, and the middle third is (or may be) just hanging in there. Both of the following statements are equally true. Up to two-thirds of the households are now saving at least as much as they should be. And two-thirds are “at risk” in that any deterioration in their situation could make it impossible for them to maintain their living standards in retirement.
In short, two key factors matter tremendously to any characterization of the problem: the heterogeneity of saving behavior across households and uncertainty concerning the right measures of wealth to use.
Areas of Uncertainty
The boomers’ prospects are also complicated by uncertainty in other areas: retirement patterns, life spans, home values, asset markets, health care costs, and the economy itself.
Average age at retirement, which fell through the 20th century for men, may start rising regardless of the adequacy of saving. Many of today’s jobs do not depend on “brawn” and can thus be done by older people. The normal Social Security retirement age will rise to 66 by 2009 and 67 by 2027 even if no further changes are made in Social Security.
Partial retirement may matter as well. Many retirees cut back on work gradually rather than abruptly. According to a study by economist Christopher Ruhm, only 36 percent of household heads retire immediately at the end of their career jobs. Nearly half remain in the labor force for at least ve years. Of workers eligible for a pension, 47 percent continue to work after leaving their career job. If people continue to work even after retirement, they will be better able to support living standards in retirement.
A related uncertainty involves life span. Expected remaining life spans of 65 year-olds have grown in the past two decades and are projected to grow further. Living longer means having to stretch a given amount of money over a longer period.
Uncertainty regarding home equity is twofold. First, how will housing prices evolve? Both demographic pressures and the reduction in tax rates in the 1980s may reduce the long-term value of housing. And, second, regardless of housing values, to what extent should housing be counted as part of household wealth? In recent decades, the elderly have been reluctant to cash in their housing equity. But baby boomers have been willing to extract housing equity and were major recipients of home equity lending booms in the 1980s and 1990s. It remains to be seen whether the boomers in retirement will act more like themselves in earlier years or like current retirees. In any case, a household with low financial assets that lives in a $300,000 house and refuses to dip into housing equity may not be considered a pressing social concern.
Asset markets too are uncertain. Equity values cannot continue to grow as rapidly as they did in 1996. And even if the boomers accumulate what seem to be sufficient retirement funds, they will, loosely speaking, all want to cash in those funds at roughly the same time. That could mean massive sell-offs of stocks and bonds that could depress asset prices. Conceivably asset prices could fall sharply, but since markets are forward looking, asset prices may instead be stagnant for a long period. Finally, the evolution of health care costs and of the economy as a whole could have a major impact on the adequacy of retirement preparations.
What’s in Store?
The retirement prospects for the baby boomers are uncertain. One issue is what policymakers and boomers themselves will accept as a reasonable goal for retirement living. More thought needs to be given as to how to assess living standards when, as a matter of biology, retirees face declining health. In addition, they typically have more leisure time and can literally substitute time for money. A second source of uncertainty is the boomers themselves. Whatever imponderables the economy as a whole may offer, baby boomers can improve their retirement prospects by saving more—that is, by reducing their current living standards.
What can government do? First, keep the fiscal house in order by reducing the long-term budget deficit in ways that do not reduce private saving. Second, the government could provide, or encourage others to provide, financial education to workers and households on how much they need to save. Third, the government should encourage people to use the many saving incentives already in place. Fourth, judicious Social Security and pension reform, especially pension reform that raises pension coverage, could help resolve these problems and raise private saving at the same time.