The Evolution of Retirement: An American Economic History, 1880-1990.

Burkhauser, Richard V.

By Dora L. Costa.

Chicago: The University of Chicago Press, 1998. Pp. xiii, 234. $40.00.

Over the last decade and a half, the National Institute on Aging and the National Science Foundation have provided substantial support to researchers at the Social History Laboratory at the University of Minnesota and at the University of Chicago to develop computer-ready, public-use versions of the early United States Censuses and other major pre-1940 federal and state data sets. Costa is one of a number of econometrically oriented economic historians whose research demonstrates that this major investment in intellectual capital has been worthwhile.

This book springs from four recent journal articles by Costa that use Union Army pension program and early Census records. The articles show how the massive Union Army pension system, which hit its peak - 30% of all Federal expenditures in 1900 - in the early years of the 20th century, influenced work and living arrangements at older ages as well as the importance of income, health, and the decline of agriculture in the long-run fall in the labor supply of older males. Costa builds on this earlier research to provide the more general reader with a consistent story of how and why retirement grew as an institution in the United States, especially over the first half of the century. She then infers from this story how the current crisis in Social Security is likely to impact on future retirement trends.

In the early chapters, Costa uses census data to show that the decline in the labor force participation of older men that has dominated modern - post-World War II - labor markets began as early as 1880. The major argument in the book is that this decline has primarily been caused by rising income at older ages. Her evidence largely comes from the Union Army pension data. Costa finds that pension income significantly increases the likelihood of labor market nonparticipation. With the elasticities from her empirical estimations, she argues that secular increases in income explain most of the increase in retirement rates prior to 1940. Costa's elasticities overestimate more recent falls in participation rates and are higher than those reported in the literature with more contemporary data. In a later chapter, Costa argues that the fall in the price of leisure due to technological change and the increased public provision of leisure-related goods explain this modern decline in the importance of income.

Costa debunks two other explanations for the fall in labor supply at older ages. First, she uses the medical records in the Union Army pension data to show that just as today, health played an important role in the decision of Civil War veterans to retire. However, she also shows that the rate of chronic disease among older workers in the first half of the century was much higher then among modern workers, implying that retirement increased over the century despite, not because of, improving health.

She then argues that a reduced demand for older workers relative to younger workers is also not a plausible explanation for increased retirement. By synthetic cohort techniques, Costa shows that while farming declined dramatically over the first half of the 20th century, this could not have increased retirement, since farmers retired at the same rate as nonfarm workers over this period.

Another chapter based on her past work finds that Union Army pensions reduced co-residence of retired veterans. This finding suggests that the desire of older persons to live independently from their children is not a recent phenomenon. When they had the means to do so, turn of the century retirees also preferred separate residences. The important policy message of these collective findings is that the century-long rise in retirement and in the independent living of retired workers has primarily been a voluntary response to increases in income at older ages.

The final two chapters provide a broader public policy context to the book. In them, Costa reviews the history of public programs targeted at older people and the demographic pressures - an aging baby boom population and increasing life expectancy - that are putting a financial strain on the current Social Security system. She acknowledges that a reversal of the century-long decline in the labor force participation of men is one solution to the system's financial problems. But she is skeptical that reductions in benefits or an extension of the early or normal retirement age is likely to have a large effect on work because today's aged are less sensitive to income changes than were Union Army pension beneficiaries. She suggests that a reduction in consumption is equally possible.

Costa makes a strong case for the importance of secular increases in income relative to changes in taste, health, or demand as the major force in encouraging retirement. But because she uses a single period rather than a life cycle approach in her conceptualization of the influence of pensions on behavior, she misses a critical component of most pension systems that also influences the timing of retirement.

It is certainly true that technological change has dramatically increased real wealth and, hence, income in the United States and that increased wealth has been used to fund more life cycle leisure. However, it is far from clear that workers would have demanded that leisure in the form of a complete exit from the labor market at older ages in the absence of a mandatory government pension program whose real wealth value fell for those who postponed acceptance past a given age. Nor is it obvious that the typical male retirement age in the United States would have continued to fall through 1985 in the face of substantial increases in life expectancy without the incentives to retire that are contained in this program. In the absence of the across-year labor leisure price effects that non-actuarially fair, defined-benefit public and private pension systems send, it is much more likely that the pure wealth effect of technological change would have been a greater reduction in hours worked per week or in weeks worked per year than in additional years of leisure at the end of life.

Costa is unable to distinguish the relative importance of this life cycle price effect from the income effect she focuses on in her work because the Union Army pension and the census data are cross sections. This is less a criticism of her empirical work than of the inferences she draws from it with respect to the importance of social security policy on the retirement age of older workers and how future policy changes may impact labor supply.

Costa's Census-based analyses end in 1990, too early to pick up the end of the long trend in earlier retirement that occurred in 1985. Since 1985, the labor force participation rates of older men in the age ranges she considers have remained constant or increased. While the evidence is far from conclusive, it appears that this shift has been caused, at least to some degree, by reducing the age-related price effects in Social Security.

My reading of the evidence suggests that retirement behavior is likely to be symmetrical. Government policies in the United States have had a major long-run impact on the ages over which leisure is consumed. When this is added to the facts that male workers reaching age 62 in the 21st century will be far healthier and have longer life expectancies and greater levels of human capital than previous generations, I am more convinced than Costa appears to be by her evidence in her final chapter that these men would substantially increase their work effort in the absence of Social Security and private pension incentives to retire at 62.

Despite these caveats, I highly recommend this book to anyone interested in the influence of government programs on retirement behavior. For those who have not read the major journal articles on which the book is based, it is a useful introduction to them. For those who have, the book puts this earlier work into a broader context that more fully explains the forces behind the evolution of retirement in the United States over the 20th century and how retirement behavior is likely to change over the next century.

Richard V. Burkhauser Cornell University