RETIREMENT

Retirement is primarily a twentieth-century phenomenon that developed through a convergence of public and private employment policies, a restructuring of the life span relative to work activity, and a redefinition of the terms of monetary compensation for work performed. It may be tempting to view retirement as the "natural" development of a social institution matched to the needs of older people experiencing declines in capacity; but the invention of a distinctive nonemployment status called retirement was not simply a response to human aging. Rather, in reconciling a transformed economy to an aging population with an increasing amount of surplus labor, an explicit policy of job distribution was produced. Retirement policies incorporated age as a characteristic that served as both a qualifying and an exclusionary principle for work and income. The fact that these policies were age-based can be linked to the social production of age as a predictor of individual capacity and potential, a production that had ideological roots in the science and larger culture of the time.

HISTORICAL DEVELOPMENTS

Whereas retirement contracts existed in both Europe and colonial America, Plakans (1989) argues that preindustrial retirement was a gradual transition. The head of a household transferred legal title to an heir in exchange for some combination of monetary payments, material provisions, and services as stipulated by the aged person or couple. These contracts were typical of agrarian economies in which land was the main factor in production; they represented the final step in a long and sometimes elaborate process of property transfer. These "stepping-down" practices were therefore most immediately linked to inheritance patterns; they could be used to ensure that family control of the land was maintained (Sorensen 1989).

Between 1790 and 1820, American legislatures introduced policies of mandatory retirement for certain categories of pubic officials. By the late 1800s, the majority of businesses still had no formal policies of fixed-age retirement. Instead, informal policies eliminated older workers from the labor force (Fischer 1977). This decline in the demand for older workers can be linked to changes in the structure of American capitalism. During the late 1800s the structure of American capitalism began to change from small-producer, competitive capitalism to large-scale corporate capitalism (Sklar 1988). Part of this reconstruction involved attempts to rationalize age relations in the workplace, a process that was embedded in a more general disenchantment with older workers and a devaluation of their skills. Indeed, the employment rates for men aged 65 and older showed a steady decline during this period, from 80.6 percent in 1870 to 60.2 percent in 1920 (Graebner 1980). According to Graebner's analysis, retirement became the impersonal and egalitarian method adopted by both public and private employers for dealing with superannuated workers. It allowed employers to routinize the dismissal of older workers, thereby restructuring the age composition of their workforces in a way they believed would enhance efficiency, a belief supported by the principles of scientific management. Pension plans legitimized this process and, at the same time, served as an effective labor control device.

The first pension plan (1875) is credited to the American Express Company, but benefits were restricted to permanently incapacitated workers who were at least 65 years old with a minimum of 20 years of service (Schulz 1976). In 1920, the first general piece of retirement legislation, the Civil Service Retirement Act, provided pension coverage for federal civilian employees. One year later, the Revenue Act of 1921 encouraged businesses to implement private plans by exempting both the income of pension and profit-sharing trusts and the employer contributions to these trusts from income tax liability. Nevertheless, coverage remained concentrated in a few industries, and 85 percent of the workforce continued to be without pension coverage.

By the 1930s, the problem of superannuated workers was coupled with the more general problem of managing surplus labor. The changing technology of the workplace helped transform the labor process. A subsequent increase in worker productivity and the growing recognition of the cyclical crises inherent in industrial capitalism broadened the concern beyond that of simple superannuation to that of job distribution and consumption capacity.

The Depression of the 1930s greatly exacerbated the growing problem of old-age poverty and unemployment. By 1935 unemployment rates among those 65 and older were well over 50 percent. Even those with pension benefits did not escape poverty; trade union plans were collapsing, and state and local governments were reducing or discontinuing pension payments (Olsen 1982). Legislative proposals for alleviating some of these problems included the Townsend Plan and the Lundeen Bill. The Townsend Plan proposed a flat $200 monthly pension for older Americans; recipients had to spend the pension within thirty days. The Lundeen Bill proposed benefits at the level of prevailing local wages for all unemployed workers aged 18 and older (including the elderly) until suitable employment was located. Neither of these plans was directly related to a retirement transition. The Townsend Plan granted equal benefits to all nonemployed persons over age 60. The Lundeen Bill focused more on job creation for workers of all ages than on limiting labor supply through age exclusions.

In 1934, President Franklin Roosevelt appointed the Committee on Economic Security (CES) to develop legislation to address the problems of oldage poverty and unemployment. The Social Security Act of 1935 offered a solution that based benefits on the level of workers' contributions to a general trust fund. Upon their retirement, covered workers (primarily those in manufacturing) could draw retirement benefits, assuming they met the age and work eligibility requirements.

For the CES, retirement referred to complete withdrawal from the labor force. As stated by Barbara Armstrong, an original member of the CES, "[r]etirement means that you've stopped working for pay." According to Armstrong, the option facing the Roosevelt administration pitted older workers against younger workers (Graebner 1980, p. 186). Retirement would reduce unemployment by moving older workers out of the labor force, allowing industries characterized by surplus labor to transfer jobs from older to younger workers. The federal government could facilitate this process by shifting the focus of older workers' financial dependency from the wage contract to a federal income maintenance program. In that sense, the Social Security Act of 1935 established what was primarily a program of old-age relief; its limited coverage and low benefit levels precluded its serving as an effective instrument of retirement. However, in establishing a measure of income support for retired older workers, the act reinforced the view that in the competition for jobs, age was a legitimate criterion, and youth had the "higher claim" (Graebner 1980). Ironically, the mobilization for World War II created job opportunities for older workers, as it did for women. Even though these opportunities proved temporary, they challenged the connection between retirement and superannuation, a connection that was asserted more emphatically when the supply of labor exceeded the demand.

During the next several decades, considerable growth in private pension plans occurred; coverage increased from 4 million workers in the late 1930s to 10 million workers in 1950 and 20 million workers in 1960. The expansion was spurred by a number of factors including the desire of firms to encourage loyalty and reduce turnover, favorable tax treatment, and the 1949 Supreme Court decision to uphold a National Labor Relations Board (NLRB) ruling that pensions were appropriate issues of negotiation through collective bargaining (Schulz 1976). During this same period, Social Security coverage was extended to more workers, and Congress continued to raise benefits in response to changes in the cost of living, although real benefit levels remained relatively low (Derthick 1979).

DECLINING RATES OF LABOR-FORCE PARTICIPATION

Early research on retirement was centrally concerned with the question of voluntarism in the retirement transition, as well as with the financial, social, and psychological consequences of leaving the labor force. Even though the expansion of both government and employer-based pensions had improved the economic situation of older people in retirement, poverty rates among the elderly were still high. By 1960, 35.2 percent of persons aged 65 and older were below the poverty line, compared with 22.4 percent of the general population. Poverty was the norm for older white women and older African Americans (U.S. Bureau of the Census 1987).

During the 1950s, the Social Security Administration began studying the characteristics of newly entitled beneficiaries. Initial reports stated that early retirement occurred primarily because of poor health and difficulties finding and keeping jobs; the availability of Social Security retirement benefits played a secondary role (Wentworth 1945; Stecker 1955). Although these studies relied on beneficiary-reported reasons for retirement, a measurement strategy that was criticized because of its susceptibility to social desirability bias, the findings cannot be totally discounted. Retirement in the 1950s was not a financially attractive status for many older workers. Given that retirement income programs offered "fixed" benefits (benefits that remained nominally the same but, with inflation, declined in real terms), the financial security of middle- and working-class retirees was in jeopardy.

During the 1950s, retirement became more common. Insurance companies led the way in developing "retirement preparation" programs, as researchers attempted to define strategies for "successful aging." In an era of postwar prosperity, retirement came to be viewed as a "period of potential enjoyment and creative experience which accrues as a social reward for a life-time of labor" (Donahue et al. 1960, p. 361). Researchers investigating the effect of retirement on life satisfaction found that "retirement does not cause a sudden deterioration in psychological health as [had] been asserted by other writers" (Streib and Schneider 1971, p. 161). Rather than rejecting retirement, advocacy groups for the elderly lobbied for improved conditions of retirement, including more generous pension benefits. Mandatory retirement had not yet become an issue. Instead, the trend was in the direction of earlier retirement, that is, before the age of 65. In 1956 women and in 1962 men were allowed to retire at age 62 by accepting actuarially reduced Social Security benefits.

During the mid-1960s, in the context of Lyndon Johnson's War on Poverty, the issue of old-age poverty was again addressed. In the Older Americans Act of 1965, Congress established a "new social contract for the aged" (Achenbaum 1983, p. 85) by specifying a series of objectives that, if met, would significantly improve the quality of life enjoyed by older people. Among these objectives was an "equal opportunity to the full and free enjoyment of . . . an adequate income in retirement in accordance with the American standard of living
. . . [and] retirement in health, honor, and dignity" (U.S. Department of Health, Education, and Welfare 1976, p. 2).

Richard Nixon's presidency inaugurated the era of modern retirement. Whereas previous amendments to the Social Security Act had brought more and more workers into the system, they had not significantly improved the level of retirement benefits (Munnell 1977). The presumption that Social Security benefits should serve as retirement income supplements rather than as the primary source of income had not been challenged. But the persistently high rates of old-age poverty lent credence to the charge that benefits were inadequate. During the decade following passage of the Older Americans Act, benefits were increased five times and indexed to changes in the consumer price index. Both the "real" level of benefits and the replacement rate of benefits to previous earnings were improved. Enhanced retirement benefits allowed workers to maintain their standard of living across the retirement transition and helped redefine retirement as a legitimate nonwork status that average-income workers could afford to enter voluntarily. During the 1970s, employer-sponsored pensions were also being reorganized. The 1974 passage of the Employee Retirement Income Security Act (ERISA) regularized vesting plans and provided workers with some protection against benefit loss. Private sector initiatives aimed at inducing early retirement were also increasingly common (Barfield and Morgan 1969). Until the 1970s, workers choosing early retirement virtually always accepted reduced benefits. During the 1970s, however, employers began to offer early retirement incentive plans. Not only did these plans pay benefits at younger ages, but the present value of the benefits often exceeded that of normal retirement benefits.

The parallel changes in labor-force participation rates and in poverty rates among the elderly are noteworthy. In the latter part of this century, labor-force participation rates at older ages declined significantly. During the 1970s, rates for men aged 55–64 dropped from 83 percent (1970) to 75.6 percent (1975) to 72.1 percent (1980) (U.S. Department of Labor 1983). In addition, at the beginning of the decade, 24.6 percent of those aged 65 and older were living below the poverty line, twice the 12.1 percent that characterized the general population. By the end of the decade, the poverty rate among the elderly had dropped to 15.2 percent, compared to an overall poverty rate of 11.7.

As Figure 1 illustrates, the changes in labor-force participation rates (derived from the Current Population Survey) differ by both age and gender. Rates for men and women aged 65 and older appear to have stabilized and perhaps marginally increased from their lowest point. Rates for men aged 55–59 and 60–64 show overall declines with recent hints of a slight upturn. In contrast, rates for women in these age ranges have steadily increased during the last quarter-century, with rates for women in their late fifties surpassing rates for men in their early sixties during the last decade.

DETERMINANTS OF RETIREMENT

Most current research examining retirement is of one of two types: national longitudinal studies that address the behavior of workers across a wide variety of occupations and industries (e.g., Quinn and Burkhauser, 1994; Hayward et al. 1989) and studies of specific firms that assess the retirement behavior of workers who share substantial commonality in workplace features (e.g, Hardy et al. 1996; Stock and Wise 1990). The former design maximizes variation in both work context and individual characteristics, using statistical controls to assess the relative impact of individual and job characteristics on retirement transitions. Because of the sampling strategy, results from these studies can be generalized to the national population of older workers. The second design—the case study—limits observations to workers who share a particular work context and attempts to explain variation in workers' responses to a common decisional matrix.

In both these designs, individualized models of retirement have been dominant, although the relative emphasis of "sociological" versus "economic" models may differ. Retirement models proposed by economists have emphasized the financial considerations involved in exchanging one income flow (e.g., wages and salary) for alternative sources of income (e.g., pensions and income from savings and investments). They have proposed measurement strategies based on the calculation of the present discounted values of the various income streams to attempt to disentangle relative effects. Sociological studies of retirement frequently focus either on the social psychological consequences of retirement or on the importance of occupational structure in shaping behavioral contingencies. Both opportunities and constraints are unequally distributed across workers. Whereas unionized manufacturing jobs may protect older workers through seniority systems, they also provide early retirement incentives through employer-based pension plans. Older workers are also vulnerable to plant closings and job dislocations that accompany mergers, downsizing, and cutbacks. This theoretical framework views retirement transitions as a career characteristic, with late-career behaviors being at least partially contingent on earlier career opportunities. In addition, retirement behavior is embedded in more general macroeconomic conditions. Rates of unemployment, changes in both government and employer pension programs, and the age structure of both the population and the labor force are implicated. Combining insights from both disciplines leads to models in which the financial trade-off is captured by the relative effects of current earnings and the present value of pension benefits; human capital (and, indirectly, the probability of alternative employment) through education, health status, skill level, and age; characteristics of career through years of seniority, history of recent layoffs, and overtime work; and family situation through marital status and the presence of children in the home.

MEASURING RETIREMENT

As the nature of retirement transitions changed, the question of measurement became more difficult. Armstrong's definition of "no longer working for pay" was being replaced by a variety of definitions oriented toward transitions out of career employment or full-time work, changes in major income sources (the receipt of Social Security or employer pension benefits), or changes in identity structures (e.g., through self-identification as a retiree). As definitions of retirement shifted toward receipt of pensions and the inclusion of part-time workers, the concerns of government turned toward the escalating cost of retirement and the advisability of delaying it. It became important to distinguish among exiting the labor force through "early" retirement, "regular" retirement, or disability, since these distinctions have implications for income replacement as well as labor-force reentry. Professionals, managers, and salespeople tend to delay retirement, whereas skilled and semiskilled blue-collar workers move more rapidly into retirement; clerical workers move more quickly into both retirement and disability statuses; and service workers experience relatively high rates of disability and death out of employment (Hayward et al. 1989). In addition, reentry into the labor force has become more common, with estimates of as many as one-third of retirees becoming reemployed, often within one to two years of their retirements. In short, the heterogeneity of what it means to be retired has increased considerably: it encompasses a broader age range; it involves diversity of income sources; and it allows for some level of postretirement employment.

ISSUES OF GENDER AND RACE

The development of retirement policy has been primarily oriented around the work careers of men, predominantly white men. The original Social Security program excluded industries in which women and blacks were concentrated. Although later amendments eventually covered these categories of workers, the benefit structure continued to reward long and continuous attachment to the labor force and to penalize workers for extended or frequent work interruptions. The temporal organization of women's lives relative to work and family, paid and unpaid labor, put women "off schedule" in accumulating claims to retirement income.

Spousal and survivors' benefits were designed to support couples and (primarily) widows during their later years. Research on women's retirement often focused on unmarried women and found that the determinants of retirement for women were similar to the determinants that had been identified for men. Although unmarried women and men differed in occupational locations, wages, health, and access to employer-sponsored pensions, these determinants appeared to sort unmarried women into retirees and workers in much the same way as they sorted men.

The pattern of women's labor-force participation has changed in recent decades, and more women—particularly more married women—are in the labor market. In fact, the trends in rates of labor-force participation for older women reflect both the increasing employment rates for successive cohorts of women and the tendency for more recent cohorts of older working women to retire at younger ages. The increase in dual-earner couples suggests that retirement decisions may be interdependent, with age differences, relative earnings, and the relative health of spouses figuring into joint decisions about careers, retirement, and postretirement employment.

Work and income disadvantages that are experienced at earlier stages of the life course cast a shadow on retirement transitions among minority group members. Lower earnings, lower job status, and discontinuity in labor-force attachment all undermine the financial platform for retirement. Work histories characterized by frequent spells of unemployment, illness, or temporary disability are linked to lower average retirement benefits and lower rates of savings and asset accumulation. In addition, African Americans are less likely to be married than whites and therefore more likely to be limited to their individual earnings and retirement resources. African Americans are more likely to exit work through disability and also more likely to continue to work intermittently after retirement. Gibson (1987) argues that disability can be used as another pathway to retirement for older African Americans, one that offers financial advantages. Because work histories can appear sporadic in old age as well as youth, establishing the timing of retirement as an event also can be difficult (Gibson 1987).

INTERNATIONAL COMPARISONS

Industrialization, economic development, demographic shifts, and politics are implicated in international comparisons of both the prevalence and the financing of retirement. Pampel (1985) reports that, among advanced industrial nations, a pattern of low labor-force participation among aged males is related to the level of industrialization and population aging. Cross-national comparisons of employment-to-population ratios demonstrate a continued decline in labor-force participation between 1970 and 1990 for men aged 55 and older. This decline is not, however, consistent across all age groups. Employment rates for men aged 55–59 have not shown the same proportional decrease as rates for men aged 60–64, or for those aged 65 and older. Compared to other Organization for Economic Cooperation and Development (OECD) countries (see Figure 2), Canada, Finland, Japan, Sweden, and the United States have relatively high rates of labor-force participation for men and women aged 65 and older.

Despite an overall downward trend in average age of retirement, nations continue to differ in the patterns of labor-force exits. Early retirement in European countries such as France, the Netherlands, and Germany remains the norm, with one-half to three-quarters of 60–64-year-old men out of the labor market (Guillemard and Rein 1993). Whereas early retirement in the United States was primarily financed through early retirement incentive programs (ERIPs) offered by private firms, more severe problems of unemployment in Europe fueled early retirement through expanded eligibility for state programs. Among some countries of western Europe and North America, disability programs also can operate as pseudo-retirement programs that allow workers to exit prior to normal retirement age. In contrast to the pattern of western Europe, Canada, and the United States, labor-force participation rates in Japan for men in all three age groups have been more resilient. Rates in 1990 remained relatively high, with more than one-third of men aged 65 and older participating in the labor force (Quinn and Burkhauser 1994). Cross-national comparisons of women's retirement patterns are more complicated. Whereas some countries show little change since 1970 (e.g., Canada, Australia, Italy, Japan, and the United States), others show patterns of labor-force withdrawal that parallel the trends for older men (e.g., Finland, France, West Germany, Spain, and the United Kingdom).

When comparing rates of labor-force participation, it is important to take into account national differences in census procedures, the definition of the labor force, and the kinds of activities that constitute "work." Because countries differ in the pathways workers take to retirement, using pension receipt as an indicator is also flawed. In Germany, for example, disability benefits and intermediate unemployment benefits also provide access to early retirement. Comparisons based on rates of labor-force participation confound country differences in full-time versus part-time employment, complete versus partial retirement, and unemployment and employment. In addition, cross-sectional figures do not allow a comparison of rates of withdrawal or reentry that would allow us to distinguish relatively stable from volatile labor markets.

Within each nation, policy development and social dynamics exert an important influence on retirement behavior. Recent debate in the United States and other countries has centered on the financial burdens of supporting a growing population of retirees and the desirability of reversing the trend toward early retirement. In 1983 the United States amended the Social Security Act to legislate a gradual increase in the age of full entitlement from 65 to 66 by 2009 and to 67 by 2027. Germany (1992 Pension Reform Act) and Italy have made similar policy changes. Japan tried to raise the retirement age from 60 to 65 but failed in their initial attempt. In 1990 Sweden succeeded in blocking the early retirement pathway through the disability fund, but in 1992 failed to abolish the partial retirement pension system. Even in France, where "old age is seen as a time of life when work is illegitimate (Guillemard 1983, p. 88) and where the legal retirement age was lowered to 60 in 1982, it is likely that the retirement age will be raised. This type of political reaction to demographic aging is one way of trying to shift the cost of early retirement from government programs to firms and individuals.

Other workplace policies linked to demographic aging, such as flexible retirement systems, are also being considered in countries like Germany, France, and Great Britain. Countries with lower rates of early exit, such as Japan and Sweden, provide an alternative approach to labor-market withdrawal which may alleviate the pressures of demographic aging. These countries structure retirement as a gradual process involving lowering wages, reducing hours, and reassigning workers. To some extent this model is already in place in the United States through "bridge jobs," in Sweden through partial retirement pensions, and in Japan through wage reduction arrangements.

DIRECTIONS FOR RESEARCH

Sociologists are also interested in the social context in which retirement decisions are made and retirement policies are developed. Family contexts, work contexts, economic contexts, and historical contexts all provide important frames of reference in which these behaviors are negotiated. To date, retirement research has refined both the measurement of concepts and the complexity of the behavioral models. Many of these refinements have involved the economic dimensions of retirement. The financial trade-off between pensions and wages, the changes in accumulated pension wealth, and the age-earnings profiles of different occupations have been captured in current models. What models of retirement continue to lack, however, is sensitivity to the social frames of reference (e.g., the shop floor, the office, the firm, the family, and the community). Studies that have addressed some of these issues (e.g., Hardy and Hazelrigg 1999) suggest, for example, that firm level features may be implicated in both the rate and the determinants of early retirement.

Retirement behavior may be primarily motivated by financial considerations. But given a threshold of financial security, perhaps the unfolding of the retirement process also involves the culture of the workplace, family dynamics, and societal values. Our first task is to theorize the social aspects of these decisions so that we can develop hypotheses. To test these hypotheses, a different data collection strategy is required—one that samples a sufficient number of observations of individual workers within sufficient numbers of organizational or, more generally, cultural contexts that can themselves be measured in terms of their salient characteristics. What has become clear is that retirement decisions are shaped by individual preferences, but that these preferences are shaped by the opportunities and constraints that workers encounter.

references

Achenbaum, W. Andrew 1983 Shades of Gray: Old Age,American Values, and Federal Policies since 1920. Boston: Little, Brown.

Atchley, Robert 1982 "Retirement: Leaving the World of Work." Annals of the Academy of Political and SocialScience 464:120–131.

Barfield, Richard E., and James N. Morgan 1969 EarlyRetirement: The Decision and the Experience. Ann Arbor: University of Michigan Press.

Derthick, Martha 1979 Policymaking for Social Security. Washington, D.C.: Brookings Institution.

Donahue, Wilma, Harold Orbach, and Otto Pollak 1960 "Retirement: The Emerging Social Pattern." In Clark Tibbitts, ed., Handbook of Social Gerontology. Chicago: University of Chicago Press.

Fischer, David Hackett 1977 Growing Old in America. New York: Oxford University Press.

Gibson, Rose 1987 "Reconceptualizing Retirement for Black Americans." Gerontologist 27(6):691–698.

Graebner, William 1980 A History of Retirement: TheMeaning and Function of an American Institution 1885–1978. New Haven, Conn.: Yale University Press.

Guillemard, Anne-Marie 1983 "The Making of Old Age Policy in France: Points of Debate, Issues at Stake, Underlying Social Relations." In Anne-Marie Guillemard, ed., Old Age and the Welfare State. Beverly Hills, Calif.: Sage.

——, and Martin Rein 1993 "Comparative Patterns of Retirement: Retirement Trends in Developed Societies." Annual Review of Sociology 19:469–503.

Hardy, Melissa A., and Lawrence Hazelrigg 1999 "A Multilevel Analysis of Early Retirement Decisions among Auto Workers in Plants with Different Futures." Research on Aging 21(2):275–303.

——, and Jill Quadagno 1996 Ending a Career in theAuto Industry. New York: Plenum.

Hayward, Mark D., Melissa A. Hardy, and William Grady 1989 "Labor Force Withdrawal Patterns among Older Men in the United States." Social Science Quarterly 70(2):425–448.

Munnell, Alicia H. 1977 The Future of Social Security. Washington, D.C.: Brookings Institution.

Olsen, Laura Katz 1982 The Political Economy of Aging:The State, Private Power, and Social Welfare. New York: Columbia University Press.

Pampel, Fred 1985 "Determinants of Labor Force Participation Rates of Aged Males in Developed and Developing Nations, 1965–1975." In Zena Blau, ed., Current Perspectives on Aging and the Life Cycle. Greenwich, Conn.: JAI Press.

Plakans, Andrejs 1989 "Stepping Down in Former Times: A Comparative Assessment of Retirement in Traditional Europe." In David Kertzer and K. Warner Schaie, eds., Age Structuring in Comparative Perspective. Hillsdale, N.J.: Erlbaum.

Quinn, Joseph F., and Richard V. Burkhauser 1994 "Retirement and Labor Force Behavior of the Elderly." In Linda Martin and Samuel H. Preston, eds., Demography of Aging. Washington, D.C.: National Academy Press.

Schulz, James H. 1976 The Economics of Aging. Belmont, Calif.: Wadsworth.

Sklar, Martin J. 1988 The Corporate Reconstruction ofAmerican Capitalism, 1890–1916. New York: Cambridge University Press.

Sorensen, Aage B. 1989 "Old Age, Retirement, and Inheritance." In David Kertzer and K. Warner Schaie, eds., Age Structuring in Comparative Perspective. Hillsdale, N.J.: Erlbaum.

Stecker, Margaret 1955 "Why Do Beneficiaries Retire? Who among Them Return to Work?" Social SecurityBulletin 18:3.

Stock, James H., and David A. Wise. 1990. "The Pension Inducement to Retire." In David A. Wise, ed., Issuesin the Economics of Aging. Chicago: University of Chicago Press.

Streib, Gordon F., and Clement Schneider 1971 Retirement in American Society. Ithaca, N.Y.: Cornell University Press.

U.S. Bureau of the Census 1987 Statistical Abstract of theUnited States 1986. Washington, D.C.: U.S. Government Printing Office.

U.S. Department of Health, Education, and Welfare 1976 Older Americans' Act of 1965, as Amended andRelated Acts, Bicentennial Compilation, March 1976. Washington, D.C.: Office of Human Development, Administration on Aging.

U.S. Department of Labor, Bureau of Labor Statistics 1983 Current Population Reports, ser. P-60, no. 148. Washington, D.C.: U.S. Government Printing Office.

Wentworth, Edna C. 1945 "Why Beneficiaries Retire." Social Security Bulletin 8:16–20.

Melissa A. Hardy
Kim Shuey