RETIREMENT, PATTERNS

The pattern of retirement at the end of the work career is shaped less by aging per se than by institutional mechanisms that provide incentives and support structures for workers’ exits from the labor force. The origins of retirement institutions across industrial societies have been traced to emergent economic and governmental conditions in the late nineteenth century. These institutions developed gradually until the mid-twentieth century, then rapidly over the next quarter-century, and are undergoing critical review and reorganization in the early twenty-first century. Factors driving these developments changed over time, and ranged from the politics of veterans’ pensions to class politics in the context of economic downturns and globalization. The twenty-first-century reorganization of retirement and other welfare institutions is being motivated by population aging, the growing insecurity of financial and labor markets stemming from global economic restructuring, and the changing nature of the family. These changes raise questions regarding the future of retirement as a standardized and permanent age-related transition from an income status based on employment to one based on transfers and assets at the end of the work career.

The institutionalization of retirement

The earliest pensions in U.S. history emerged in both public and private contexts. In the public sector, Union Army veterans’ pensions began benefiting northern white males in particular before 1900 (Skocpol). By 1900, thirty-five percent of white males age fifty-five to fifty-nine were receiving Civil War pensions (Costa). And while ill health and unemployment were factors in these retirement rates (Graebner), the availability of these pensions probably motivated their recipients’ exits from the labor force. Other early pension schemes are traceable to the private sector, where they were developed by employers to control the age composition of their workforces and to ensure labor stability. The American Express Company initiated a pension plan in 1875 for incapacitated workers who had been with the firm for twenty years or more. And in 1884 the Baltimore and Ohio Railroad instituted a pension plan that, in many respects, anticipated the major structure of twentieth century pensions. It called for mandatory retirement at age sixty-five after a minimum of ten years’ service. Age, years of service, and salary level were the elements of the benefit calculation.

The first three decades of the twentieth century brought a slow development of retirement plans although the number of workers covered by pensions, health insurance, profit-sharing, and other employee benefits grew steadily (Jacoby). Industrialization consisted of more than production lines; it included the development of personnel management and the growth of unionization, both of which would shape the boundaries of the regular work career with negotiated rule structures regulating hiring, promotion, wages, and income maintenance (including health insurance and retirement). Over time, these workplace regimes created age-structured work careers (Henretta).

The passage of the Social Security Act in 1935 stimulated the more rapid growth of pension plans by establishing minimum standards. Its setting of sixty-five as the age of eligibility eventually made that age the most common age for retiring (Costa). And, somewhat paradoxically, wage control policies during World War II further stimulated the development of more elaborate employee benefit packages (termed ‘‘the hidden payroll’’ by the U.S. Chambers of Commerce in 1947) to permit employers to compensate their most valued workers without violating wage limits. These highly valued workers constituted internal labor markets whose long-term commitments to their employers were shaped, in part, by back-loaded pension plans. Such plans encouraged long tenure but also presented strong incentives to retire by a specific age. Unions also negotiated or provided for these plans for their workers. Referred to as defined benefit plans, they were, in effect, earnings predictably derived from formulas linking age, years of service, and peak salary levels.

The institutionalization of retirement is best characterized as the regularized exit from the workforce when income changes from wages and salaries based on employment to public transfers like Social Security and private assets including employee pensions. In industrialized societies, retirement has become a universal transition among workers and their families. However, it has changed as economic fortunes have varied, the political climate has shifted, and the structural and demographic compositions of the workplace have been transformed. These changes are addressed below.

From early retirement to variable retirement

Economic prosperity followed by economic turndown in the post–World War II period led to changes in retirement patterns. By the end of the 1960s a trend toward early retirement emerged, especially among workers covered by private pensions in major manufacturing and related sectors. Early retirement meant that more and more workers began to voluntarily and permanently exit the labor force prior to Social Security eligibility, in large part because of eligibility for private pension benefits. Economic downturn and restructuring between the early 1970s and the early 1990s further accelerated early retirement, but in this period it was also motivated by incentive packages used by employers to trim their liabilities via plant closings and workforce downsizing (Hardy et al.). Two streams of early retirees developed by the 1980s, one representing the most privileged category of ‘‘pension elites’’ from professional, managerial, and skilled sectors and the other representing workers in declining industries whose early retirements were accelerated by employers’ efforts to downsize and reorganize their production and distribution systems.

These economic cycles coincided with the growth in government regulation of pension plans beginning with the Employee Retirement Income Security Act (ERISA) in 1974 and extending through the Retirement Equity Act of 1984, and the Pension Protection Act and later amendments of ERISA in the 1990s. These legislative actions extended minimum standards of reporting, disclosure, vesting, and benefit management, and sought to regulate plan termination policies by holding employers liable for their ‘‘pension promises.’’ These rising pension liabilities, coupled with growing market uncertainties, provoked employers to abandon the long-term contracts implicit in back-loaded, defined benefit plans and to adopt new pension instruments. The rapid spread of these new instruments, called defined contribution plans, redesigned the pension landscape at the end of the twentieth century. Defined contribution plans are retirement accounts invested by workers in a mix of equity, bond, or money market funds that do not promise a specific benefit level. Workers are primarily responsible for contributions and the financial risks associated with different investment mixes. Employers have limited responsibility and do not always contribute on behalf of employees. The accounts are portable across jobs and can usually be borrowed against, under strict rules of repayment and penalty. Thus they are highly individualized and reflect a departure from the defined benefit plans, born in the nineteenth century, that standardized the work career and the exit from it.

The effect on retirement patterns of these more widely adopted plans is not well established. However, the trend toward early retirement has reversed (Quinn; O’Rand and Henretta). Four patterns appear to have emerged more strongly than at earlier times. Workers are leaving their major jobs later, although still before normal Social Security ages. More workers, especially in the public sector, are reducing their work hours rather than leaving the labor force completely, following what has come to be called ‘‘phased retirement.’’ They are moving to post career jobs, sometimes referred to as ‘‘bridge jobs.’’ And they are returning to work after retirement. Accordingly, retirement is increasingly less ‘‘crisp’’ as a life transition. Jan Mutchler and her colleagues report that as early as the mid-1980s a significant minority of men age fifty-five to seventy-four who were observed over twenty-eight months followed patterns characterized by repeated exits and entrances and spells of unemployment. These patterns appear to have increased over the 1990s (Herz).

The volatility of labor and financial markets in the 1990s and its direct bearing on the performance of defined contribution accounts and on workers’ perceptions of economic security, are potentially important sources of the variability in retirement behaviors. Employers and employees are renegotiating the employment contract. The implicit ‘‘lifetime employment’’ model associated with industrialization and personnel policies over most of the twentieth century is being replaced with a more ‘‘contingent’’ model. Higher rates of job mobility, job displacement, and ‘‘retirements’’ follow. However, these retirements are increasingly less likely to be crisp and permanent. The growing responsibility of workers to manage their own retirements with less institutional support increases expectations of more diversity and heterogeneity in retirement behavior.

However, factors besides labor and financial markets probably also play important roles in these complex patterns (Kim and Moen). The first is that the workforce has become steadily more heterogeneous in its composition: more women and more minority groups participate in the labor market than during the peak period of voluntary early retirement. Women and ethnic minority workers are more likely to have less stable wage/salary histories, more job shifts, multiple spells of unemployment, and higher rates of contingent work. These workers are less likely to be covered by pensions, and when they are covered, they are more likely to be covered by defined contribution plans. There is evidence that these lower wage groups contribute proportionately less to their pensions and are more risk-averse in their choice of investment mix (Bajtelsmit and VanDerhei). In addition, these workers are less likely to be covered by retiree health insurance, another employee benefit that has influenced retirement patterns. Workers with this coverage tend to retire at higher rates. Trends beginning in the 1990s suggest that access to retiree health insurance is declining rapidly for all workers (U.S. Department of Labor) and may disappear for all but the most privileged labor sectors.

The changed demographic composition of the U.S. labor force has had implications for retirement patterns above and beyond the effects of differential access to employee benefits. Disparities in health over the life course allocate workers down different paths to retirement. Racial health disparities account significantly for pre-retirement labor force patterns. Black men and women are more likely to enter retirement via a disability pathway (as disability insurance recipients) or as previously unemployed (Flippen and Tienda). The concentration of Hispanic and African-American males in manual and blue-collar jobs with greater exposure to environmental hazards and occupational injury systematically selects them into disability categories prior to retirement eligibility (Hayward and Grady).

Finally, the recent trend toward increased active life expectancy, especially among white aging cohorts, may become a stronger force to further differentiate retirement behaviors. Active life expectancy is the average number of years of impairment-free life estimated for a population, based on the distribution of aged-based prevalence of functional impairment. Active life expectancy has increased across recent cohorts of those age sixty-five and older, meaning that the prevalence of functional impairments is occurring at later and later ages (Manton and Land), with a notable disparity in this trend between whites and Hispanics and African Americans (Hayward and Heron).

The improved health of current and future aging cohorts may have significant implications for retirement, especially in the context of an aging society. Population aging results from increased life expectancy and declining fertility, with fertility decline the more important factor. It is reflected in the increased ratio of older to younger population groups (e.g. in the ratio of those aged sixty-five and over to those younger or to those of working ages eighteen to sixty-four). Aging societies are confronted by the fiscal implications of these changing ratios. How can fewer workers support more retirees? The trend toward increased active life expectancy may in fact provide a solution to this problem. Healthier, older workers may choose to remain at work, and employers may choose to retain these workers in one fashion or another (e.g., through later retirement, phased retirement, or even post-retirement contingent reemployment). The coincidence of increased active life expectancy with public policies that delay eligibility for Social Security to later ages and that eliminate earnings limits while collecting Social Security, and with private policies that place more savings responsibilities on workers, has the potential impact of diversifying retirement even more, with major subgroups of early permanent retirees, intermittent retirees, later retirees, and so on. Consequently, a dominant pattern such as early retirement may be supplanted by multiple patterns.

Gender and retirement

Gender has differentiated retirement behavior over the twentieth century, although women’s and men’s work histories have become more similar. The household division of labor and women’s limited opportunities for pension acquisition based on their own employment resulted in the past in the prevalent tendencies of women to retire (1) earlier than men or jointly with their husbands; (2) as spouses, survivors (widows), divorced dependents, or welfare recipients (Supplemental Security Income recipients) whose employment histories do not qualify them for worker Social Security benefits; (3) with greater and increasing dependence on Social Security as income over retirement; and (4) at higher risk for poverty after retirement, especially following the death of a spouse and when reaching the oldest age categories.

Employer pensions and Social Security account for 80 percent of median incomes of retirement households (Clark and Quinn). However, the significant majority of retiree households with any employer pension income fall above the median total income of all households. Moreover, less than one-third of women over age sixty-five receive employer pensions, whereas nearly half of men over sixty-five receive pensions; women’s pension benefit levels average about 60 percent of men’s, mirroring preretirement wage ratios (O’Rand and Henretta). One consequence of this distribution of pension income across households and gender groups is women’s greater dependence on social retirement benefits (i.e. Social Security). When husbands with pensions die, their pensions may not continue to support their survivors over the remainder of retirement. About one-third of husbands still do not elect joint-survivor options, which offer them the opportunity to reduce their initial benefits and spread them longer into the future to support their survivors (Smeeding). Accordingly, as widows age, their incomes from private pensions held by their deceased spouses (if they had them) decline.

Extensions of the Social Security Act since 1935 have slowly progressed to meet the needs of older women, beginning with widows in the 1939 extension and finally reaching divorced spouses in marriages that lasted ten years or more in the 1983 amendment. However, the system still places women at risk for poverty with its survivor benefit policies. Women tend to outlive men and survive an average of fifteen years as widows. Social Security benefit rules result in a significant reduction in benefits following the spouse’s death, with women who have retired as workers rather than as dependent spouses often penalized more as widows. While this bias toward traditional male breadwinner couples can be criticized as unfair, the benefit reductions to both groups of widows are nevertheless at odds with women’s well-being, especially as out-of-pocket medical costs increase over the remainder of life and as other sources of income diminish (Smeeding).

Joint retirement. One source of variability in women’s retirement patterns now, and increasingly in the future, is their retirement to their own pensions and Social Security based on longer work careers over the life course. More than half of women between the ages of fifty-five and fifty-nine and nearly 40 percent of women in their early sixties were participating in the labor force by the late 1990s. Projections of future cohort participation patterns at these ages foretell significant increases in these rates (Bianchi). And, currently, women and men are about equally likely to participate in pension plans; approximately half of each group participates. Finally, approximately two-thirds of women currently retire as workers rather than as dependent spouses under Social Security (U.S. Social Security Administration). Accordingly, women’s retirement as dependent spouses is decreasing, although the continuing gender-based wage and pension gaps will slow their retirement income parity with men.

These changes in the economic roles of women suggest that the dynamics of decision-making in retirement may be shifting in a direction that will produce even more heterogeneity in couples’ retirement behaviors. Joint retirement may change from a process in which wives retire ahead of or with their husbands, based on their dependent statuses, to a more complex decision based on balancing independent opportunity sets and common preferences, especially within the context of the changing pension environment described earlier. Research on retirement of worker couples finds strong patterns of synchronization or joint exit among couples retiring between the 1970s and the 1990s (e.g., Henretta et al.; Gustman and Steinmeier; Blau). The studies also show that wives are more responsive to their spouses’ exits. Yet, the most recent patterns reveal that wives’ labor exits are also highly influenced by their own market characteristics, specifically their own earnings, health insurance coverage, and pension eligibility (Honig). The extent to which these characteristics will become more important in the retirement decisions of future cohorts is a matter of conjecture.

Solitary retirement. One final pattern of retirement with gender-related implications is solitary retirement. The growing proportions of never-married and divorced and unremarried persons entering retirement are changing the composition of the retired population. Historically, widowhood has been the predominant solitary status among retirees. However, the Social Security Administration has projected that by 2020 the widowhood rate will decline to 31 percent from 42 percent in 1991; less than half of women over age sixty-two (46 percent) will be married; and nearly one-fifth (19 percent) will be divorced, up from 6 percent in 1991. The likelihood of remarriage among women is lower than among men, and this disparity increases with age. Thus, among women, solitary retirement will include women without spouses and with fewer benefits from marriage, including disposable shared assets and social support. Eligibility for divorced spouse Social Security benefits does not provide adequate support against poverty, and poverty rates in the older population are highest among widowed, divorced, and never-married women (Smeeding).

The future of retirement

Population aging has proceeded faster in other advanced societies than in the United States, although the aging of the post–World War II baby boom cohorts is increasing the U.S. rate dramatically at the turn of the twenty-first century. Early retirement in Europe was a response primarily to public institutions erected to control the age composition of the labor force, making room for younger workers (Kohli et al.). Since the 1980s, however, the fiscal burden of these regimes has motivated even the most advanced welfare states (e.g., Sweden, Germany, the Netherlands) to develop flexible extensions of disability and unemployment compensation policies to manage high unemployment among younger age groups and to introduce more privatized pension schemes, including defined contribution accounts following the U.S. model, in order to ameliorate the crisis of public pension financing (Esping-Andersen).

Some scholars anchor the crisis of the welfare state primarily in the changing family and economic stagnation following global restructuring. Esping-Andersen argues that the welfare state was constructed on three institutional foundations: family reciprocity, market distribution, and state redistribution. Family reciprocity refers to the nonmonetary exchange of goods and services in families on which welfare programs have been predicated. The existence of unpaid domestic work by family members, especially wives and children, has always been a presumption of welfare policies. Social Security policies in the United States, described earlier, exemplify this taken-for-granted aspect of welfare policies. Except in the most socially democratic states like Sweden, the breadwinner (single wage earner) model has linked the labor market and the family directly to welfare programs on the basis of the industrial nuclear family with a gendered division of labor (Ginn et al.).

However, the market has changed: the decline in wages or jobs for men’s and women’s increased labor force participation has changed economic roles in the family; structural forces have increased the demand for female labor and decreased the demand for male labor. The family, in turn, also has changed: declines in fertility, delays in marriage, the increase in divorce, childbearing out of wedlock, and geographic mobility away from families of origin have transformed the ‘‘family reciprocity’’ system. In the absence of the family’s capacity to absorb social risks related to the welfare of its members and of the labor market’s capacity to sustain full employment, the three-legged stool of the welfare state is collapsing and impelling changes in family and labor policies, including retirement policies.

Alan Walker has argued further, and cogently, that the future of retirement institutions is integrally connected with the health status of aging populations: ‘‘If the health of workers is maintained then they will be willing and able to extend their working lives’’ (p. 14). Recent trends in increased active life expectancy in the United States and other countries raise hopes along these lines. However, health status over the life course is also supported by the three-legged stool of the welfare state. Health disparities begin early in the life course and become accentuated with differential exposures to life course hazards and health maintenance systems. In the United States, where health and economic disparities are among the highest among advanced countries, the market dominates health care over the life course and is available primarily through employment. Those outside the insured employment system are disadvantaged and at risk of ill health (Landerman et al.). The price of formal care over the life course—from premature births through nursing home care at the end of life—is rapidly increasing. In a world of diminishing informal support, the future of health maintenance and productive aging is less hopeful.

Health insurance systems are confronting crises of financing similar to those facing pension systems. Responses to these crises are familiar: increased privatization and healthcare rationing. Privatization shifts the system to the market, where efficiency criteria operate against universal health delivery and equal opportunity.

In short, interdependent demographic, market, and state forces have produced highly diverse life courses in advanced societies, including variable retirement patterns. The United States perhaps has the most variable retirement patterns because of stronger market and weaker state policies. However, all countries face the challenges of population aging, global markets, increased inequality, and family transformation that will require new views of the life course and new institutions to reconstruct it.

Angela M. O’Rand

See also Employee Retirement Income Security Act; Employment of Older Workers; Pensions; Population Aging; Social Security; Welfare State.

BIBLIOGRAPHY

Bajtelsmit, V. L., and VanDerhei, J. L. ‘‘Risk Aversion and Pension Investment Choices.’’ In Positioning Pensions for the Twenty First Century. Edited by M. S. Gordon, O. S. Mitchell, and M. M. Twinney. Philadelphia: Pension Research Council, University of Pennsylvania, 1997. Pages 45–66.

Bianchi, S. M. ‘‘Changing Economic Roles of Women and Men.’’ In State of the Union: America in the 1990s. Vol. 1: Economic Trends. Edited by R. Farley. New York: Russell Sage Foundation, 1995.

Blau, D. M. ‘‘The Labor Force Dynamics of Older Married Couples.’’ Journal of Labor Economics 16 (1998): 595–629.

Clark, R., and Quinn, J. F. The Economic Status of the Elderly. Medicare Brief no. 4. Washington, D.C.: National Academy of Social Insurance, 1999.

Costa, D. The Evolution of Retirement: An American Economic History, 1880–1990. Chicago: University of Chicago Press, 1998.

Esping-Andersen, G. Social Foundation of Postindustrial Economies. New York: Oxford University Press, 1999.

Flippen, C., and Tienda, M. ‘‘Pathways to Retirement: Patterns of Labor Force Participation and Labor Market Exit Among the Pre-Retirement Population by Race, Hispanic Origin and Sex.’’ Journal of Gerontology-Social Sciences 55B (2000): S14–S27.

Ginn, J.; Street, D.; and Arber, S., eds. Women, Work and Pensions: International Issues and Prospects. Buckingham, U.K.: Open University Press, 2001.

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

Gustman, A., and Steinmeier, T. Retirement in a Family Context: A Structural Model for Husbands and Wives. National Bureau of Economic Research Working Paper no. 4629. Cambridge Mass.: NBER, 1994.

Hardy, M.; Hazelrigg, L.; and Quadagno, J. Ending a Career in the Auto Industry. New York: Plenum, 1996.

Hayward, M. D., and Grady, W. R. ‘‘Work and Retirement among a Cohort of Older Men in the United States, 1966–1983.’’ Demography 27 (1990): 337–356.

Hayward, M. D., and Heron, M. ‘‘Racial Inequality in Active Life Expectancy Among Adult Americans.’’ Demography 36 (1999): 77–91.

Henretta, J. C. ‘‘Uniformity and Diversity: Life Course Institutionalization and Late Life Work Exit.’’ Sociological Quarterly 33 (1992): 265–279.

Henretta, J. C.; O’Rand, A. M.; and Chan, C. ‘‘Joint Role Investments and Synchronization of Retirement: A Sequential Approach to Couples’ Retirement Timing.’’ Social Forces 71 (1993): 981–1000.

Herz, D. ‘‘Work after Early Retirement: An Increasing Trend among Men.’’ Monthly Labor Review 118 (1995): 3–9.

Honig, M. ‘‘Married Women’s Retirement Expectations: Do Pensions and Social Security Matter?’’ American Economic Review Papers and Proceedings 88 (1998): 202–206.

Jacoby, S. Employing Bureaucracy: Managers, Unions and the Transformation of Work, 1900–1945. New York: Columbia University Press, 1985.

Kim, J. E., and Moen, P. ‘‘Moving into Retirement: Preparation and Transitions in Late Midlife.’’ In Handbook of Midlife Development.Edited by M. E. Lachman. New York: John Wiley and Sons, 2001. Pages 487–526.

Kohli, M.; Rein, M.; Guillemard, A. M.; and Van Gusteren, H., eds. Time for Retirement. Cambridge, U.K.: Cambridge University Press, 1991.

Landerman, L. R., et al. ‘‘Private Health Insurance Coverage and Disability Among Older Americans.’’ Journal of Gerontology-Social Sciences 53B (1998): S258–S266.

Manton, K. G., and Land, K. C. ‘‘Active Life Expectancy Estimates for U.S. Elderly Population: Multidimensional Continuous Mixture Model of Functional Change Applied to Completed Cohorts, 1982 to 1996.’’ Demography 37 (2000): 1–13.

Mutchler, J. E.; Burr, J. A.; Pienta, A. M.; and Massagli, M. P. ‘‘Pathways to Labor Force Exit: Work Transitions and Work Instability.’’ Journal of Gerontology-Social Sciences 52B (1997): S4–S12.

O’Rand, A. M., and Henretta, J. C. Age and Inequality: Diverse Pathways Through Later Life. Boulder, Colo.: Westview Press, 1999.

Quinn, J. F. ‘‘Retirement Trends and Patterns in the 1990s: The End of an Era?’’ Public Policy and Aging Report 8 (1997): 10–15.

Skocpol, T. Protecting Soldiers and Mothers: The Political Origins of Social Policy in the United States. Cambridge Mass.: Belknap, 1992.

Smeeding, T. M. ‘‘Social Security Reform: Improving Benefit Adequacy and Economic Security for Women.’’ Aging Studies Program Policy Brief no. 16. Syracuse, N.Y.: Maxwell School of Citizenship and Public Affairs/ Center for Policy Research, Syracuse University, 1999.

U.S. Department of Labor. Retirement Benefits of American Workers: New Findings From the September 1994 Current Population Survey. Washington, D.C.: U.S. Government Printing Office, 1995.

U.S. Social Security Administration. Annual Statistical Supplement to the Social Security Bulletin. Washington, D.C.: U.S. Government Printing Office, 2000.

Walker, A. ‘‘The Future of Pensions and Retirement in Europe: Towards Productive Ageing.’’ Hallym International Journal of Aging 1 (1999): 3–15.