individuals: by providing jobs to those who want to work and by providing pension income to those who do not. The ability and willingness of private employers to offer jobs and pensions has been changing in the last two decades. Large changes have occurred, not only in age-based legislation, but in the economy's industrial base, the economic platform that provides both work opportunities and post-retirement income. Among other changes are the widely noted decline in the relative share of the large-firm, highly unionized sector and the expansion of the service sector.
This sectoral shift offers both opportunities and dangers to older individuals. On the one hand, employment in the service sector appears to be much more open to older Americans (Ruhm, 1990; Quinn, Burkhauser, and Myers, 1990), so that older individuals should have greater access to jobs in the future. On the other hand, they may need to work for more years if the expanding sectors remain relatively ungenerous in their pension offerings. Indeed recent evidence indicates that pension coverage in the United States declined in the 1980s, driven in part by relative employment declines in high-pension-coverage sectors (Woods, 1989; Parsons, 1991b, 1994; Bloom and Freeman, 1992; and Even and Macpherson, 1994a), although preliminary evidence from the May 1993 pension supplement indicates that coverage rates have stabilized in the 1990s (Woods, 1993; U.S. Department of Labor et al., 1994). At the same time, there has been a shift in the composition of pension coverage away from the traditional final-salary-based, retirement annuity programs (defined benefit plans) and toward savings-like, lump-sum programs (defined contribution plans) with payouts dependent on the returns-to-plan assets.3 These compositional shifts can also be partly explained by the relative decline in industries that traditionally offered defined benefit plans, although both the general decline in pensions and the shift toward defined contribution plans are evident within industrial sectors as well. The trend toward defined contribution plans introduces an unwelcome element of uncertainty into policy planning because we lack broad experience with such plans.
What are the strategies underlying employers' retirement and pension policies? How and, of equal importance, why do these strategies vary across firms, workers, and market structures?4 Although one can occasionally uncover direct evidence on employer intentions (Gratton, 1990), usually this information must be gleaned indirectly from observations of the age-related elements of employment conditions, including maximum age-of-hire restrictions, compulsory retirement rules, actuarially unfair pensions, age-restricted eligibility for retirement plans, and special retirement incentive programs or ''windows." The motivations for age-based policies are often ambiguous. With pensions, for example, the firm has two plausible objectives: (1) to serve as an efficient conduit for fringe benefits that its workers value, for example, as a mechanism for securing tax-deferred retirement income, and (2) to structure compensation in a way that induces desirable behavior from workers at least cost, shaping compensation