Retirement Decisions of Women and Men in Response to their Own and Spousal Health

This dissertation studies the impact of individual and spousal health on the retirement decisions of both spouses in dual-earner families. The survival analysis techniques (Kaplan-Meier estimator, a Cox proportional hazards model, a Weibull specification) are used to analyze eight biennial waves (1992-2006) of the Health and Retirement Study (HRS), a nationally representative panel survey of the U.S. population over age 50. The modeling of retirement ages of both spouses and the time interval between their retirement dates provides a more accurate picture of the family retirement process than has been available in previous analyses. Of the various causes of early retirement, the onset of work disability has the biggest effect. The effect of functional disability is comparable, while major health events and chronic illnesses have smaller but statistically significant effects. The research findings also highlight the importance of correctly dating the exact timing of health deterioration to alleviate "justification bias" in empirical estimation of health effects. The estimates of the spousal health effects demonstrate that the onset of a husband's work disability can lead to an earlier age of retirement not only for the husband himself but also, through joint retirement, for his wife. This effect can be defined as a "joint early retirement" phenomenon. Husbands whose wives become ill also tend to retire at earlier ages. The "joint early retirement" phenomenon is more pronounced among families where the spouse remaining in the labor force is a low earner. The cost-of-illness estimates for indirect costs (productivity lost through an early retirement) of different health conditions are calculated at the individual and societal (adjusted for prevalence of conditions) levels. From a societal perspective, the most costly work disability conditions are those related to musculoskeletal and heart and circulatory conditions. At the individual level, the productivity losses are greatest from work disability related to heart and circulatory conditions and cancers and tumors. The total family productivity lost due to a husband's (a wife's) work disability is also estimated. The results of this dissertation have important implications for public policies related to labor force participation, Social Security reform, and health care.

Document number: RGSD-243

health status

21460