Incomes and labor force activity are important influences on the health and welfare of our older citizen. Retirement behavior is a central determinant of incomes, and is a major descriptor of changes in labor force activity. The proposed project is a retirement study. It is designed to provide updated information on retirement behavior and related policy issues that otherwise will not be analyzed for years. A major emphasis is placed on the role of pensions in influencing retirement and on the proper measurement of pension incentives. The proposed study will estimate five versions of retirement models in order to incorporate into the study the recent developments in retirement analysis, to determine the models one against the other, and eventually to create a model which incorporates important strengths of these alternative approaches. Data from the Panel Study On Income Dynamics (PSID) and the Survey of Consumer Finances (SCF) will be used to analyze retirement behavior for more recent cohorts than have been analyzed with the Retirement History Survey (RHS), providing updated retirement analysis for a nationally representative sample sooner than would be available from the new Healthy and Retirement Survey. Thus the study will provide estimates of retirement behavior for more recent cohorts than those analyzed in other nationally representative studies. Moreover, the study will measure the likely effects on retirement of the changes in pensions over time and differences among demographic groups, and will consider the issues raised by errors in measuring pensions. The Survey of Consumer Finances has a matched panel of employer provided pension data. Particular emphasis will be given to the employer provided plan descriptions available with the SCF. Given firm and individual reported pension data, it is possible to analyze the effects of errors in measurement, when pension plan provisions are misreported by covered workers, on the parameter estimates that are obtained. Trends in pensions and their implications will also be analyzed. In addition, the project will use information available on these surveys, but not available in the older surveys, to anticipate and address a number of methodological issues which otherwise are unlikely to be addressed until the first few waves of the Health and Retirement Survey become available.
