This project will investigate the differences in assets held by persons of differing health status and the short-term effects of health on savings of elderly persons in the community. Using data from the Survey of Income and Program Participation, this study will compare the levels of total net worth and the portfolio of assets held by persons distinguished by aspects of their health, including self-reported health status, reported limitations in activities of daily living and mobility, reported days limited to bed, and prior hospitalization. Regression estimates, conditional on retirement status and corrected for possible bias in the data, will be made to estimate the effects of health on the level and composition of wealth. The study will also estimate saving rates for elderly persons using the longitudinal data from the survey. After grouping individuals and households by their reported characteristics, saving rates will be estimated using mean log rates of accumulation and a ratio estimator, both of which are consistent in the presence of some reporting errors in wealth. While reporting errors make individual level estimates of saving suspect, a simple regression model of the effects of the health variables on saving rates will be estimated, and the results compared to the previous grouped estimates. Results from the study can be used to refine simulation models which predict the future income and wealth of the elderly and their demand for private health and long term care insurance.
