|Essays on Retirement and the Residential Choice of the Elderly
|Year of Publication
|Doctor of Philosophy
|Michigan State University
|East Lansing, MI
The dissertation consists of three chapters concerning the well being of the older adult population in the United States. All three essays use data from the Health and Retirement Study (HRS), a longitudinal survey on health, retirement, and aging. The first chapter studies the relationship between retirement and time transfers within the family. We use cross-sectional variation in the need for caregiving to assess whether caregiving affects the retirement decision. For instance, parents' inability to take care of certain tasks for themselves, or accumulated and new health problems of spouses could lead to a demand for care from working individuals. We do not find evidence that potential parental care-need accelerates retirement transitions. Also, it does not seem that people retire early to provide care for sick spouses. However, post-retirement health insurance coverage is important in early retirement, in that when people have access to such coverage, they retire sooner if a family member happens to be in ill health. The second chapter assesses the potential usefulness of subjective expectation information in micro data by documenting the relationship between moving expectations and subsequent moving realizations among the United States population ages sixty-five and older. We find that the subjective probabilities of moving are very important in predicting future moving, even once demographic information known to be associated with the propensity to move is added to the analysis. Motivated by the observed relationship between the reported subjective probabilities and actual moving propensities, we hypothesize that when people are asked for a subjective probability they report the true probability conditional on available information, plus some random noise. We look at the proposed model's implications regarding which population groups are better at predicting future residential moving. However, we fail to substantiate the hypothesis, and therefore, cannot conclusively identify individual characteristics associated with better forecasting. The third chapter examines the long contested issue of whether the elderly draw down their housing wealth during retirement. In examining whether housing wealth declines during retirement, we emphasize exploring heterogeneity across population groups in housing wealth adjustments. Our analysis demonstrates that for non-mover retirees there is no systematic decline in housing equity. But for retiree-movers there is a decline in the median housing equity starting at age 71, and a decline in the mean housing equity from age 76. We find evidence of significant heterogeneity in housing equity adjustments at retirement. Nearly a quarter of the retiree-movers report that they are moving to downsize, and they do. Those with low non-housing wealth and with low income reduce housing equity significantly more than their respective counterparts. Retiree-movers experiencing widowing or divorce reduce housing equity substantially more than those without similar experience. Our findings are, in fact, largely consistent with the existing evidence in the literature regarding downsizing in later life. This study does not disprove or bolster either side of the debate on the role of housing wealth in financing retirement needs. But it highlights that the choice of emphasis regarding which side of the debate holds is often reliant on how one chooses to interpret what is in the data.