An examination of the impact of federal disability programs on family labor supply: Evidence from the Health and Retirement Study

This set of essays recognizes the wide-ranging impact that a disabling condition can have on the family and thus examines the behavioral responses of spouses in addition to the individual with the health condition. The data used for all three essays are from the Health and Retirement Study (HRS)--both the public release data as well as Surface provides description only. Full text is available to ProQuest subscribers. Ask your Librarian for assistance. access earnings histories from the Social Security Administration and the Internal Revenue Service. Essay I uses a panel from the HRS in conjunction with a number of different nonexperimental econometric techniques to analyze the effects of a male\u27s disability application on the wife\u27s earnings and labor force participation. The econometric techniques employed in the essay range from the traditional OLS estimation, which is the most restrictive model, to a fully-interactive nonparametric cell matching technique, which imposes no functional form restrictions on the data. Essay I finds the impact of the husband\u27s disability application on his spouse\u27s labor supply to be dynamic. Behavioral changes are observed both before and after application and the size of these effects differ depending on the time to application. After accounting for the dynamic effect of disability, the different models provide similar results and follow the same trend. The different dynamic techniques show women reducing their labor supply well before disability application and that this pre-treatment effect increases on through the application date. Essay II compares the health and demographic backgrounds of successful and unsuccessful applicants to Social Security Disability Insurance and Supplemental Security Income. The essay also looks at the earnings for these groups and their trends in the years leading up to application. It finds that like successful applicants, denied applicants report having health problems at a high rate. In the years leading up to application, denied applicants actually earn and remain in the labor force at substantially lower rates than accepted applicants. Furthermore, after restricting the sample to those with strong earnings histories, a divergence between the two groups is observed as the application date approaches, with earnings for rejected applicants falling at a faster rate than for the accepted group. These patterns hold, even after controlling for health and other observable characteristics. Essay III once again examines women\u27s responses to declines in their spouses\u27 health. Here the duration from the onset of the male\u27s health condition until the wife changes income categories is examined. A discrete-time proportional hazard model (with and without corrections for heterogeneity) is used to estimate the wife\u27s probability of switching to a lower (or higher) income group in the years following the onset of her husband\u27s condition. Essay III finds that for those moving from the low to high-income category, the hazard function is initially declining, but there is positive duration dependence for a number of the post-onset years. This pattern is missed with the model that does not account for unobserved heterogeneity. For those moving from the high to low-income category, negative duration dependence prevails