Three Essays on Social Security Disability Insurance.

Abstract

This dissertation studies two central aspects of the Social Security Disability Insurance (DI) program: economic factors influencing application decisions and the program's employment effect.

The first chapter documents that most applications for DI occur after a job loss, and most of them report a health shock shortly beforehand. However, applicants typically do not apply immediately after having stopped working. Classifying non-employment episodes ending with DI application by beginning events and comparing their first and last month reveals notable differences in terms of job search, transfer recipiency, and other factors likely to influence applications decisions for DI.

Building on these insights, the second chapter examines how Unemployment Insurance affects the decision to apply for DI. Theoretically, higher UI benefits may make it less necessary to obtain DI cash benefits, but they also facilitate applications by providing income support during the application process. Workers may also believe that UI take-up reduces the chances of a successful DI application. These predictions are tested by jointly estimating a hazard model of DI application and UI take-up. Results show that both UI take-up and a higher UI benefit level reduce the probability of applying for DI, which suggest that application decisions for DI are sensitive to short-term income flows. However, increasing UI benefits to reduce expenses for DI is not cost-effective because such a policy does not target workers at risk of applying for DI well enough.

The third chapter, which is joint with John Bound and Timothy Waidmann, turns to the employment effect of the DI program. Between 1990 and 2005, enrollment for DI rapidly increased while employment rates for men with disabilities decreased, suggesting a strong employment effect. However, studies using denied applicants find a low to moderate employment effect of DI for the majority of applicants. In order to reconcile these findings, changes in total employment are decomposed into population and employment changes for non-participants, DI beneficiaries, and denied applicants. Results for men with disabilities show that the DI program played at most a moderate role in explaining the continuing employment decline, especially after the mid-1990s.