Alexander Gelber, University of Pennsylvania and NBER
Timothy Moore, George Washington University
There have been increasing efforts to encourage Social Security Disability Insurance (SSDI) beneficiaries to work. The generosity of the SSDI benefit schedule may itself have important work incentive effects, as the rate at which previous earnings are replaced [in the conversion from Average Indexed Monthly Earnings (AIME) to the Primary Insurance Amount (PIA)] decreases from 90% to 32% to 15% to 0% as AIME increases. We propose to examine how the benefit schedule affects work activity using a regression kink design. Specifically, we will examine the relationship between the PIA-to-AIME ratio and employment of SSDI beneficiaries just above and below the three bend points created by this schedule in order to estimate the marginal effects of these changes on individuals earnings, the probability of trial work periods, and the probability of exiting SSDI due to improved health. We will also examine how these relationships vary by age, sex, type of disability, and business cycle conditions.