Association for Public Policy Analysis & Management

There have been renewed calls from policymakers to raise the retirement age to improve the financial viability of Social Security and help reduce the government's budget deficit. Such policy tends to reduce the prospective lifetime retirement wealth of beneficiaries conditional on income and life expectancy. This paper revisits the Social Security reforms of 1983, which provides an exogenous source of variation in retirement wealth across birth cohorts, to isolate the causal effects of a wealth shock on the labor supply of elderly workers in various stages of their life cycle. The key components of the amendments were an increase in the full retirement age (FRA) in two-month increments from 65 to 67 years old, and the related measured hike in the penalty for early benefit claimants from 20 to 30 percent starting with workers born in 1938. The adjustments were phased in gradually but the reforms were designed such that workers born between 1943 and 1954 are subject to the same FRA or penalty. Using data from the Current Population Surveys from 1993-2015, this paper exploits the nonlinearity in the design of the Social Security amendments to carefully estimate the impact of a large, potentially unanticipated wealth shock on labor outcomes. The identification from the difference-in-difference models rests on the assumption that, after controlling for age, education, or macroeconomic conditions, any observed trend-discontinuity in the labor supply of workers born before and after 1938 is due to the corresponding change in the prospective social security benefits. Preliminary results reveal increased labor supply in the pre-retirement age of 55-60, near-retirement age of 61-65, and post-retirement age of 66-70. A separate analysis of female workers with partners reveal that women contributed to increasing family income as well, with the strongest response observed in the near-retirement age group. The persistence of the effects past the normal retirement age, particularly among the lower-educated workers, suggest that the wealth shock have had effects beyond the normal working years of the elderly and that the lower-income classes are being hit harder.