In sum, the aim of studying Old Age Assistance is twofold: understanding the role of OAA in explaining early retirement trends, and exploring how means tested benefits affected retirement behavior in the past. The present day applicability of these labor supply estimates cannot be directly inferred from this work. Major changes in the health status, wealth, and annuitization of post-retirement income make for very different living and working conditions for older individuals today. Nonetheless, research on OAA offers a unique and potentially relevant perspective.
My approach in this paper is to use the rich individual level data available from the 1940 and 1950 Censuses to estimate the impact of OAA benefit levels on labor force participation. Parsons (1991) also analyzed the impact of OAA. He estimated that 50% of the increase in retirement rates between 1930 and 1950 could be attributed to OAA. By including 1930, however, Parsons was compelled to use aggregated state level data because individual records from the 1930 Census are not available. Using individual level data instead makes it easier to control for personal characteristics correlated with labor supply and to study how OAA affects individuals with different characteristics. The Census data is still limited, though, because it does not include information on actual or potential OAA benefits. The available OAA data are average state level benefits. Nevertheless, controlling for individual covariates, state and year fixed effects, and state level measures of economic prosperity, I estimate a strong negative effect of OAA. If OAA benefits had not been raised during the 1940s, participation rates would have risen slightly instead of falling, as strong economic growth was encouraging people to work longer.

These results are broadly in line with Costa (1995), who estimated a sizable income elasticity of retirement at the turn of the century, also based on policy variation affecting the incentive to retire.

The rest of this paper consists of five sections. Section I describes the OAA program, Section II reports trends in labor force participation, and Section III discusses the empirical approach and data. Section IV analyzes the estimates and presents evidence against the possible endogeneity of state benefits. Section V concludes.