This paper explores whether the Old Age Assistance program, the first means tested transfer program for the elderly, induced individuals to retire from work. OAA was strongly means-tested, benefit levels were generous relative to income levels of the day, and OAA recipiency rates were high. Using individual records from the 1940 and 1950 Censuses, I estimate a substantial effect of OAA on labor force participation. Yet, economists have had trouble attributing the much larger decline in participation that followed to Social Security, even though Social Security came to be received by most of the population. A major problem in quantifying the impact of Social Security or pensions on retirement arises because benefit levels do not vary across the population randomly but depend on past earnings. Similar difficulties have hindered investigation of the current Supplemental Security Income program for the poor elderly. Studying OAA gets around this problem because benefit levels varied across states. The impact of OAA is estimated here from the relationship of changes in participation to changes in benefit generosity within states between 1940 and 1950. The conclusion is that the growth of means tested transfers for the elderly played a significant role in the trend towards earlier retirement.

The estimates suggest that participation would have risen slightly, instead of falling, if not for the 20% increase in real OAA benefits between 1940 and 1950. The incentive to withdraw from the labor force remained even while the prospering economy acted to encourage individuals to stay in the labor force, mitigating the decline in participation. Additional data on slightly younger men offers evidence that benefits were not set endogenously with labor supply trends.

The impact of expansions in the generosity of Social Security during the following decade is consequently predicted to be large, explaining at least one-third of the decline in participation during the 1950s.