The estimates also indicate that farm residents were more than a third less sensitive to OAA than the rest of the population. This is consistent with Fisher (1946), who reported that, while 22% of the aged population were farm residents in 1940, only 16% of OAA recipients were farm residents in the twenty-two states studied in 1944. The lower levels of OAA recipiency among farm residents might arise because they had higher assets or household income precluding eligibility.43 Another possibility is that individuals left the farm when they retired and began to receive benefits, in which case the increase in benefits might have contributed to the move away from farms.

Lastly, nonwhites exhibit the greatest sensitivity of any group to OAA, along with the largest decline in participation during the decade. Part of the explanation is, no doubt, much lower income and wealth levels among racial minorities. Another factor might be declining labor demand in the South associated with the mechanization of cotton picking in the 1940s, work which had been largely done by blacks. This pattern might raise concerns that changing labor demand was correlated with changing state benefit levels, which would be reflected in the labor supply estimates here.44 For that reason, I re-estimated the relationship with whites only, as reported in column (6) of Table 7. The coefficient is almost identical and the conclusions about OAA are unchanged when nonwhites are excluded from the sample.

One last specification explores how OAA affects the employment rate, rather than the labor force participation rate. Focusing on participation supposes that job seekers did not receive OAA, perhaps because the process of applying for OAA was costly or because a job seeker was less likely to be deemed eligible. But if, instead, job seekers tended to receive OAA until they found employment, then the impact of OAA would be felt on employment rates more directly than on participation rates. A specification with the employment rate on the left hand side is reported in column (7) of Table 7. The estimated magnitude of the OAA main effect turns out to be quite similar, as does the aggregate impact of OAA. The aggregate employment rate rose by one percentage point, from 46.0% to 47.0%, between 1940 and 1950. The estimates suggest that if benefits had not risen, then the employment rate would have risen further to 48.2%, an additional 1.2 percentage points. Thus, the conclusions about OAA are as strong, whether considering its impact on the employment rate or the participation rate.