Effects of Welfare Time Limits on Enrollment and Work: New Evidence from Welfare Reform Experiments

• In Event: Enrollment and Policy Effects of Assistance Programs Including SNAP, TANF, and Head Start

Abstract

The enforcement of a 60-month time limit on cash assistance under the 1996 welfare reform sparked a substantial and ongoing debate over whether such a policy benefits or harms disadvantaged families. A lifetime limit on welfare receipt may incentivize recipients to seek employment, thereby encouraging them to escape poverty and achieve economic self-sufficiency. However, cutting benefits before recipients reach financial independence may result in substantial economic hardship and leave them in a poverty trap.

This research reexamines how a strict time-limit policy affects welfare participation and labor supply among single mothers, using a previously unexplored source of data in the time-limit literature: experimental data from Connecticut’s Jobs First Program (JFP). To isolate the time limits’ effect from other TANF policy changes, I adopt the identification strategy first developed by Grogger and Michalopoulos (2003, hereafter GM03), which leverages the age of the youngest child to isolate the effect of time limits from those of the broader reform. Both under AFDC and its successor, TANF, families lose eligibility for cash welfare once their youngest child turns eighteen. Therefore, only families whose youngest child was age thirteen or younger when the 60-month time limit was implemented were affected by the time limit policy. Families whose youngest child was older than thirteen serve as a counterfactual control group, as they experienced the other components of the reform but were not bound by the time limit.

Compared to prior papers, results from the JFP data reveal novel and somewhat contrasting findings: Single mothers exposed to welfare time limits did not reduce their welfare use. Importantly, more people in Connecticut began to combine both welfare and work after the time limit was introduced. This finding is robust across various specifications, although it is not consistent with the predictions of GM03's model assuming rational and forward-looking individuals, and previous empirical findings.

Regarding the mechanism, I find suggestive evidence that Connecticut’s unusually generous earned income disregard policy might be the reason behind these different results. Connecticut's JFP granted the full benefit amount for working families as long as the family’s earnings remained below the federal poverty line. To test whether Connecticut’s EID rule drives the observed results, I estimate the “simulated” number of welfare-eligible months that would have occurred if the Connecticut sample had instead been subject to Florida’s less generous EID rules. When using this simulated welfare use as the dependent variable, the estimated effect of the time limit on welfare use becomes negative, consistent with the findings from the Florida sample. This simulation exercise suggests that if Connecticut’s EID rule had been less generous, welfare use would likely have declined during the pre-time-limit period following the implementation of the time limit policy. This new finding indicates that the effects of welfare time limits are context-dependent and should be understood within the broader context of other policies.

Author

• Seojung Oh, University of North Carolina, Charlotte
  Presenting Author