Reassessing the Impact of Welfare Reform on Female Employment and Labor Supply

Jeffrey T. Lewis, University of Maryland

The anomalous paper that does not conclude that TANF boosted female employment is Schoeni and Blank (SB). I determine whether or not, accepting the SB model but examining a different time period or utilizing a different data set, one can uncover evidence that TANF increased female employment. Examining a shorter time period (1989-2000) or using a different data set (CPS-ORG) than SB, I find employment effects of TANF. The researchers who do not restrict their attention to some group of single women (Moffitt and SB) run regressions in which they pool women from different education groups into the same sample. Interacting the reform dummies with the education groups, the authors expect reform effects to be concentrated among dropouts. They do not, however, interact every independent variable with the education dummies. Running regressions fully stratified by education group, I conclude that previous results in the literature are sensitive to model specification.