Abstract

This paper estimates the treatment effects of the two different training strategies followed by "Welfare-to-Work" (WTW) programs: Labor Force Attachment (LFA) and Human Capital Development (HCD). A key problem in estimating the effects of these two types of programs is that the available data sources do not identify which individuals have been subject to training. This paper presents a non-experimental econometric methodology that allows to identify the treatment effects even under unknown treatment status. The results, from estimations based on administrative data for California, suggest that the LFA programs have short term positive effects that fade out around two years after entry into welfare, while the effects of HCD programs are non-significant or negative, and the effects of being on welfare without receiving any training appear as negative in the first two years after welfare entry, and become positive afterwards. After welfare reform, however, the effects of both training programs appear as non significant, with the effect of no training strongly positive. This might be a consequence of the interaction of the new environment after welfare reform with the WTW programs.