Evaluating the impact of a mandatory job search program: evidence from a large longitudinal dataset

Abstract

We assess the impact on unemployment duration and re-employment wages of a large scale employment program implemented in Portugal during the late 1990's and early 2000's. Based on a dataset covering over 1.5 million individuals, we construct a pooled set of treatment and control groups to infer the effect of the program. We test the program effectiveness on two outcome variables: the unemployment duration and wages after unemployment. Based on the class of difference-in-differences matching methods, the estimates of the average effects of the treatment on the treated point towards: (i) a reduction in unemployment spells smaller than 1 month, rather small in the context of long unemployment spells observed in Portugal and (ii) typically a reduction, if statistically insignificant, of the re-employment wages.