Too bad to benefit? Effect heterogeneity of public training programs. IZA Discussion Paper 3240

This study analyzes the treatment effects of public training programs for the unemployed in Germany. Based on conditional propensity score match-ing methods we extend the picture that has been sketched in previous stud-ies. Besides estimating average treatment effects on the treated, we explicitly concentrate on treatment effects for different sub-groups of participants with respect to vocational education. Our results indicate that the effects of par-ticipation on employment, unemployment, and support probabilities involve effect heterogeneity. In particular, low-skilled individuals notably benefit from participation in the most important program type (occupation-related or gen-eral training). In context of the recent reform, we thus draw the conclusion that increasingly selecting the fittest unemployed into this program type is a questionable approach