This paper evaluates the effects of Public Sponsored Training in East Germany in the context of reiterated treatments. Selection bias based on observed characteristics is corrected for by applying kernel matching based on the propensity score. We control for further selection and the presence of Ashenfelter's Dip before the program with conditional difference-in differences estimators. Training as a first treatment shows insignificant effects on the transition rates. The effect of program sequences and the incremental effect of a second program on the reemployment probability are insignificant. However, the incremental effect on the probability to remain employed is slightly positive.


Bergemann, Annette
Fitzenberger, Bernd
Speckesser, Stefan


Evaluation of active labor market policy in East Germany, nonparametric matching, conditional difference-in-differences, employment dynamics, Ashenfelter&acutes Dip, bootstrap