This paper is concerned with the effects of training on the future employment probability of trainees. During the transformation process of the East German economy after reunification, employment decreased by a large amount. Therefore, many workers in East Germany participated in training programs which were heavily subsidized by the labor administration. However, at this point a thorough evahiation of such labor market programs does not exist. This paper attempts to fill this gap with regard to the impact of participation in training on the future employment probability of the trainees. Our analysis is based on the Arbeitsmarktmonitor Ost, a large labor market panel data set used by the German labor administration to examine the development of the East German labor market. We analyze the effects of two types of training: training within a firm and training in an institution outside of the firm, where the worker is employed. In our empirical analysis, we are very careful about the intriguing selection issiies involved, and we also take account of the severe attrition in the data set. Neglecting selection effects would lead to the result that training outside of the firm shows strong negative effects on employment probabilities whereas training in the firm shows strong positive effects. However, when taking account of selection effects, the results are completely reversed. Contrary to what is often alleged, we find in particular that training outside of the firm shows a considerable positive effect on the employment probability during the sample period.
training employment probabilites selection dynamic random-effects probit