Evaluating the employment effects of public sector sponsored training in East Germany

Abstract

This paper investigates the employment effects of Further Training in East Germany for 1990 to 1999. We estimate the effect of treatment-on-the-treated for single programs and program sequences. We correct for selection bias by applying nearest neighbor or kernel matching together with a conditional difference-in-differences estimator. This estimator is based on the long-run preprogram situation to take account of Ashenfelter's dip. As outcome variable, we use employment levels and transition rates. Our results indicate that that modelling the transition rates is more appropriate and that on average training did not enhance the employment prospects of participants. The basic results are robust to the different matching approaches and to the use of conventional or bootstrap standard errors.

Cite article

Bergemann, A., Fitzenberger, B. & Speckesser, S. (2001): Evaluating the employment effects of public sector sponsored training in East Germany. Conditional difference-in-differences and Ashenfelter's dip. 37 p.