Evaluating the dynamic employment effects of training programs in East Germany using conditional difference‐in‐differences

Abstract

Die Discussion Papers dienen einer möglichst schnellen Verbreitung von neueren Forschungsarbeiten des ZEW. Die Beiträge liegen in alleiniger Verantwortung der Autoren und stellen nicht notwendigerweise die Meinung des ZEW dar.Discussion Papers are intended to make results of ZEW research promptly available to other economists in order to encourage discussion and suggestions for revisions. The authors are solely responsible for the contents which do not necessarily represent the opinion of the ZEW.Download this ZEW Discussion Paper from our ftp server:ftp://ftp.zew.de/pub/zew-docs/dp/dp0441.pdf Non-technical SummaryAfter German unification, Active Labor Market Policies (ALMP) were implemented at a large scale in East Germany in order to prevent unemployment. These programs mainly consisted of training and temporary employment schemes. In 2002, more than a decade later, the German Federal Employment Service (Bundesanstalt für Arbeit, BA) still spent around e 20 Billion (≈ 0.9% of the GDP) for ALMP. About 50% of this budget goes to East Germany even though the labor force in East Germany is less than one sixth of Germany as a whole. Quite a significant share of the labor force in East Germany has been participating in programs of ALMP since 1990. In light of persistently high unemployment, the effectiveness of ALMP in East Germany is often questioned.Contributing to this debate, we estimate the employment effect of public sector sponsored training programs in East Germany at the individual level for the time period 1990 to 1999. Training programs were intended to provide skills to the individuals that are in demand in a market economy but not in sufficient supply due to the former educational system. Consequently, regarding the number of participants, training was the largest ALMP program in East Germany. Our empirical analysis focusses on the group of individuals who belonged to the active labor force in 1990. This group was hit fully by the transformation shock. We use data from the Labor Market Monitor for the state of Sachsen-Anhalt (Arbeitsmarktmonitor Sachsen-Anhalt LMM-SA).This paper first provides an overview on training programs in East Germany and develops the methodology for estimating treatment effects of training programs based on a dynamic employment model. Then, the evaluation of the program involves estimating separately the effects on the outcome variables transitions between employment and nonemployment depending on the employment status in the previous month. When evaluating the average effect of labor market policy for the treated individuals, it is important to estimate the counterfactual average nontreatment outcome for the treated individuals based on similar nontreated individuals in order to avoid a possible bias induced by the selection into the programs. We implement nonparametric kernel matching based on the estimated propensity score for treatment. This way, we intend to control for selection bias based on observed characteristics. In a second step, we take before-after differences in o...

Get Training or Wait? Long-Run Employment Effects of Training Programs for the Unemployed in West Germany

“…Effectively, we run a nonparametric local linear kernel regression (Heckman, Ichimura, Smith, Todd (1998), Pagan and Ullah (1999) Bergemann et al (2004)) which comprises a weight function w N l (i, j) that gives the higher weight to individual j in the eligible subset of the l-group, {Tũ = l}, the stronger his similarity to participant i in treatment k regarding the estimated propensity score and the starting month of the unemployment spell. Based on a local linear regression of the outcomes in the eligible l-group on the estimated propensity score and the starting month of unemployment, the estimated ATT can be written as…”

Employment effects of the provision of specific professional skills and techniques in Germany

“…In a dynamic setting, one has to take account of the timing of events, see also van den Berg (2003, 2004). Our matching estimator is implemented using local linear matching (Heckman/Ichimura/Smith/Todd, 1998) with the crossvalidation procedure suggested in Bergemann et al (2004).…”

“…Effectively, we run a nonparametric local linear kernel regression (Heckman/Ichimura/Smith/Todd, 1998;Pagan/Ullah, 1999;Bergemann et al, 2004) which can be represented by a weight function w N 0 (i, j) that gives the higher weight to nonparticipant j the stronger his similarity to participant i regarding the estimated propensity score. The estimated TT can be written as…”

Education, unemployment and migration

“…When reinterpreting our distribution of education costs as distribution of success probabilities, we should expect that too many people participate in highly subsidized education and training programs. Thus, it is not astonishing that evaluation studies often …nd zero or small impacts of such programs on subsequent employment probabilities and earnings of the participants (see, e.g., Heckman et al, 1999;Bergemann et al, 2004;Lechner et al, 2005;Albrecht et al, 2005).…”