Analyzing an Active Labor Market Program in Germany: A Regional Approach—An Attempt to Use Propensity Score Matching for the Estimation of Causal Effects on the Level of Counties and Independent Cities

Abstract

The Institute for Work, Skills and Training was assigned to evaluate a labor market program aimed at the integration of long-term unemployed individuals aged 50 or older. The integration should have been achieved not only by training and coaching of individuals, but also by building regional networks between labor market stakeholders within a region. To appraise the success of the action undertaken on the regional level in the sense of causal effects, an observational study was used. The experimental-and control-groups were built using propensity score matching. The matching was done using not individual-level data, but data on the regional level because of missing individual data and the aims of the program. The mean growth of the number of employees subject to social insurance contributions over time was chosen as the outcome of the program. The findings are that observational studies are suitable to estimate the causal effects of active labor market programs on the regional (macro) level if individual data are missing or if the aims of the program cannot be observed on the individual level.