Since many years the analysis of labour market policies in Germany is of high interest for researchers and politicians alike. Results are, however, in several cases contradictory or ignore important aspects. This research project applies recently developed flexible estimators that take into account data quality problems. In our applications we use German administrative individual data of the Institute for Employment Research (IAB), Nuremberg and the German federal pension insurance, Würzburg. Data problems in these data often have a different nature than errors in survey data. In cooperation with projects [A1], [A3], [A4] and [C1], we develop and implement appropriate statistical models for unemployment duration and labor market transitions which also take into account the specific problems of the IAB data and the data from the German pension system. This project focuses on the application and the implementation of methods. Our analysis will therefore provide new insights to the understanding of micro- and macroeconomic labor market processes in Germany. We will derive empirically founded recommendations for labor market policy in Germany.Link to the homepage of the research group: www.stochastik.math.uni-goettingen.de/forschergruppe/index.php