This paper assesses the importance of reverse causality when evaluating the impact of training duration for unemployed workers. We use planned duration as an instrumental variable for actual duration. Our results suggest that the potential endogeneity of exits seems to be only relevant in the...