In many Western countries, a sizeable group of people live on welfare benefits for a long time. Many of them suffer from mental health issues. This paper studies the labor market and mental health effects of an activation program targeting these long-term inactive people. We exploit the staggered implementation of the program in a difference-in-differences design. We find that the activation program hardly affects labor market outcomes. However, for those on mental health medication prior to the start of the program, the use of mental health medication substantially drops in the years following the start of the program. This effect is particularly pronounced for men. We also study spillover effects on the children of those targeted by the program, finding some suggestive evidence for improved learning and mental health outcomes.