What makes a good caseworker?

How do caseworkers affect job finding and what characterizes a productive caseworker? To answer these questions we exploit variation coming from the fact that many local employment offices in Sweden assign job seekers to caseworkers based on their date of birth. We couple this identification strategy with fine-grained administrative data on both caseworkers and job seekers. Estimation of caseworker fixed effects reveals sizable variation in overall caseworker value-added. Female caseworkers perform better than male caseworkers and caseworkers with two years of experience outperform caseworkers with less experience. Cognitive ability and personal experience of unemployment are not related to caseworker performance. Based on the actions taken by the caseworkers we show that caseworker strategies are important. Analyses of caseworker–job seeker matching show that matching based on previous labor market experiences or gender leads to better outcomes.