A large fraction of disadvantaged young individuals makes money from crime, and yet their labor market is usually studied ignoring the presence of the criminal sector. Consistent with the empirical evidence, I build a two-sector model of frictional labor search to study crime and employment outcomes jointly, and estimate it using a panel data set of young offenders. I use the model to illustrate sizable interactions between sectors, and show that, aiming at the same reduction in crime, policies targeting the legal labor sector generate larger welfare gains, and at a comparable cost, than traditional policies focusing on the criminal sector.