Understanding whether criminal behavior is “contagious” is important for law enforcement and for policies that affect how people are sorted across social settings. We test the hypothesis that criminal behavior is contagious by using data from the Moving to Opportunity (MTO) randomized housing-mobility experiment to examine the extent to which lower local-area crime rates decrease arrest rates among individuals. Our analysis exploits the fact that the effect of treatment group assignment yields different types of neighborhood changes across the five MTO demonstration sites. We use treatment-site interactions to instrument for measures of neighborhood crime rates, poverty and racial segregation in our analysis of individual arrest outcomes. We are unable to detect evidence in support of the contagion hypothesis. Neighborhood racial segregation appears to be the most important explanation for across-neighborhood variation in arrests for violent crimes in our sample, perhaps because drug market activity is more common in high-minority neighborhoods.