Hot Spots Interventions at Scale: The Direct and Spillover Effects of Policing and City Services on Crime

Bogotá intensified two forms of state presence to make high-crime streets safer: policing and municipal services such as cleanup and lighting. We randomly assigned 1,919 high crime “hot spot” streets to 8 months of doubled policing, greater services, both, or neither. At this scale, statistical power exceeds prior studies by an order of magnitude. But spillovers in dense networks cause interference between units and “fuzzy clustering.” We show valid hypothesis testing requires randomization inference. After accounting for spillovers, intensive policing or municipal services alone do not lead to statistically significant increases in security on hot spots. Together, however, both forms of state presence have large and statistically significant impacts on security, and suggest increasing returns to state presence. But total crime directly deterred was modest and, using data from all streets, we cannot reject the hypothesis that directly deterred crimes were displaced to neighboring streets. Nonetheless, there are some indications of a fall in total violent crime. These results contrast with prior studies concluding that intensive policing has large direct effects and positive spillovers.