Hot Spots Interventions at Scale: the Direct and Spillover Effects of Policing and City Services on Crime
88 Pages Posted: 12 Oct 2017 Last revised: 1 May 2018
Date Written: April 17, 2017
To reduce crime, cities focus state resources in high-crime “hot spot” streets. A U.S. literature argues that this approach, especially policing, has large direct effects and beneficial spillovers - a finding consistent with criminal rents being concentrated in a very small number of streets. We need large samples, however, to accurately assess local spillovers. An experiment in Bogotá, Colombia is an order of magnitude larger than previous opportunities to evaluate such policies. The city identified 1,919 hot spots and randomly assigned them to either 8 months of doubled policing, greater municipal services (cleanup and lighting), both interventions, or neither. At this scale, however, spillovers in dense networks can introduce bias and complicate variance estimation through “fuzzy clustering”. We show valid hypothesis testing requires randomization inference. We find that intensive policing or municipal services alone had small direct effects on security in hot spots, but together both interventions reduced crime substantially. Nonetheless, the evidence points to both interventions displacing crime to neighboring streets. On average, aggregate crime appears to have increased in Bogotá, and we can rule out decreases in aggregate crime of more than 1-2%. These results are consistent with a situation where illegal opportunities are not highly concentrated in a city, a large proportion of crime (especially property crime) has a sustained motive, and criminals shift activities to places where they have lower risks of detection.
Keywords: Crime, Violence, Police, Public Services, State Building, Spillovers, Field Experiment
JEL Classification: K42, O17, E26, J48, C93
Suggested Citation: Suggested Citation