Abstract:
We worked with the City of Bogotá to study the effects of intensifying state presence on violence and perceptions of the state on the most violent city streets. States commonly use a mix of intensified armed force and public services to pacify violent areas and build legitimacy, in situations ranging from inner city policing to counterinsurgency. This is a particularly common and celebrated approach in high crime areas of US cities. While backed by evidence from randomized trials, these trials have been largely US-based and also fairly small in scale. Besides questions of statistical power, this raises the question of whether these policing and state buillding strategies work at scale, and also whether they have positive or negative externalities for surrounding areas. To investigate, we worked with the Mayor’s office and police in Bogotá to identify the 2% (or 1,919) highest-crime streets, or "hot spots”, and assigned these streets to either: a two-thirds increase in police presence for 8 months; basic municipal services (e.g. cleanup and lighting); both interventions; or neither. Since violence could simply be displaced, we structured the design to be able to estimate spatial spillovers within several radii. We find weak evidence of a 5-10% reduction in crime and violence, with no evidence of displacement to neighboring streets. These impacts seem to be greatest when the treatment is most intense, when streets receive both the policing and municipal services interventions, especially in the first weeks of the intervention, and in the highest crime streets. Even then the results are somewhat fragile and have low statistical power. We also see no evidence of an improvement in attitudes towards the police or city government. The results suggest that intensive, highly targeted, shorter term rotating interventions are promising, but that the existing literature may overstate the benefits of intensive policing for reducing violence and increasing state legitimacy.