Increasing state and police presence in the least governed areas of Bogota: a hotspot policing approach

Publication Details

3ie Funded Evaluation,DPW1.1044. A link to the completed study will appear here when available.

Chris Blattman, Donald Green, Daniel Mejia, Daniel Ortega, Santiago Tobon
Institutional affiliations
None specified
Grant-holding institution
None specified
Latin America and the Caribbean
None specified
Anti-corruption/ Governance
Gender analysis
Anti-corruption/ Governance
Gender analysis
Equity Focus
None specified
Evaluation design
Randomised Control Trials (RCT)
Ongoing 3ie Funded Studies
3ie Funding Window
Development Priorities Window 1


The study evaluates real-time impacts of various anti-crime interventions in Colombia.


In Bogotá, one of Latin America’s largest cities, 59 per cent of residents report feeling insecure. Newly elected Mayor Enrique Penalosa has made reducing crime and violence one of the central aims of his term, which runs from January 2016 to December 2019. However, there is little evidence on how the city government and police can reduce crime and build legitimacy cost-effectively. The results of the current study are expected to help the city to redeploy police and resources.

Research questions

The evaluation will answer the following questions:

  1. What is the relative effectiveness of coercive and non-coercive interventions to increasing order and state legitimacy?
  2. Are these interventions cost-effective?
  3. What are the positive and negative spillovers of these interventions into the rest of the city?


Intervention design

The study includes two interventions. The first is a hotspot policing intervention that consists of increasing the dosage of patrolling time from about 55 minutes per day per hotspot street segment to 90 minutes, divided in six entries of 15 minutes each. Police patrols will be given specific instructions on how to distribute entries during the day. The activities the police perform while patrolling will be standard (e.g. criminal record checks, door-to-door visits to the community, arrests, drug seizures). For hotspots in the control group, police will not receive any special instructions and will be free to patrol as they see fit.

The second intervention is informed by the broken windows theory which suggests that crime may reduce through maintenance interventions such as cleanup and garbage pickup. However this has not been tested rigorously on a large scale in any country. This intervention will consist of sending a municipal team to selected hotspots to clean up streets, to signal state presence and order. The municipal team will be charged with repairing street lights, cleaning graffiti, and collecting garbage every few weeks.

Theory of change

The first intervention, hotspot policing, increases police presence in hotspots through the reallocation of police presence from street segments with less crime to those with more. Potential criminals and the community will then become aware of the increased police presence in these streets. The increased time police patrols spend in these segments increases the likeliness of apprehension and punishment. Because potential criminals are more likely to get caught, the expected cost of engaging in criminal activities in these areas rises. For some individuals, these higher costs will now outweigh the benefits of committing the crime so they will no longer commit the crime, leading to a decrease in crime. The second intervention aims to reduce street disorder and create an environment of lawfulness. Potential criminals will become aware of the improved physical environment and believe that police presence and other enforcement efforts are stronger at this location. Therefore, the subjective perception of apprehension and punishment will rise, increasing the cost of engaging in criminal activity and thus decreasing crime.

Evaluation design

The study uses a randomised controlled trial design where street segments are randomly assigned to receive the intervention or not. To identify the experimental sample for the intervention, past crime data and verifications with police stations were used to identify 1,919 highest crime street segments. In addition to the experimental evaluation sample, data were also collected on a random sample of 550 no-hotspot streets. A street lighting and citizens survey will be conducted as well. As part of a bigger project, the current evaluation is aimed at estimating both the direct and indirect (spillover) effects of a largescale hotspot policing intervention and a largescale municipal services intervention.

Additional publications

Media coverage

Scroll to Top