Key takeaways: Coordinated randomized controlled trials across six contexts reveal that community policing interventions were inconsistently adopted, did not improve citizen-police relations, and failed to reduce crime.
Geographical Regions: Africa, Asia, and Latin American and the Caribbean
Type of study: Comparable experimental studies in six countries
Does community policing reduce crime and improve trust between the police and citizens? Through a coordinated randomized control trial in six countries across the Global South, the authors evaluate whether introducing community policing alleviates conditions of insecurity. Although the authors find that the interventions increased police compliance with community-policing initiatives, they detect no evidence that the reforms reduced crime, nor do they find evidence of increased trust in the police. Taken together, the results suggest that the promise of community policing may only be achieved in the context of broader structural reforms.
More than one quarter of the world’s population lives in conditions of insecurity due to high levels of crime and violence. The burden of reducing crime and insecurity typically falls to the police. But widespread public mistrust of the police hamstrings its effectiveness by undermining citizens’ willingness to collaborate in the coproduction of safer communities. Community policing has emerged as a proposed solution to this friction, a reform designed to improve dialogue by involving citizens directly in the police process. The core of these programs revolves around an increase in foot patrols, frequent town-halls, and problem-solving strategies to take action on concerns raised by citizens.
Evidence on the efficacy of community policing initiatives is mixed. Studies showing positive effects—documenting reductions of crime, in particular—come primarily from North America and Western Europe. The authors evaluate the efficacy of community policing in contexts across the Global South, settings in which public perceptions of police capacity and integrity are low, and in which crime and police abuse plague many communities.
To evaluate the efficacy of community policing initiatives, the authors conduct coordinated randomized controlled trials in six contexts: Brazil, Colombia, Liberia, Pakistan, the Philippines, and Uganda. In each site, the authors collaborated directly with the relevant local police agency, which implemented a locally appropriate community policing intervention. The interventions had a core set of elements across all six contexts, including the convening of town hall meetings and the introduction of foot patrols. In three sites, the intervention included the convening of meetings between neighborhood watch groups and the police. In another three sites, interventions introduced new mechanism for citizens to report crime and police abuse. Finally, in three sites, problem-oriented policing was introduced, a set of strategies for resolving underlying problems leading to insecurity.
At each site, the authors randomly assigned the introduction of these reforms to some police beats or neighborhoods (“treated”), while keeping police practices unchanged in other areas (“control”). The size of these study areas varied across contexts. Drawing on citizen and police surveys, as well as administrative data, the authors measure the effect of community policing by comparing four classes of outcomes across treated and control units: crime, citizen attitudes toward the police, citizen cooperation with the police, and police attitudes and behaviors.
The authors first establish that the intervention led to an uptick in community policing practices in treatment units. Many, often well-attended community meetings were held in every site. Though citizen awareness of community meetings increased substantially and the frequency of police patrols increased slightly, neither effect is distinguishable from no effect. The level of compliance varied across contexts: the authors find no increase in community policing practices in treated units in Brazil.
Turning to the main outcomes, the introduction of community policing generated none of the anticipated effects. In the main analysis, the authors detect no evidence that community policing affected crime victimization or police abuse as measured by both surveys and administrative data. Nor do the authors find any effects on citizens’ attitudes: community policing does not appear to affect citizens perceptions of insecurity, perceptions of the police, or willingness to cooperate with the police.
These overall null effects do not obscure regional variation: community policing did not lead to the expected changes in any of the six sites. However, the authors find some evidence that community policing moved some certain auxiliary measures of citizens’ attitudes toward the police. In Liberia and Pakistan, for instance, the authors find that community policing boosted perceived police intentions, while citizens’ perceptions of police capacity increased in Colombia. Why does community policing fail to deliver better outcomes? The authors are able to rule out several possibilities. First, community policing does not appear to displace crime to neighboring communities. Second, the null effects do not appear to be sensitive to the intensity of intervention. The authors detect no differences in the efficacy of community policing across contexts with short interventions (Pakistan, 6 months) and those with longer interventions (Philippines, 17 months).
Community policing in the Global South does not, by and large, deliver the benefits claimed by its advocates. It does not appear to reduce crime, and it does not in most cases lead to improvements in citizen trust in the police. Given the cost of community policing interventions, the findings presented here suggest that concerned international development practitioners and policymakers should direct their resources elsewhere.
The authors identify several possible reasons why community policing fails to alleviate conditions of insecurity across the Global South, all of which offer new levers on which concerned policymakers can pull to improve policing. In particular, three structural constraints may have prevented substantive change.
First, the failure of community policing may have stemmed from the insufficient buy-in from senior police leadership. The research teams generally partnered with local police forces to implement community policing; in many cases, these actors were hamstrung by inadequate commitment from higher-level police leadership. Ground-level leadership thus faced incentives to deprioritize community policing responsibilities. Generating buy-in at every level, from beat cops to station commanders to national police leaders may be needed for successful reform.
Second, officers were often unwilling or unable to address issues raised by citizens during community meetings. Although citizens were generally concerned with so-called lesser crimes—such as domestic abuse, harassment, and fraud—senior police officials encouraged officers to pursue investigations into high profile major crimes. Here, the authors underline the importance of aligning officers’ promotion prospects with the needs of the citizens they are tasked with serving. Third, the frequent transfer and rotation of police often interrupted the effective roll out of community policing. Officers who were transferred in after the program’s inceptions did not possess the requisite training and had a difficult time establishing a rapport with community members during short stints. A hallmark of modern bureaucracies, this suggests that the frequent transfer and rotation of police is a critical hurdle to be removed for the effective implementation of community policing and other police reforms.