Blair said he and his co-author Michael Weintraub, an associate professor of government at Universidad de los Andes in Bogotá, chose to focus on a Latin American city because using military forces to police high-crime civilian areas is relatively common in that region.

"If you travel to almost any country in Latin America, you will see soldiers performing tasks that, in the United States, we would typically associate with police officers,” Blair said. “Things like running patrols, interrogating suspected criminals, setting up roadblocks… it's very common, and it's very controversial."

To learn more about the impact of this ubiquitous yet understudied practice, Blair and Weintraub worked with the mayor's office, the Colombian armed forces and Innovations for Poverty Action Colombia to evaluate the impact of a military policing intervention in Cali. 

In 2018, Cali had one of highest rates of violent crime rates in Colombia, with 46.7 homicides per 100,000 residents — more than triple the rate in Bogotá. In an effort to reduce crime, the Colombian government developed a program called Plan Fortaleza, which involved recurring, intensive military patrols targeting high-crime areas in the city. 

According to Blair, both the city and the military were interested in evidence-based approaches to crime reduction and were supportive of the idea of an empirical analysis of the program’s impacts. Cali's security and justice secretariat reached out to Weintraub "to see if he would be interested in setting up an impact evaluation of this intervention that they planned on running.

"They were convinced that we would find this approach effective," Blair said. “They thought a rigorous evaluation would give them ammunition — so to speak — to argue that the way they do policing is effective."

Monitoring an intervention

Blair described the study as “probably the most ambitious data collection effort I’ve ever undertaken.” 

To start, the city created a kind of randomized trial, arranging the setup of military patrols at a diverse set of locations and times to more accurately assess their effectiveness. The researchers then worked with the city police to access data from the coroner’s office and other sources — allowing them to capture the true prevalence of violent crime and its effects in Cali before, during and after the interventions. (Blair said crime data can be inaccurate because not all victims report crimes to the police.)

Finally, the research team released two surveys, one during the intervention and another afterward, reaching about 10,000 residents in total. 

"We also had civilian monitors with the soldiers at all times, monitoring and helping them navigate to the particular blocks they were supposed to patrol,” Blair said. “And those monitors recorded their observations of what was going on. So we have really nice data on where the patrols actually went, how long they spent at each location and what the soldiers did while they were there.”

When the researchers compared the patrolled areas with a control group that had not been patrolled, they found a null result. In other words, the research team found that, contrary to what city leaders had believed, there was scant evidence that Plan Fortaleza had meaningfully reduced the prevalence of crime.

There was similarly little evidence that the intervention had introduced new human rights abuses in Cali, Blair said. The team examined reports from the attorney general’s office, in addition to  firsthand observations from the civilian monitors. In neither source did the researchers find any evidence of abuse or misconduct by members of the military.

Blair said the surveys did show evidence of a rise in human rights abuses by police officers. However, he said, it’s difficult to know whether that represented a genuine rise in police abuse or whether that was merely an artifact of more arrests in light of the increased military presence.

"It's possible that what the intervention did was create more opportunities for police officers to be abusive without making them more abusive than they already are,” he said.

Blair was reluctant to say the results of this study could be generalized to apply to other areas, though he noted parallels between Cali and other settings where military policing has been tried. He said he hopes the study inspires other scholars across the globe to lead their own rigorous studies on the effectiveness of military interventions in civilian areas, enriching the evidence base and helping leaders make decisions that balance their constituents’ safety and liberty.

"Currently, there is a slim body of work that addresses this question, which is why I'm really proud of this study," Blair said.