A police raid in the Jacarezinho favela of Rio de Janeiro killed 27 Brazilians, prompting a demand from Brazil’s Supreme Court and U.N. human rights officials for an investigation into the use of police force. Jacarezinho is a stronghold of one of Brazil’s most powerful drug trafficking gangs, the Red Command (Comando Vermelho, or CV). It is also home to thousands of lower-income Black and brown residents not affiliated with the gang.

What happened? The May 6 police raid — supposedly driven by intelligence information — was intended to target drug traffickers but quickly turned into an assault on the community. Police claimed to kill criminals, but 20 of the 27 killed were not suspects. Police fired at residents seated in metro cars and, according to eyewitness reports, executed alleged suspects. At least one killing during the six-hour assault took place in front of a child.

“They’re all criminals,” Hamilton Mourão, right-wing vice president and former army general, said in a news conference after the massacre. “[Criminal territorial control is] a serious problem in the city of Rio de Janeiro that we will have to resolve someday or the other.”

What happens when governments try to contain drug trafficking by using excessive force?

It often backfires. Leading research across the Americas shows that unconditional tough-on-crime policing does little to contain drug trafficking organizations. In fact, violent crackdowns often exacerbate rather than reduce violence. My research studying the effect of a temporary ban on police raids in Rio de Janeiro found that when compliance with the ban was high, the number of civilians killed by police decreased by 66 percent, while overall homicides in the community decreased by 19 percent. The ban on police raids led to broad reductions in crime — and saved more than 100 lives in just 30 days.

What happened during this ban?

On June 5, 2020, the Supreme Court of Brazil temporarily banned police raids “for humanitarian reasons” during the coronavirus pandemic. The ban arose amid public outcry after police shot and killed João Pedro, a 14-year-old boy playing with friends in his home. For the first few months of the ban, police followed the rules and ceased the raids.

In my study, I examined the consequences of the ban for both police and civilian violence. First, I looked at the effect of the ban on police killings and shootings. I combined daily precinct-level data from Rio de Janeiro’s Public Safety Institute and Crossfire (Fogo Cruzado), a data journalism nongovernmental organization that reports on shootouts. I compared violence levels in the 30 days before and after Pedro’s shooting, which immediately preceded the ban.

My estimates show that police killings decreased by at least 66 percent after the ban, totaling 111 fewer lives lost at the hands of the police during this 30-day period. And shootings involving the police in general decreased by 50 percent as well.

Second, I measured whether the ban affected violence that does not involve the police. I again used Public Safety Institute data to measure daily precinct-level homicides, extortion and property crimes, and used the Crossfire data to measure shootings among civilians.

I found that homicides dropped by 19 percent in the three months after the ban, with at least 39 fewer lives lost over 30 days. I saw no discernible impact of the ban on civilian shootings and other crimes (robbery, theft and extortion). This decrease in homicides was larger in communities that were more likely to be targeted by the police.

Why might a ban on police raids reduce violence overall?

These findings confirm popular assumptions that a vast majority of police use of lethal force occurs during raids. Banning police raids reduces opportunities for police to use violence; regular patrols and responsive policing are not where most police violence happens.

The absence of police raids — at least in the short term — didn’t lead to violent rampages committed by drug trafficking organizations. Although Rio de Janeiro’s former governor Wilson Witzel said in 2019 that “if we don’t kill those armed with a rifle, they will kill innocent people,” my study found the opposite.

There are several possible reasons that banning raids could reduce criminal violence. All start from the same place: Police raids can quickly and unexpectedly alter the balance of power between criminal groups and create uncertainty. When law enforcement disrupts criminal activity, it can lead to a contagion of violence, igniting turf wars, invasions and bloody power struggles. Likewise, in the absence of disruptive law enforcement activity, turf wars may decelerate, with the government playing a less aggressive role.

Experts have already begun making predictions about future criminal conflict in Jacarezinho. Some predict that the raids were political, deliberately weakening the Red Command drug trafficking gang so that protection racket-style militias — potentially including groups that support Brazil’s president — find inroads in the community.

Is banning police raids the solution?

My results strongly suggest that the police operation ban reduced unnecessary bloodshed in the short to medium term. Additional statistical tests I conducted rule out other possible explanations driving the decrease in violence after the ban: covid-19, the public outrage around João Pedro’s death and the seasonality of violence. These factors alone could not explain the decrease in police killings, homicides and shootings after the ban.

To be sure, a ban on police raids — unaccompanied by other public policies — isn’t enough to be a long-term solution to public security. But as a tearful lawyer and community leader pleaded to the camera in the aftermath of the massacre, “Did this raid end drug trafficking in Jacarezinho?” The answer is no. My findings add to the mounting global evidence that law enforcement strategies that “fight violence with violence” can be unnecessarily violent and not keep civilians safer.

Jessie Bullock (@jessiebullock) is a PhD candidate in government at Harvard University researching organized crime and clientelism in Latin America.