How do you root out bad cops without changing the behavior of good cops?
That’s a question explored in a forthcoming academic paper on policing the police by Harvard economist Roland Fryer and co-author Tanaya Devi. Given the current nationwide protests and mob violence ignited by the death of George Floyd in Minneapolis, it’s a subject on the minds of many.

In 2016 Mr. Fryer released a study of racial differences in police use of deadly force. To the surprise of the author, as well as many in the media and on the left who take racist law enforcement as a given, he found no evidence of bias in police shootings. His conclusions have been echoed by researchers at the University of Maryland and Michigan State University, who in a paper released last year wrote: “We didn’t find evidence for anti-Black or anti-Hispanic disparity in police use of force across all shootings, and, if anything, found anti-White disparities when controlling for race-specific crime.”

Mr. Fryer said in an interview that the new paper is an extension of his earlier research. Although it seemed clear to him that racial disparities in police shootings stemmed primarily from racial disparities in criminal behavior, police departments continued to be investigated, and he suspected these investigations weren’t having the intended effect. In fact, he noticed what he suspected was a pattern that warranted further study. After surveying more than two dozen federal and state probes of police departments across the country, the pattern became clear. When police were investigated following incidents of deadly force that had gone viral, police activity declined and violent crime spiked. It happened in Ferguson, Mo., after Michael Brown was shot by an officer. It happened in Chicago after a cop gunned down Laquan McDonald. And it occurred in Baltimore after Freddie Gray died in police custody.

Mr. Fryer stressed that it isn’t the investigations themselves that are the problem so much as the circumstances under which they are launched. Investigations that weren’t prompted by well-publicized events resulted in little change in police behavior and violent crime. “But when I look at cities in which the investigation was preceded by a viral event,” he said, “homicide goes up considerably. Total crime goes up considerably.” What happens, he said, is that police effectively pull back. They don’t stop doing their jobs, but they become less proactive and curb their interactions with civilians.

In Chicago, there was a 90% drop in police-civilian contacts immediately after the announcement of an investigation, and “Baltimore literally went to zero” after a probe was announced there, he said. In cities where these contacts fell the most, homicides increased the most. Sadly, the decision to launch departmentwide state and federal inquiries into the deaths of Brown, McDonald and Gray resulted in numerous additional deaths. Mr. Fryer said that because of changes in police behavior following investigations in these and other cities, “my estimates show that we lost a thousand more lives, most of them black as well, because of an increase in homicides.” The protesters and their political allies insist that policing is the problem, but when police pull back, black communities are hit hardest.

“This is not to say that police departments shouldn’t be investigated,” Mr. Fryer added. “But to quote [former Mayor] Rahm Emanuel in Chicago, ‘investigations have to be done with police, not to police.’ ” One alternative is to target individual officers for wrongdoing rather than putting entire departments under a cloud. Federal officials also could be more patient in letting local investigations run their course before Washington gets involved. Mr. Fryer hopes his results will “encourage introspection on the trade-offs involved when we increase scrutiny on police departments.” At the moment there’s a lot more venting and posturing than introspection.

Protesters have decided to vilify the police. Rioters have decided to take advantage of the protests. And the media have expressed little interest in putting this tragedy in context. The activists tell us that what happened to George Floyd is commonplace and racially motivated, but the empirical evidence points in the opposite direction. Camera phones and social media may give fatal encounters between cops and black suspects more attention, but anecdotes are no substitute for hard data.
And now we know how scapegoating law enforcement can backfire in ways that do the most harm to our most vulnerable communities. “I never would have guessed that if police stopped putting in the effort, that homicides would change like this,” said Mr. Fryer. “You hear some people say ‘Oh, we want to police our own neighborhoods, get out.’ No, you don’t want that. I guess I always knew it was a foolish idea, but I didn’t realize it was this deadly.”