On June 16, the Department of Justice released the findings of its multiyear investigation into the Minneapolis Police Department (MPD). In a press conference, Attorney General Merrick Garland said that the DOJ has “reasonable cause to believe” that the MPD “engages in a pattern or practice of unlawfully discriminating against Black and Native American people in enforcement activities.” The finding made front-page news in the national press, which declared that the “scathing” report exposed “racist” and “discriminatory” policing. We read the report with keen interest, knowing how challenging it is to demonstrate discrimination, especially with non-experimental data. The report contains glaring oversights that overturn its conclusions.
First consider that, in 2018, MPD arrested nearly 9,000 men but fewer than 3,000 women. Of course, based on this statistic, hardly anyone would conclude the police discriminate against men. A commonsense perspective backed by criminological knowledge suggests that sex differences in criminal behavior account for the disparity.
The careful analyst should apply the same logic when evaluating evidence of racial disparities in law enforcement. Though they constitute 18 percent of the city’s population, blacks were responsible for 88 percent of Minneapolis homicides in which the race of the offender was known to the police. The rate of any violent crime among black Minneapolis residents is 27 times higher than among white residents. Given these massive differences in criminal behavior, the disparities in policing are, if anything, smaller than one might expect: the DOJ found that MPD stops black people 6.5 times more frequently than white people, and that “MPD used force against black people at nine times the rate that it used force against white people, given their shares of the population.” (The last part of the quoted sentence indicates the DOJ considers population size a suitable denominator when assessing disparities in enforcement; yet based on this logic, the MPD discriminates against men.)
What does the DOJ report say about racial differences in criminal offending in Minneapolis? Absolutely nothing. The DOJ Civil Rights Division spent years investigating the MPD but did not bother to examine the nature of the crime problem in the city. The report provides all kinds of auxiliary information about Minneapolis: the reader is told that “the median Black family in the Twin Cities earns just 44% as much as the median white family” and given a detailed account of the history of housing discrimination in the city, which dates back to 1910. Yet the report never recognizes the overrepresentation of African Americans in serious violent crime, both as victims and offenders.
Parts of the DOJ report do consider what the authors describe as “race-neutral” explanations for the observed disparities. In social-science terms, they used a multivariate analysis to examine some of the claims about racial discrimination. The purpose of this methodological approach is to see what happens to the initial disparity once additional factors related to the outcome of interest are taken into account.
Suppose the police are more likely to use force in stops involving black people. One possible race-neutral reason for this disparity could be that, in their encounters with the police, black civilians are more likely to possess a weapon, try to flee, or threaten the officer. These situational characteristics would justify the use of force; to show discrimination, a credible analysis must refute such explanations. The goal of multivariate analysis is to include a large number of salient factors” and see how much disparity remains once other factors are considered. In short, this method attempts to make all else but race equal.
In the multivariate analysis of the police use of force, the DOJ investigators adjusted for racial differences in a large number of behavioral factors, such as aggression towards the officer, disorderly behavior, fleeing police, and possessing a weapon. What happened to the initial nine-fold disparity once these race-neutral factors were taken into account? This is what the report stated: “After controlling for these factors, we found MPD discriminates when using force during stops. For example, . . . MPD used force during stops involving black individuals 24% more often (at roughly 1.24 times the rate) than it did during stops involving white people at similar times who reportedly engaged in similar behavior.” Based on this result, the report concludes that “the racial disparities we found are unlikely to be based on race-neutral factors. MPD unlawfully discriminates against black . . . people it stops by using force disproportionately.”
This interpretation of the analysis has two problems. First, as depicted in Figure 1—and contrary to what the report suggests—nearly all of the black–white disparity in the police use of force owes to race-neutral factors. Compare the size of the blue bar (“baseline disparity”) to the size of the orange bar (“adjusted disparity”). The blue bar depicts the amount of disparity in the police use of force once the race-neutral factors were included. Specifically, the results show that as much as 86 percent of the black-white disparity in the police use of force is due to behavioral differences. The second set of bars in Figure 1 indicates that 93 percent of the black-white disparity in vehicle searches owes to race-neutral factors.
Figure 1. Multivariate results of racial disparities in policing
Second, it is incorrect to suggest—as the report does—that the small amount of disparity that remains unexplained by the race-neutral factors implies discrimination. The multivariate adjustments are limited to the data available in the MPD’s administrative records. They do not account for “unobserved heterogeneity,” social-science jargon for differences that data couldn’t capture—a common limitation with this line of research. In the presence of unobserved heterogeneity, it is prudent to recognize that much of the residual disparity could disappear with more complete and accurate information about all relevant race-neutral factors. (Multivariate adjustment can help get closer to the truth, but it is no substitute for an experimental design under which all else really is equal.)
Why the oversights? Why did the DOJ choose to ignore racial disparities in criminal offending and victimization? Why did the study claim that race-neutral factors did not explain racial disparities in enforcement, when those factors account for almost all of the observed disparities? And why did the report insist that the remaining disparities imply discrimination? Beyond statistical questions, why did the report’s discussion of racial inequalities focus on black and white residents, ignoring other racial minorities? And why did the report never mention that the majority of the city’s black residents always opposed defunding the police, as the city promised to do following George Floyd’s death?
The investigation was the responsibility of the DOJ’s Civil Rights Division, the two leading members of which have supported calls to defund the police. In her previous role as the chief executive of the Leadership Conference on Civil and Human Rights, associate attorney general Vanita Gupta testified before the Senate Judicial Committee that it was “critical for state and local leaders to heed calls from Black Lives Matter and Movement for Black Lives activists to decrease police budgets and the scope, role, and responsibility of police in our lives.” Echoing this point of view, the current head of the DOJ’s Civil Rights Division, Kristen Clarke, penned a 2020 opinion piece in Newsweek that advocated “defunding policing operations that have made African Americans more vulnerable to police violence.” She noted that the federal government could refuse to fund police departments with a “history of violence or racial disparities in enforcement.”
Given the views of those in charge of the Civil Rights Division, the glaring omissions and misinterpretations plaguing the MPD report do not appear accidental. A critical appraisal of the report suggests that the real purpose of this long and expensive investigation was to establish a narrative that the MPD is a racist organization. The report did not seek the truth. It was a political stunt disguised as an impartial investigation.
The stage is now set for the Minneapolis Police Department to be placed under a federal consent decree. These measures impose highly bureaucratic regimens on police departments, forcing them to spend more time on paperwork than on policing. The best available evidence on the impact of DOJ investigations suggests they are unlikely to confer any benefits and very likely to undermine public safety. Research by Tanaya Devi and Manhattan Institute senior fellow Roland Fryer found that investigations carried out in reaction to “viral” incidents, such as the killing of George Floyd, have produced large increases in serious crime. The authors estimate that such investigations “caused almost 900 excess homicides and almost 34,000 excess felonies.”
All Americans must be protected from police abuse and other forms of unconstitutional policing. This is not controversial. Citizens should also be safe, however, from predatory criminal victimization, regardless of where they live, work, or go to school. The DOJ seems convinced that relatively infrequent incidents of police violence pose a more substantive threat to African Americans than the endemic problem of gun violence in disadvantaged neighborhoods. Facts suggest otherwise. In 2021, the number of unarmed black people killed by the police was 11—a fraction of 1 percent of the total homicide victims within the African American population.
Both the MPD and DOJ are entrusted to uphold the rule of law and to protect public safety without prejudice. By maligning the MPD with unsubstantiated allegations of racial discrimination, the DOJ has fueled mistrust of police among Minneapolis residents—making them more likely to take the law into their own hands. And by providing an ideologically skewed account of policing in Minneapolis, the DOJ has undermined its own credibility.
Photo by Glen Stubbe/Star Tribune via Getty Images