Some Thoughts on the Fryer Paper
April 01, 2018
Michael Brown. Eric Garner. Tamir Rice. The unjust deaths of these unarmed black men and child at the hands of police officers sparked a national movement proclaiming that Black Lives Matter. For Roland Fryer, an economist at Harvard University, they prompted him to wield his expertise to examine racial differences in police uses of force. In a forthcoming paper, he finds that blacks are more likely than whites to endure non-lethal levels of force—such as slapping, tasering, or pushing to the ground—in interactions with police in New York City. However, Fryer finds no evidence of racial bias in police shootings—a lethal use of force—in his data from ten police departments across the country. He has remarked that this lack of evidence for discrimination is the most surprising result of his career.
News outlets were quick to jump on this paper. The New York Times published an infographic and article, with the title “Surprising New Evidence Shows Bias in Police Use of Force but Not in Shootings”.
Like the data that came before, Fryer’s data actually do indicate that blacks are more likely to endure police shootings. For example, the likelihood of black men ages 15-34 to be killed by police is about 10% according to recent data gathered by the Guardian and VICE. According to Fryer’s data, the likelihood is even higher at 15%.
This seems to show that Fryer finds racial bias in police shootings, since blacks are more likely to be shot. His investigation, however, does not end there. Fryer relies on econometric methods that tease out the different factors potentially leading an officer to shoot. Accounting for variables such as the alleged crime, demographics of the encounter location, or the civilian’s behavior during the encounter can help isolate the actual impact of racial bias itself, as opposed to other factors.
To do this, Fryer analyzes a massive amount of police reports. He uses data from ten U.S. cities including Los Angeles and Houston. Armed with a team of research assistants, he creates new datasets with an unprecedented amount of detail—coded from the reports—so his analysis can include as many of those other variables as possible.
Despite his thorough methods, Fryer’s finding that there is no evidence of racial bias in police shootings still cannot be trusted. There is possibility for bias, omission, or obfuscation in police reports. There is also the issue of the limited sample of U.S. cities. Houston is home to a progressive, racially diverse police department that has been publicly recognized for their efforts to combat discrimination. Other police precincts that volunteered their data may also feel like they have little to hide. Cities like Chicago, which has been investigated by the Justice Department for widespread racial bias, may yield different results. Given that this data come from a specific set of cities and precincts, any analysis cannot claim to explain American policing as a whole.
The data also do not capture the probability of a police interaction in the first place. Most people would agree that blacks are more likely to be stopped by police compared with whites, but Fryer’s experiment only looks at police use of force on individuals who have already been stopped. Based on a 2013 analysis by the New York City Public Advocate’s office, the likelihood that police find a weapon after stopping a white New Yorker was twice that of blacks. Let’s assume that police do not discriminate racially in shootings after stopping a civilian, but that blacks are stopped twice as often as whites because of racial bias. That implies that out of the general population, blacks are still shot at twice as often as whites as a result of racial bias. In not capturing police interaction data, an important piece of the puzzle is still missing.
Ultimately, Fryer’s finding on police shootings is not the noteworthy part of this paper. Out of police uses of force, shootings comprise only a tiny fraction. The rest are non-lethal uses of force, which happen on a much greater scale and affect many more people. In this category, Fryer finds clear evidence for racial bias. Using the massive amount of NYC Stop and Frisk data—5 million recorded encounters and over 100 variables to control for—and the same rigorous techniques, he finds that blacks are 21.2% more likely to endure non-lethal levels of force than whites.
To reconcile his findings, Fryer argues that police use lower levels of force with more abandon and racial discrimination, but are cautious about taking potentially lethal action across all races. Doing so would incur a greater cost, such as possibility of punishment or media attention. Fryer notes that there is still a possibility that police use excessive force on black civilians. To address this, he argues that police consequences of using unnecessary levels of force should be increased, even if they are not shootings. As Fryer states, “Black Dignity Matters.”