In some parts of the US, police kill black people at a rate six times higher than they kill white people. The differences are most stark in the northern Midwest, especially Chicago, and in north-eastern states like New York.
Protest movements like Black Lives Matter have highlighted the disproportionate killing of black people by US police, and called for major changes in policing practices. However, official data on police killings can be unreliable. The database run by the Bureau of Justice Statistics is known to undercount deaths, partly because police forces don’t have to contribute data. That makes it harder to stop the killings.
Gabriel Schwartz and Jaquelyn Jahn at Harvard University compared police killings in different regions of the US between 2013 and 2017. They used data from Fatal Encounters, an independent organisation that gathers public and media reports of killings, and fact-checks them.
The researchers assigned each death to one of the US’s 382 “metropolitan statistical areas”. These are “cities and the areas surrounding cities”, says Jahn, and reflect where people spend most of their time.
Rates of police killings varied widely. For the overall population, the highest rates of killings were in south-western states like California and New Mexico, where more than 1 in 100,000 people were killed by police every year. In the north-east, rates were often lower than 0.3 people per 100,000.
However, the pattern changed when the team looked for differences linked to ethnicity. In south-western states, police killed black people 1.81-2.88 times more often than they killed white people. In the north Midwest and north-east, the disparity was often more than 2.98. In the Chicago metropolitan area, black people were killed 6.51 times more often than white people.
“They are showing for the first time that there’s a lot of variation by place in racial inequalities in police killings,” says Justin Feldman at New York University. That in turn should help us understand why some places have such large disparities, and how to reduce the deaths, he says.
Risk of death
Schwartz and Jahn’s study is the latest of a raft of studies showing that black people in the US are killed by police more often than white people. Young black men are at highest risk. A 2019 study found that black men aged 25-29 were being killed at rates between 2.8 and 4.1 in 100,000.
Neighbourhoods are also a factor. Death rates are highest in poor neighbourhoods and neighbourhoods with high non-white populations, but black people are at higher risk of being killed in white neighbourhoods.
There is evidence that the killings have wide-ranging effects beyond those killed and bereaved. A 2018 study found that the killings had a harmful impact on the mental health of the wider black population.
Some scholars and commentators do still claim that there is no racial inequity in police killings. In 2019, David Johnson at the University of Maryland and his colleagues published a study in PNAS claiming no evidence of anti-black disparities in police shootings.
Similarly, African-American economist Roland Fryer, also at Harvard University, has argued that there is no evidence for racial disparities in police shootings. These studies have received widespread media coverage.
The problem is that these studies focus solely on people who interact with the police, for instance by being stopped, says Feldman. The underlying argument is that if black people commit more crimes, a higher rate of police killings would follow.
“You can’t do that in a valid way,” says Feldman. “If there’s racial bias in why police stop people or investigate crimes in the first place, it’s going to obscure the racial bias in police shootings or police killings.” The 2019 study has received multiplecritiques from other scholars because it didn’t account for this problem.
There is evidence that police stop black people more often than the stop white people. For instance, under New York City’s stop-and-frisk policy, black and Hispanic people were stopped more than white people, even accounting for estimated differences in crime rates. Furthermore, a 2015 study found that rates of police killing don’t follow crime rates.“