It depends on the measure
Mapping Police Violence (MPV) says:
Unarmed black people were killed at 5x the rate of unarmed whites in 2015
This is a real statistic, but it is important to understand what it says. For every million white people and every million black people, a certain number of unarmed people of that race are killed by police. That number for black people is about five times as high as the equivalent number for white people. All that said, even just looking at unarmed people, their data shows more whites (259) than blacks (244) were killed.
That source did their own research, but it is consistent with other sources. For example, the Washington Post found:
In a Post analysis looking at population-adjusted rates, unarmed black men were seven times as likely as unarmed whites to die from police gunfire.
That said, there are some weaknesses in the MPV analysis if someone wants to claim it proves racism.
While looking through the data, I found one case where a former police officer is accused of murdering his ex-girlfriend. While she was black and unarmed, it seems something of a stretch to call this a police killing. He certainly wasn't on-duty or acting in any kind of official capacity at the time. A double murder and suicide, a backyard brawl, and other off-duty examples also pollute the data.
In one case, the decedent had been drowning a police officer when killed. While technically unarmed, his use of water as a weapon would likely have resulted in the officer's death if he had not been stopped.
The data doesn't say whether the killer was white or black. We usually discuss these statistics when there is a case involving a white police office and a black decedent. The statistics don't tell us how common or rare that is.
The MPV source includes:
- an innocent bystander or hostage killed
- a pedestrian or motorist accidentally hit by a police car or passengers in a vehicle chased by police with no weapon on them
- drivers or passengers accidentally hit by a police car
To claim racism here, you have to argue not that police are deliberately targeting black people to be killed but that they are more negligent when potential victims are black than white. That's certainly possible, but it is rarely what people are discussing.
Are there other explanations? Certainly. By the same kind of metric, crime is more likely to occur in black neighborhoods than white neighborhoods. As such, situations where an innocent bystander is killed are more likely to happen in black neighborhoods (where the criminals are).
Some measures that support explanations other than racism:
More white people than black people are killed by police. In the Washington Post data this was 494 to 258 (a ratio of just over 1.9:1).
From the Department of Justice as published by National Review:
- 22.4% of violent crimes are committed by blacks.
- 42.9% of violent crimes are committed by whites.
Going back to the Washington Post data:
- 26% (258 of 990) police killings involved black decedents.
- 50% (494 of 990) police killings involved white decedents.
Commentary notes that this gives a similar rate for blacks and whites when compared with their relative crime rates.
Of course, not all decedents in police killings are involved in violent crimes. The real use of that number is as a proxy for police contacts by race, which is not generally tracked.
An alternate source of data where analysis showed similar conclusions. That only studied urban areas where more people are black. And it only looked at homicides.
Black neighborhoods are poorer, may spend less money on police training and salaries, and may use police in roles where wealthier neighborhoods would call in specialists.
ProPublica has data on the race of the police officers. 78% of those killed by black officers are black. Contrast that with 46% of those killed by white officers.
More whites than blacks are killed by police, but there are more whites than blacks overall. Taken by population, blacks are more likely to be killed by police than whites, but blacks are more likely to be involved in crimes. Normalizing by ratio of police killings to ratio of violent crimes by race, blacks and whites are about equally likely to be the decedent in a police killing.