This post has been corrected.
Republican presidential contender Donald Trump crossed what some might consider a line in his racially charged campaign this weekend. On Nov. 22, he tweeted what purport to be statistics of the proportions of black and white Americans murdered, respectively, by blacks, whites, and police, but which are flagrantly and demonstrably false.
The data supposedly comes from the ”Crime Statistics Bureau—San Francisco,” an entity which, as Buzzfeed pointed out, does not exist. Here are homicide data by race for 2014 from the FBI—the 2015 data are not out yet, since 2015 is not over. The clearest discrepancy between these figures and Trump’s is that the overwhelming majority of white American murder victims are killed by other white Americans.
Another way to look at these numbers is to take deaths as a share of the victims’ population (based on 2014 population estimates from the US Census bureau). That’s a measure of the risk to any person of dying at the hands of each ethnic group. Here again, whites are at far higher risk of being killed by other whites than by blacks. It does, however, underscore that a black person runs a far higher risk of being killed by another black person than for any other combination.
Now to the data regarding police. One thing that makes Trump’s table clearly bogus is that it lumps in killings by police with all the others, though not all police killings are homicides. Another is that, as has been well (and scandalously) documented, there are no reliable national statistics for police killings in the US. But the Guardian’s “The Counted” project has been tallying police killings in 2015 based on news reports and crowd sourced information.
Police have indeed killed nearly twice as many white as black Americans:
But look at those killings as a share of each ethnic group’s population, and it’s a very different story: black people are at three times as high a risk of dying at the hands of police as white people.
The fact that Trump feels free to publish such wildly inaccurate and racially inflammatory numbers—without, so far at least, any visible backlash from his own party—may not surprise anyone who has been following American politics for the last few years. But it feels like a line has been crossed nonetheless. Are there really no limits?
Correction: An earlier version of this post miscalculated the numbers in the fourth chart—death rates at the hands of police as a share of population. Black Americans are three times, not 11 times, as likely as white Americans to be killed by police.