The column is actually dumber and more offensive than the headline - sure cops are more likely to kill black people, but have you considered that the black people are poor? - but of course no one argues that police brutality is "always about race."
There's a lot of bad empirical work which is basically "if you control for income (often badly) in this dumb regression a lot of the effect of race seems to disappear!" Very generally you can see the problem of doing this for something like "police killing black people during traffic stops" as the police don't have your tax returns, they just know you're black.
Reviewing every piece of criminology and policy empirical work is beyond the scope of this very fine blog, but I'll just submit that advanced math and stats doesn't feature prominently in their training. Economists do this work, too, but they don't know anything about criminology, and their models often beg the question to some degree ("beg the question" used correctly here, pedants)! It's easy to find one study that proves your fucking point, especially if you are unable to evaluate it. So, you know, always a problem.
Doesn't take a genius to understand that brutal cops will brutalize (and kill) people they think are less likely to "matter" - to the press, to people who can hire lawyers - and that isn't just minorities, of course, but...