> Another possible explanation is that Blacks are more likely to be criminals, and these stops represent police getting bad guys off the street. [..] City-wide, 68% of no-arrest stops city-wide were of Blacks, the same proportion as arrest stops, which means that Blacks were no more likely to be criminals than others who were stopped.

If, despite being stopped more often, the stop is equally likely to result in an arrest (I'm assuming that's what "no more likely to be criminals than others who were stopped" means, since if they were less likely, the author would surely mention it), doesn't that support this explanation? Of course there are problems with this, since an arrest itself is up to the cop's possibly biased judgement. Which brings me back to

> There is more policing in those beats, they argue, because that’s where the crime’s happening. There’s a logical problem with this argument, because more crime will always be found where more policing is done. But leaving that aside

How quickly he moves on! But there is a solution: look at crimes not affected by over-policing, such as homicide, or look at victimization surveys [1], that don't involve the police. Finding the results is left as an exercise for the reader.

[1] https://en.wikipedia.org/wiki/National_Crime_Victimization_S...