Between January 1994 and January 1999, on the Maryland stretch of Interstate 95, a black motorist was three and a half times as likely as a white motorist to be stopped by the police and searched for drugs. You might be tempted to conclude that the police were out to get blacks. The truth is more sinister: Actually, the police were out to get drug dealers.
OK, I have two things to explain. First, how can we know what motivated the police? Second, why is animus toward drug dealers worse than animus toward blacks? Let me take these one at a time.
Regarding police motives, the key observation comes from a recent study by Professors John Knowles, Nicola Persico, and Petra Todd at the University of Pennsylvania: Among those motorists who were stopped and searched, about one-third were found to be carrying drugs—and this proportion was the same for blacks as it was for whites. You might be tempted to conclude that blacks and whites have equal propensities to carry drugs. But think again: Blacks carried drugs just as frequently as whites did even though they were three and a half times as likely to be caught.
If you believe that people respond to incentives, you must conclude that in the absence of racial profiling—if blacks and whites were treated equally—a greater proportion of blacks would carry drugs. In that sense, and at that time and in that place and for whatever reason, blacks had a higher propensity than whites had to carry drugs.
(Of course, this argument makes sense only if blacks are aware that they're being stopped more frequently than whites—but given the ubiquity of complaints about racial profiling, this seems like a natural assumption.)
So, do police stop blacks because they're black, or do they stop blacks because it's the best way to get a lot of drug convictions? The answer, say Knowles, Persico, and Todd, is that only the second hypothesis can explain why that "one-third" statistic is the same for both races.
Here's why: Think about what happens if the police single-mindedly maximize the number of drug convictions. They start by focusing their attention on the group with the higher propensity to carry drugs—in this case blacks. That makes it easier for white drug carriers to slip through the net, and harder for black drug carriers. Because people respond to incentives, the number of white drug carriers grows and the number of black drug carriers shrinks.
That process continues until whites and blacks are carrying drugs in equal proportions. At that point (which could have been reached long before the data started being collected in 1994), there's no reason to crack down on blacks any further, but no reason to ease up on them either. (And any temporary easing would quickly lead to a discrepancy between black and white conviction rates and a return to the equilibrium.)
If the police actually bore some animus toward blacks, you'd expect them to go further—cracking down on blacks to the point where very few blacks would dare to carry drugs. Then we'd see a lower conviction rate for blacks than for whites. Instead we see equal conviction rates, which suggests that the police concentrate on stopping blacks right up to the point where it helps them increase their conviction rates and no further.
By that standard, it's not blacks but Hispanics who have cause for complaint. Stopped Hispanics are only about a third as likely to be carrying drugs as stopped whites or stopped blacks. Why would the police stop a Hispanic who has a one-ninth chance of carrying drugs instead of a black or white who has a one-third chance of carrying drugs? Arguably, it's because they have something against Hispanics.