Racism Without Racists (LAPD version)

October 31, 2008
Posted by Jay Livingston

Can you have racially discriminatory outcomes without racist motives or intent?

Los Angeles police are much more likely to stop blacks and Latinos than they are to stop whites. And when they stop someone, they are more likely to frisk or search minorities than whites. Here’s a graph from the ACLU report that collected the data. The principle author is Ian Ayres. (The full report and data set are here.)

William Bratton, chief of the LAPD says flatly, “This department does not engage in racial profiling, has not. We have significant safeguards built in to protect against that.”

I believe him. But then how do you explain the data?

One commenter at the Freakonomics blog, where Ayres aired his findings, suggests that the crucial variable is not the racism of the police but the demeanor of the suspect. Maybe minorities, especially young males, act in a way that sets off the warning bells. That’s also what the police union president seems to mean when he says that the ACLU study is “an exercise that might work on a spreadsheet at Yale, but doesn’t work on the streets of Los Angeles.”

Ah yes, the streets.  The standard cop argument is that number-crunchers don’t know what’s really going down on the street. Cops know. Cops have that sixth sense, born out of years of street experience. It tells them whether someone is “clean” or “dirty.” Maybe they can’t put it into words, maybe they can’t lay it out so that lawyers in expensive three-piece suits and judges in black robes will recognize it as probable cause. But the cops know. They know who to stop, and they know who to search.

At least, that is the conventional wisdom . . . in the precinct and to a great extent in the media. (Do we ever see a movie where a cop’s strong but intuitive suspicions are wrong?) If cops are stopping and searching more minorities, it must be because minorities are more likely to be carrying illegal drugs and weapons. And the cops can tell.

Or can they? The data also show that the police searched a lot more innocent minorities than innocent whites. Cops searching blacks were about 40% less likely to find weapons than when searching whites.

This discrepancy certainly suggests that cops, wittingly or not, are discriminating against minorities. Ayres himself seems to favor that explanation.
The department should require that all existing and new officers take the Implicit Association Test (IAT) . . It produces a measure of unconscious bias . . . . For example, the black/white IAT produces a measure of whether an individual has unconscious negative associations with photographs of African-Americans relative to photographs of whites.
But I have different explanation. Mine is also race-based, but it doesn’t assume that police are racists or that they are, consciously or unconsciously, biased against blacks and Latinos. It’s just that the cops’ street sense, their ability to read people, doesn’t work so well across racial lines. That shouldn’t come as a surprise. We know that eyewitnesses are far more reliable in identifying people of their own race than people of another race. And just as we have trouble reading faces across race lines, we may also have trouble reading behavior.

If I’m right, then same-race searches should have a higher “hit rate.” And they do, regardless of the race of the suspect.
the racial disparities in the likelihood of arrest were substantially lower when at least one of the stopping officers was the same race as the suspect.
I picture a scene where a pair of cops, one black, one white, stop a young black suspect. They question him briefly. The white cop wants to throw the kid up against the car and search him, but the black cop restrains him. If I’m writing the dialogue, I don’t have the black cop warn about racism (“Watch it, Harry. We don’t want any Rodney Kings here,”). I have him say calmly but assuredly, “Take it easy, Harry. This kid’s clean.”


mike3550 said...

Jay, This is really fascinating. I also wonder how much of it might be due to another form of unconscious racism in that minorities tend to live in communities with higher crime (and therefore police are more cautious and/or willing to be more aggressive). Since racial residential patterns are so segregated, this would also drive up the stop & frisk rates of minorities for racial, but not racist, reasons.

Jay Livingston said...

Mike, you should download the report if not the data. Ayres included neighborhood variables. "the regressions in Table 4 are sufficient to show that the racial disparities of Table 1 are not merely a by-product of African Americans and Hispanics living in high-crime neighborhoods. Even after controlling for the crime rates, we find large and statistically significant disparities in the stop rate."

"The middle column of Table 23 adds in controls related to the attributes of the specific stop—such as whether the stop took place at night, whether the stop took place on the weekend, the division where the stop occurred, the role (driver, passenger or pedestrian) of the suspect and the assignment (traffic, patrol or other) of the officer. Adding these stop attribute controls slightly lowers the racial disparity estimates. The black frisk disparity is reduced to 39.9% and the Hispanic frisk disparity falls to 28.2%."

newsocprof said...

Ayres wrote an op-ed in the LA Times last week on this -- it was good. Then some yahoo from the police officer's union wrote in the next day or so saying that Ayres should have included information on whether the stops yielded an arrest or found drugs or whatever... I haven't read the report (it's not really shocking to me that the LAPD profiles, I'd be more likely to read a report that showed they don't), but it looks like Ayres did account for this.

Nice to see that the union representatives don't take the time to read the report before writing a public letter to the LA Times to condemn it!

mike3550 said...

jay, I will download and look at the report because it sounds really interesting. I'm glad that they looked at the neighborhood characteristics and it's interesting to know that it didn't do that much to reduce disparities.

Thanks again for posting this and for letting me know that there was even more interesting material in the report than what you already described.

PCM said...

This assumes the data has same reliability in white and black neighborhoods, which I doubt.

Also, the trigger to frisk is higher in less violent neighborhoods, because you're less likely to assume people are armed, which would increase the hit rate.

Just food for thought.