Are Illinois State Troopers Using Racial Profiling?

Rob at the 26th St. Bar Association is calling me out to post something about the statistics revealed in a report by the ACLU of Illinois about racial profiling by Illinois State Police. As Rob summarizes from the report:

In essence, blacks are “asked” if it’s ok to look around their vehicle three times more often than whites and Hispanics are “asked” twice as many times as whites if the ISP can just poke around. And you know what? It’s the whites with the drugs!

The results show that during consent searches, 8.85% of whites searched had contraband (pot, usually) on them, and only 5.06% of blacks and 1.12% of Hispanics have contraband.

I guess I should acknowledge that we have these numbers because the Illinois legislature passed a law requiring the police to gather the data. Whatever else I may say about them, they get points for promoting transparency.

That said, I don’t find the numbers to be terribly useful. There’s so much we don’t know, which means we can spout all sorts of theories without reaching any useful conclusions. Which is what I’m about to do. Don’t say you haven’t been warned.

If I wanted to be super-contrarian, for example, I could argue that the numbers show that minorities are better at hiding their dope. (Fans of The Wire may remember Herc complaining to Kima that the white dope dealers are so stupid that he feels no pride in catching them.)

We can do some more realistic theorizing if we look at the massive state report the ACLU used to get their numbers and pull some data from page 776. Here’s a table of the data reported by the Illinois State Police:

	Caucasian	African American	Hispanic	Total
Stops	364377	82954	35933	483264
Search Requests	591	421	196	1208
Found Something	46	19	2	67
% Searched	0.1427%	0.4521%	0.4981%	0.2222%
% Found	8.8462%	5.0667%	1.1173%	6.2384%
% Holding	0.0126%	0.0229%	0.0056%	0.0139%

Note the “% Holding” line, which shows the percentage of all people stopped who were found to have drugs. Blacks stopped for traffic offenses are clearly found to be holding drugs more often than whites. This could reflect their behavior, or it could be the result of uneven enforcement. In any case, blacks stopped in traffic are being arrested for drug possession at a higher rate than whites.

Free market economists have pointed out that racism is often punished by the market. Racist policies are often inefficient, and companies that indulge in them will pay a price. Although police activities are not a market situation, we can still see what may be a similar effect: The desire to search blacks and Hispanics seems to be hurting the department’s stats.

Since whites have the highest rate of getting caught with drugs, the optimal drug-finding strategy is to search only whites. This table projects what the Illinois State Police stats would look like in 2007 if they didn’t search the 421 blacks and 196 Hispanics and instead used their resources to search 617 more whites:

	Caucasian	African American	Hispanic	Total
Stops	364377	82954	35933	483264
Search Requests	1208	0	0	1208
Found	94	0	0	94
PCT Searched	0.2917%	0%	0%	0.2199%
PCT Found	8.8462%	0%	0%	8.8462%
PCT Holding	0.0258%	0%	0%	0.0258%

Without increasing manpower costs, they could increase drug arrests by 40%

To project this data, I assumed the rate at which drugs are found would remain a constant. There are two good reasons why that might not be the case, both of which are other theories to explain the data.

First of all, the the lower success rates for searching blacks could be a case of diminishing returns. Presumably, when Illinois State Police officers pick which 0.14% of white-operated cars to search, they pick the ones they believe most likely to have drugs. If told to triple the number of white-operated cars they search—thus matching the black search rate—they will presumably have to search cars that they previously guessed were less likely to have drugs. If their guesses were good, then having to search more cars will lower their success rate. Since they are already searching more black-operated cars, maybe this explains the lower success rate completely.

In this case, the optimal strategy is a generalization of the previous one: They should adjust their search rates to equalize success rates across races. That they clearly are not doing so suggests that state troopers are searching black-operated cars for reasons other than trying to make the most drug arrests.

Racism is certainly one possible explanation, although it’s hard to be sure of the effect. Perhaps state troopers are systematically overestimating the criminal tendencies of blacks, or they are trying harder to arrest blacks than whites, or they are using the searches themselves to harass blacks.

We can also propose some more benign theories, such as that whites are being caught with pot, but minorities are being caught with harder drugs, or that minorities are more likely to be members of inner-city drug gangs transporting cocaine and heroin as part of a criminal enterprise. Police might reasonably search more frequently to find harder drugs or to take violent gang members off the street.

Another important factor affecting the rate of successful drug searches is that drivers will presumably take the search rates into consideration when deciding whether or not to carry drugs. Since blacks are three times as likely to be searched as whites, it’s no surprise that fewer of them carry drugs.

The same theories should also apply to the differences between whites and Hispanics. The higher search rate for Hispanics should produce diminishing returns, and it should discourage Hispanics from carrying drugs, both of which should result in a lower rate of finding drugs when searching Hispanics.

This brings us to an interesting anomaly: Since blacks are searched at a higher rate than Hispanics, blacks should carry drugs less often than Hispanics. However, the statistics show that blacks are four times as likely to be caught with drugs as Hispanics.

This inversion suggests that blacks have a higher likelihood of drug possession than Hispanics, for any given level of enforcement. I’m not saying drug possession is some sort of racial trait, but it can’t be explained by the differences in police behavior captured in this data. Perhaps blacks are culturally more likely to carry drugs than Hispanics, or maybe state troopers tend to encounter blacks in high-crime areas where they are more motivated to conduct searches.

Chicago Police numbers (page 298) tend to support this second theory, because although blacks and Hispanics are 5 and 4.5 times more likely than whites to be searched, respectively, the rates at which drugs are found are much closer together—12.7% for whites, 10.7 for blacks, and 10.4% for Hispanics—which makes sense because it all happens in the same city. Then again, blacks form a significant part of the power structure in Chicago, so maybe that’s forcing CPD to handle racial issues better.

(Side note: The reports also include numbers for other ethnic groups. Interestingly, in 2007 neither the Illinois State Police nor the Chicago Police searched the cars of any American Indians. This suggests that American Indians in need of cash could hire themselves out as drug mules.)

The reason I have so many theories and so little proof for any of them is that there’s not enough data in this study. It suggests there’s something racial going on without pinpointing the problem. But maybe it’s a guide to where to look next…