What’s the Rumpus?

A hungover MT had some thoughts on the racial profiling dialogue occurring due to NY subway "random" searches.  This reminded me of everyone’s fav economist’s treatment of racial profiling.

If the U.S. decided it wanted to eliminate crime altogether, theoretically we could.  We could end murder, rape, theft, what have you, at the expense of becoming a police state, losing our individual freedoms, sacrificing our economy, etc.  Most of us agree that this is "too high" a rate of police inquiry.  A false-positive inquiry costs society.  Racist racial profiling is when one group incurs this cost at a higher rate due to prejudices. 

The issue at hand is to what extent, if at all, race should be used in suspect profiling.  On the one hand, police-work entails learning from experience, following hunches, getting a feel for criminals, etc.  Levitt gave as an example a veteran highway patrolman in the Midwest who was known for his ability to spot drug traffickers transporting cocaine to Chicago, and was regularly able to single-handedly confiscate more drugs than any other department in the state.  When asked how he did this, the officer claimed he was pulling over drivers who fit the profile:  Latino male, mid-thirties, with a dirty car or truck, often with Florida plates.  On the other hand, many police officers have proven themselves to be racist thugs, incapable of not abusing power.

A policeman’s job is to apprehend criminals, no matter what their race.  From watching Cops, it seems some criminals, whatever their race, are a little easier to spot than others,  e.g. Ronnie Dobbs. 

You can assume that people of every race will be apprehended somewhat (I have yet to hear the hyper-cynical argument that racist white cops avoid apprehending whites they know to be criminals).  Then, in addition to the criminals apprehended in common, racist cops will differentiate themselves from non-racist cops by apprehending minorities at too high a rate.  Paul Sperry argues that politically-correct policies result in non-minorities being targeted at a rate that is relatively "too high" — NYC police should stop searching "Girlscouts and grannies." 

So what is "too high" a rate for either group?  Levitt approaches this head on.  It is unlikely for criminals to have the exact demographic distribution as the population as a whole (and even if that did occur, it isn’t a "necessary and sufficient" condition).  Cops need to target whomever they are going to target, but once they do so their actions can be proven to have been prejudicial by their success ratio; i.e. the relative rate of correctly targeting suspects.  Example:  cop A searches 100 cars, 50 black and 50 white.  Of those 100 searches, 25 result in narcotics arrests.  Of those 25 arrests, 20 are white and 5 are black.  Cop B also searches 100 cars, this time 80 black and 20 white, resulting in 25 narcotics arrests, 20 black and 5 white.  Who’s racist?  Cop A is the racist cop; he’s targeting blacks.  Even though he searches as many white and black cars, he found 4 times as many white criminals as black.  Cop B is a fair cop.  When he searched 4 times as many black cars as white, he found 4 times as many black criminals as white.  Cop A’s success rate with catching black criminals is only 20%, versus 80% success with whites.  Cop A needs to stop targeting blacks needlessly and start targeting whites until his ratios balance, like Cop B, whose success rate with both blacks and whites is 25%. 

By the way, this has no policy implications for NYC subway searches.  Police have yet to find a single bomb-laden person, much less enough terrorists to imply whether they are correctly profiling their suspects.  In my opinion, since they probably will never catch a bomb, the idea of searching randomly is as good as any.  At least it stands a chance to deter in general, whereas searching according to profiles will result in a substitution effect. 

I have a feeling this post will irk some of you (especially fellow Berkeley-oriented folk).  I wrote this over the course of a day, a sentence at a time, while doing other work.  I am not sure it reads well.  And you may wonder why my relatively progressive ass just said that Cop A who searches an equal number of black and white citizens is more racist than Cop B, who searches 4 times as many blacks than whites.  If this post needs clarification, please comment or email me. 

This entry was posted on Tuesday, August 2nd, 2005 at 3:13 pm and is filed under Current Affairs, Science. You can follow any responses to this entry through the RSS 2.0 feed. You can leave a response, or trackback from your own site.