The Economics of Racial Profiling

Several years ago I built a simple model of racial profiling to see where it might lead. My original question was basically, if you were a shopkeeper and implemented a racial profiling security strategy, would it work? Specifically, if you alienated your black customers, who were profiled because of black criminals, could you remain profitable. The answer was yes. I've updated the model just a touch since then, but the basics are all the same. Here's the spreadsheet:

In month zero, you observe the first scenario in your store. You are setting up an experiment to determine whether or not to set up a racial profiling system for your new security guard. Your cameras roll and you don't bother any of your patrons. If they steal, they steal. You let them all walk. You have 1000 whites patronize your store. They spend an average of 12 dollars per trip and steal about 72 dollars worth of merchandise. Blacks spend less on average, are a much smaller percentage of the gross income of your store, but according to your observations, steal at double the white ratio. Others spend about on par with whites but less overall than blacks. One of the others rips you off big time, but per capita by racial group this is below both the black and white crime rate.

So on the basis of the rate of crimes committed by race, you decide to profile only the blacks. The results are twofold. The first is that you alienate half of your black customers, still one black gets away with lifting about $5 worth of merchandise. The second is that you reduce overall crime by 40%, and normalize the rate of crimes between blacks and whites. In the end your bottom line is that you have lost only about $65 in revenue, and maintained better than 93% of your customers.

Is that an acceptable loss?

Part of the problem here is that by identifying crime rates by race and observing the difference, you set up a standard by which some crime is justified, in this case, 'white' crime. by such a standard some race is bound to be overly persecuted in this case, 'blacks'. the very act of initiating a crackdown on criminals *by race* even if the statistics 'justify' it, is to set up a differing standard by which individuals are judged in the justice system. this is racist even if this the actions are restricted to the class of known and observed criminals. You end up treating one race of criminals worse than another race of criminals.

In fact, racial profiling is not restricted to a population of criminals. The effects are felt against the general population. In this case you alienate the innocent black general population as well as the criminals. By profiling the black population you are in fact treating all blacks as if they were black criminals, which we have already established are getting a worse deal than white criminals.

However, if you are only concerned with profit, it's clear that you can maintain such a racist policy with a minimal impact on your bottom line.

There are a number of variables which I think should be added to the analysis. Today, I'm not so interested in Korean groceries, which prompted this study, but the War on Terror. It's clear that if we were to apply racial profiling to Arabs boarding airplanes, the situation would be worse in terms of effect on innocents. Firstly, it is a whole lot less likely that an Arab would be hijacking a plane than shoplifting from a store. So the crime ratios would be a lot lower. Similarly, Arab passengers would not have so many convenient alternatives as black shoppers. So the second observation would not be too far different from the first.

Just a little food for thought.