Without a doubt, one of the most fascinating and disheartening chapters in The Logic of Life is chapter 6 on "Rational Racism".

A test was done at the University of VA. Students signed up and were paid to be part of the experiment, which was done via a faceless web based interface. They were divided in two groups. Employers and Employees. The employees were randomly assigned colors: Green or Purple. The experiment was done in 3 phases.

First, the employees were asked to decide whether to spend a specified sum on an education, which would improve their chances on a test.

Then came the test, which, as it turns out, was a simple roll of the dice but a weighted roll which favored those who decided to spend on an education. Choosing not to invest in an education did not mean one would get a poor "score" but it did increase the chances while investing in an education improved one's chances of a good score.

Then came the hiring decision where each "employer" was presented two pieces of info on the employers: His test score and his color. The employers were then faced with a choice: Hiring a worker who had an education got him extra dollars and hiring one without an education got him docked dollars.

These three steps were done 20 times. Each round, employees had to weigh the cost and benefits of investing in an education against the potential pay off

As the rounds went on information started to accumulate and patterns began to emerge as the web interface revealed average scores and hiring rates of the two color groups. This was useful info because the employers were basically playing a game of odds and used the information they had to get the best possible odds of bonus cash. Likewise, employees had to weigh the cost and benefits of investing in an education against the potential pay off each round.

Results?

The first round was color blind since green or purple had no meaning. But from the second round on, employers had more info on color. As it turns out, more green students paid to get an education in the first round and this information started to emerge in the second round. Hiring rates were also available now by color. Remember, higher scores don't mean the employee had an education and since it was whether they had an education that got them paid, the employers started using the data to look for hints to get that pay off. On the other side, employees started to see whether gamble of an education was paying off...literally. Greens kept investing more and more since they had good hiring prospects. Purples did not. Why bother? It wasn't paying off. A vicious cycle ensued all the way to round 20.

After the experiment, the students all met face to face to discuss the results. Many employees (purple) were genuinely angry and frustrated when they confronted the employers. "Why wouldn't you hire me??" was a common question. The employers wouldn't hire them because they didn't invest and purples wouldn't invest because nobidy would hire them. The small hint of color from the first round, totally random, began to drive the logic of the gamble and rationally so.

The ;long and the short of it is that employers became "discriminatory" by color in a totally rational way...regardless of whether these college students thought of themselves as racist.

They chapter goes on to discuss this sad story in real life detail.

This is what Harford calls "Rational Racism".

Harford also recalls an experiment mentioned in "Freakonomics" where researchers sent out random resumes in the thousands with a varying quality with black and white sounding names. White names received 50% more call backs. Yes, better resumes got a lot more call backs...but only for whites! Blacks had a negligible difference.

Basically, there are two types of discrimination: Taste-Based (Bigotry) and Statistical (Rational Racism). Harford contends that statistical discrimination, while less repulsive on a moral level, it is far more dangerous and harder to correct because it pays off...unlike simplistic Bigotry...which could eventually harm the employers. Regardless, it could take time and still harmful.

Well this is a set up of the chapter. Which goes on into much more detail and examples examining statistics on slower progression through school even though the gap at the Kindergarten level is mainly social due to a higher probability of coming from harder up-bringings. He also devotes time to the idea acting white brought up by Obama at the 2004 convention, which contends that African American culture and social incentives have more to do with the continued gap than outright discrimination from whites. Harford goes on to cite further studies that give weight to this idea by conducting Q&A experiments with students. In short, unlike with whites, popular black students did not get a bump because of better school grades. In fact, more popular black students tended to have mediocre or low test scores. Black students with better test scores get mistrust from their peers since it shows a sign of wanting to leave the group.

He cites the idea that receptionist at an Ad firm hurts his/her chances of advancement within the firm if becomes known that secretary is taking night classes to get a better career in say, LAW. The boss may pass over the secretary since the boss does not trust the prospects of the receptionist staying with the firm. Why bother? Hmmm...being stigmatized for wanting to do better.

In short, we need to change the incentives of blacks who are stuck in this vicious cycle.

Explosive stuff. But sad.

Here, I pass off to Marginal Revolution , which is doing a book forum on the book. Bryan Caplan is the guest contributor for this chapter....and he tends to disagree somewhat with Harford on some of his conclusions ....strangely enough.

Great Stuff and very though provoking.

I welcome your comments.

UPDATE: Tim Harford replies to Caplan.