""But discrimination should also show up in another way. Employers who prefer not to employ workers because of their sex or the color of their skin are likely to lose money: Employing stupid white men when you could be employing smart black women is not a profitable human-resources policy. Employers might nevertheless do this, either because they do not realize that their prejudices are costing them money, or because they do not care. If so, discrimination is easy to detect in principle: Just note that the profitable firms will be the ones employing more women or workers from an ethnic minority."

Again, poorly thought out and sloppy application of economic and social science thinking. Has the author considered that discrimination may in fact be far harder to detect if social mores are such that a firm that DOES hire minorities is punished by the market? This is not unusual at all. Simplest example: Put yourself in the 1950s. How many African-American or Asian flight attendants would you expect to see? The cost would undoubtedly be lower, so by simplistic analysis above, we should be able to find firms hiring and making more money by having minorities on flights. Problem is of course, when discrimination is widespread and nasty, as it was and still can be, a firm that tries to grab these cost savings would have a negative PR backlash and hence lower profits, not higher.

So, what is this piece trying to say? As a paean to the economics profession for finding ways to detect discrimination, seems pretty weak based on above. Is author trying to say that discrimination will be wiped out by economic forces? Again, demonstrably false -- see above.

This column is getting old. Author keeps taking on fundamental and incredibly complex topics and slapping a bit of Micorecon 101 on top of them and calling this an insight. Please try harder.

Why do you assume that if an airline tired to hire black flight attendants, there would be a racial backlash? White people had no problem in the 50s (and earlier) when blacks attended to them on trains, taxis, and ships. And if an airline was venturing to Africa or the Caribbean, how do you know there in fact weren't black flight attendants? Maybe Mr. Hartford isn't the one who needs to go back and refine his argument.

Suppose 1950's air travel does doesn't illustrate an instances where hiring minorities would reduce profits due to the racial prejudices of its customers. Use your imagination and come up with another. Failed? Hot dog stand outside of a local KKK "watering hole."

I think the statistically significant effect of race on positive-negative reactions from customers is well documented in studies similar to the ones conducted on employers. So it's not a huge gamble to bet on a backlash. Your point is valid though-- the service industry is not a good example, as service jobs can be seen as reinforcing racial stereotypes rather than subverting them. Jobs that carry prestige or a sense of power, though, are a different story.

Take trains: Pullman porters made good money, but were seen as top-notch menials, and therefore black men were acceptable. Engineers, on the other hand, were seen as powerful and skilled, and therefore those jobs went overwhelmingly to white males.

Here's another example-- you hire a woman and she then outperforms your male employees. Do you advance her to management because she is the best qualified candidate?What if doing so could affect the morale of the business? Despite your own wish to be fair, sexist workers may not take orders from a woman, making her unsuitable for management based only on the failings of the other candidates! Purely economic concerns (the performance of the business) can both help and hurt minorities and women.

The US Armed Forces pioneered a lot of the studies in this area. They found a huge discrepancy between the reviews of black servicemen and their advancement rates. Despite recognizing that some of these men were top-notch soldiers who deserved to be officers, COs weren't comfortable reccommending them.

Thanks for that well thought-out reply. I wonder, though, what those studies ultimately prove beyond the immediate short-sightedness of certain people.

I also wonder about the difference between perception and reality. Has there ever been a case of a major business or corporation going out of business because they lost customers due to their hiring or promotion of black people, women, gay people, handicapped, or whatever? It seems like every time a big company does do something like that (everything from Major League Baseball and the US military integrating to NBC's decision to show "Will and Grace" to Barak Obama's run as a possible POTUS candidate), good things tend to happen. There will always be some bigots out there who won't change, but the long-term benefits seem to outweigh the costs.

If someone at a board meeting says "we shouldn't promote him because he's a ____ and our customers won't like it," just how valid is that mentality in 2008?

The US Army found that they had to set quotas to get advancement, and then put in place strict and formal orders about how command violations of orders from Black Officers would be treated.

It was only after decades of this sort of quota driven affirmative action that you were able to see someone like Colin Powell make it to the Joint Chiefs rank.

As for corporations - this is harder to evaluate because the market forces are that much more complex. American Family Association claims to have had an impact on Ford through their boycott. But of course Ford has face both a cyclical downturn, production and leadership issues as well as model designs that have not had favorable response in the public

Disambiguating such differences is almost impossible.

Your target market and the size of that market make a big difference in this assessment. If your company markets to say Southern Christian Radio, odds are you are not going to hire as many women or minorities in senior positions and you very much will not hire gays.

If your target market is the African American community of a nothern urban center (Chicago, NYC etc), you will be less likely to hire latinos and jews as your front-line sales staff.

If you don't follow those generalizations, will you go bankrupt? Not likely. Will you have measurably less revenue? Not easy to demonstrate.

That's why broad studies like the one in the article are so important.