Response to Comments on Affirmative Action--Posner
The principal criticism of my posting is that "merit," understood as doing well on exams (especially timed exams), is too narrow a basis for admission to college or law school and that affirmative action is a way of rectifying the mistakes caused by the overemphasis on that too-narrow criterion. My view is that reference to "merit" and "meritocracy" is misleading. A person is not "better" because he's a better exam-taker; for that matter, he's not "better," more "meritorious," because he has a higher IQ than someone else. The issue regarding standardized testing is whether it's a good predictor of college or graduate school performance. If it is, then people who do badly on the test, but are admitted anyway because of affirmative action (or because they're good athletes), are going to do poorly in college or graduate school and cluster at the bottom of the class.
Now maybe though they cluster at the bottom of the class, they do well professionally because grades are not a good predictor of performance in the "real world." So the argument would be that blacks from poor families do badly on the SAT and in college and law school but nevertheless do well professionally, because SATs and LSATs and the rest of the educational testing apparatus are poor predictors of professional success.
Now it would be odd if race were the explanatory variable here. That is, if you took two people otherwise identical in upbringing, parents' occupations, etc., but one happpened to be white and one black, on what theory would standardized tests underpredict the black's professional success relative to the white's? Presumably the relevant variable in explaining black-white test differences would be not race as such but such factors as parents' education, household income, early schooling, etc.--factors that might well be correlated with race, but that would not be identical with race. If parental income or some other such variable is thought to cause students who have in fact great professional talent and prospects to underperform in standard tests, then that would be an argument not for affirmative action on the basis of race, sex, ethnicity, etc., but for affirmative action on the basis of parental income or the other nonracial factor that was causing the difference in test scores. This is resisted because the colleges don't care about students from poor families, etc.; they just want a certain percentage of blacks.
There is a special factor at work in law, the profession with which I'm most familiar, that casts particular doubt on the wisdom of racial affirmative action. That is the fact that to become a practicing lawyer, you have to pass the bar exam--another standardized timed test but one for which you can't substitute a take-home exam or a term paper. The black pass rate on the bar exam is shockingly low--something like 15 percent, compared to more than 60 percent for nonblack exam takers. I cannot see the sense of bending law school admissions standards in favor of applicants who are unlikely to be able to enter the profession after spending $100,000 or more for three years of law school tuition.
A number of comments mentioned "diversity" as a valid ground for affirmative action in admissions to college and law school. I agree that one benefit of college education is meeting a more diverse group of young people than one might have encountered growing up in one's particular community, which might be a lily-white suburb. But the relevant diversity is not in the color of one's skin, but in attributes which, to repeat, while they may be correlated with race, are not identical to it. There are black people who really aren't different from white people, and it is unclear how their presence increases the diversity of a student body.
Several comments from the right side of the political spectrum of our readership accused me of having a double standard--favoring or at least being willing to tolerate some discrimination against whites (i.e., some affirmative action) but not willing to tolerate discrimination against blacks. I plead guilty to the double standard. I do not think discrimination against blacks by whites, and discrimination against whites by whites, are symmetrical phenomena. A dominant group may discriminate some against its own members--that's what affirmative action is--but it's not going to go too far, whereas discrimination by the majority against a minority is likely to be far worse and more injurious.
I am always interested and pleased when the comments go in unexpected directions, focusing on what I had thought distinctly peripheral aspects of my posting. I had said that I thought a theatrical producer should be permitted to refuse to hire a white actor to play Othello, or a black actress to play Desdemona. Several comments pointed out that there have been theatrical productions in which a white played a black, a woman a man, etc., and they noted that in Shakespeare's time, because women weren't permitted to appear on stage at all, female roles were played by adolescent boys--and since there were virtually no blacks in England and almost certainly no black actors, Othello was played by a white. So a white male was playing opposite another white male and why shouldn't that be permitted today? Well certainly it should be permitted, but the question is whether the producer should be deprived of choice in the matter.
It is further true, as one comment points out, that while sex can be a "bona fide occupational qualification" under federal antidiscrimination law--and so a producer can insist that Desdemona be played by a woman, whatever Shakespeare might have thought of that--there is no BFOQ for race. I consider this rule of law mistaken. It seems to me that, at least if one is speaking of producers in the private sector, "discrimination" in the form of matching an actor's race, etc. to that of the character he or she is playing should be permitted. The impact on vocational opportunities for members of racial and other minorities is likely to be small. Of course there are not as many black characters in drama as there are white ones, but then there are not as many blacks in this country as there are whites. And, on the other side, matching the physical appearance of the actor with that of the character he's playing is important to an audience's understanding and enjoyment of a play. In my view, that benefit, together with the principle that producers and other creative persons should have maximum freedom from government restrictions in deciding what to present to their audience, outweighs the cost to those minority actors who may occasionally lose an opportunity to play someone of a different race. Indeed, it seems to me that freedom of expression requires no less.