I was waiting until the final post in this series showed up, but it was supposed to be Monday and hasn't appeared, so I'll just go ahead and post links to all the posts so far. Rick Sander has been blogging about his research on affirmative action at law schools at The Volokh Conspiracy. A lot of it matches up with things I already knew, but I picked up some interesting facts from his observations and arguments.
The opening post explains where he's going and points out that he's politically liberal with a strong history of supporting civil rights. I get the sense that this recent work of his converted him to the view that affirmative action is harmful and that he hadn't thought so earlier. Part 1 argues that three common views (and statements by practitioners of it) are just plain false: "(a) the preferences are small and not automatic, (b) race is one of a myriad of factors taken into account to create a diverse class, and (c) everyone admitted is fully qualified to do well at the school". Part 2 discusses the negative effect of affirmative action in law schools in terms of grades, graduation rates, and acceptance into the bar. Part 3 looks at the negative impact of all the prior effects in the job market. Part 4 predicts, with real numbers as the basis, what would happen if affirmative action would be removed. I already believed a lot of these arguments in the general case, but the way he's done this with law schools in particular and with hard data seems to me as if it should be pretty convincing even to those who start out believing these policies are overall helpful. Read on for more detailed analysis.
One interesting revelation is that the University of Michigan law school is even more by the numbers than the undergraduate admissions process. Why that's important is that Justice O'Connor's opinion for the two cases involving those programs in the summer of 2003 held up the policy at the law school because it was supposedly less by the numbers, which she and Justice Breyer thought made the difference between a quota-like policy and one that is constitutionally allowed. (For why such a ridiculous minority of two won the cases, see the post I just linked.) Sander argues that O'Connor and Breyer elevated form over substance. The policies weren't much different, and anything wrong with the undergrad case is substantially no different from virtually any affirmative action program for admissions. I agree.
Sander's primary argument is that affirmative action pulls black students up a tier or two to schools they otherwise wouldn't be able to get in, and their grades are lower as a result, the drop-out rate is higher than it otherwise would be, and they are less likely to be up-to-speed to pass the bar on the first time (or at all). That's an obvious harm of affirmative action, and it's hard to resist the conclusion that it has this effect, once you look at the numbers he presents. I couldn't detect any error in reasoning. He even considers whether name recognition of a better school in the job market can make up for this, and he calculates (based on statistics of who is actually hired by what firms) that it might barely balance it out for top tier schools with household names but that this effect does not come close to balancing it out with schools that are lower, where someone might not really be aware of how, say, Syracuse University and Boston College stand in relation to each other in the rankings.
The grades make a much bigger difference here than the school name, and the fact that some don't even get that far due to not passing the bar also can't be counterbalanced by anything to do with name recognition, because they're already out. So going to a lower-tier law school really can be a boon. Getting into a school that barely accepts you may not be the best thing, and affirmative action that pulls someone up to higher than where they could otherwise get in may genuinely be harming the person. (I want to point out that this argument is based on the numbers for law schools and won't as obviously and immediately also apply to undergraduate or other graduate programs, so this isn't going to be as easy an argument to make when I get to that point in my series on affirmative action).
Sander then asks what would likely happen if race preferences were removed from law school altogether. He estimates that the 45% rate of black law students' passing of the bar on the first try would increase to something like 74%. This is a guess, he admits, but it's based on real numbers. What this means is that the number of black lawyers could nearly double without affirmative action in law schools, and those black lawyers would be far more accomplished in their understanding of the law and the practice of the law. The average salary for black attorneys would be a good bit higher, and the distribution of black lawyers would range much more fully over the spectrum of possible positions, whereas now it seems to be focused more within government and business positions rather than in top firms.
What's interesting after all this is that he doesn't suggest removing race preferences altogether. Instead, he says it would be good to have a limit on it. No more than 4% of any incoming class can be given race preferences. Beyond that total, black law school applicants will have to face normal standards. This doesn't anger the political opponents of removing affirmative action anywhere near as much, but it has much of the effect of removing affirmative action. See the fourth post for more on why he thinks it will do this. My worry is that this comes dangerously close to a genuine quota. If you're setting a limit on how many blacks (or others, of course, though that's his focus) can be admitted through race preferences, admissions programs that already do more than that will insist on using all 4% of it. That will turn into a de facto quota. It's not technically a quota, since it's a limit. A limit doesn't allow you to go over but doesn't require that you do any. A quota, on the other hand, insists that you do all of it. As the official policy goes, this is right, but as it would almost assuredly be carried out it seems to be substantially what a quota system would do. That makes it much harder for me to want to sign on to something like that.