Binders are Better than Blinders

Joseph Fishkin

The most interesting moment in last night’s presidential debate, to my mind, was this: the most thoughtful defense of affirmative action I have seen in a presidential debate—and it was from Mitt Romney.  Moreover, this was a defense of affirmative action as it is actually practiced by the American companies that are serious about affirmative action.  Too many of our fights about affirmative action in this country take place in a kind of fantasyland where the employer or school always knows with perfect precision, before practicing affirmative action, who are the most qualified candidates—and then “affirmative action” means rejecting the most qualified in favor of less qualified candidates.  In reality, affirmative action often looks much more like what Mitt Romney says he did as Governor: making an affirmative effort to seek strong candidates from outside the usual channels, or to give some applicants a second look, out of an acknowledgment that one’s process otherwise seems to result in a pool with an overwhelming demographic skew (in Romney’s case, an administration of basically all men).

So I have been somewhat dismayed to watch today as various commentators, internet meme-makers, and gleeful Democrats turn Romney’s “binders full of women” into a punch line in the morning-after debate coverage.  Sure, it was inartfully phrased.  And yes, Democrats have plenty of justification for arguing that while Romney may have hired women as Governor, his actual policy proposals are not so helpful to the cause of equal employment opportunity for women.  For instance, he did not support the Ledbetter Fair Pay Act, and on the other hand did support the Blunt Amendment, which would have allowed all employers (not just religious groups) to kill contraceptive coverage for their female employees as a matter of “conscience.”  All that is fair game.  But I hope we do not miss the opportunity to have a little moment of appreciation for the usefulness, and frankly the importance, of the kind of affirmative action that Romney says he practiced as he took office as Governor.

For instance, we might ask: Why would Romney’s aides have initially come up with a pool of candidates that was essentially all male?  There are a number of different stories, none of them particularly good, but none of them particularly shocking.  We could talk about the subconscious bias in who leapt out to those aides as a strong, qualified candidate; we could talk about network effects in Massachusetts political circles, where who your friends are matters a lot, and most women are outside the old boys’ network.  (Speaking of which, what is up with Massachusetts politics?  For a Democratic state, it seems almost never to elect women to Congress.)  In any event, getting on Romney’s initial list required passing through a particular kind of bottleneck* in order to win those key aides’ approval.  It would seem that the shape of that bottleneck was, for whatever reason, not one through which women could easily pass.  On Romney’s account, he decided to do something about that, which is to his credit.

The alternative—a race-blind, sex-blind, generally-demographics-blind insistence that one’s process (whatever it may be) is perfectly meritocratic, and any diversity must be a sign of lowered standards—is just a fantasy.  Moreover, in politics, demographics are political—the optics of an all-male administration send an important signal, even if no such signal was intended.  I think most Republicans know both of these things, and furthermore, that their actions often reflect this knowledge.  But they rarely will admit such thinking publicly.  Publicly, we are told that Clarence Thomas, for instance, was chosen simply because he was the best jurist for the job, and without any consideration at all of the optics of replacing Thurgood Marshall with an all-white court.  Publicly, we are told that Sarah Palin was simply the most meritorious candidate for the job of Vice President, for reasons that had no connection to any efforts to diversify the Republican ticket.  I find these assertions silly—and I say that as someone with tremendous respect for Justice Thomas, who (contrary to ill-informed liberal stereotype) has turned out to be a powerful and important voice on the Court.

And so I appreciate that Governor Romney has come out and said that he practiced affirmative action.  Perhaps it is an easier thing for him to say in the case of sex than race, I do not know.  But he said it, and that’s important.  File it away just in case a Romney-appointed Justice is ever in a position to be the fifth vote for making what Romney says he did illegal.  We’re not at that point yet—current law on voluntary affirmative action in employment seems more permissive than the parallel body of law in higher ed—but certainly with a Romney victory, and the right retirements on the Court, that could change faster than you can say “binders full of women.”

So let’s talk a little more about what Romney says he did.  Local journalists in Boston, I should note, have actually been critical of some aspects of Romney’s account.  They note that the lists of qualified women were produced and carefully vetted by an outside group called MassGAP before Romney's election rather than at his behest; that Romney did not put any women in the most important jobs; and that the proportion of women declined over the course of his time in office (and has now risen rapidly under his Democratic successor).  But to my mind, these are relatively modest criticisms.  Romney did, it seems, order some affirmative steps to counteract what would have been an administration composed almost exclusively of men.

Now of course, this is a bit of a zero-sum game.  If Romney recruited more women, some men lost out.  Were they “more qualified?”  Would a Romney administration have been more effective at accomplishing its goals if Romney had stuck with his aides’ original list?  The two questions differ; the answer to the first is “maybe” but the answer to the second is extremely doubtful.  Merit is not a neat, tight calculus for most jobs, including jobs like these; it is hard to tell precisely which candidates will do best—especially over the long run, once they have the additional experience that comes with growing into a job or an education.

Sonia Sotomayor has said (and there is no reason to doubt) that she benefited from affirmative action as an eighteen year old, admitted to Princeton.  A couple of decades later, anyone who saw one of her oral arguments at the Second Circuit could tell you that you would be hard-pressed to find a more perceptive or incisive jurist.  To state the obvious, people learn things.  “Merit” is not some quantum deep in eighteen-year-olds’ souls, so that those who score highest on a standardized test at eighteen will continue, if everyone is given the same opportunities, to win every future contest of every other sort.  That’s not how life works.  Young Barry Obama was similarly a person who, by his own account, benefited from affirmative action.  He has turned out pretty well.

Most of our court battles about affirmative action take place in arenas where there are large-scale processes of admissions or selection—decisions on an industrial scale.  Think college admissions at the Universities of Michigan and Texas, or firefighter selection procedures in New Haven or New York City.  When you make decisions on a large scale about many candidates, you need to standardize both your procedures and your substantive standards.  This is a good and necessary thing, and one that actually may help reduce some important kinds of bias.  But it also sometimes creates an illusion of precision in the evaluation of merit, an illusion that we can really tell with great reliability who are the more “qualified”—and who will ultimately perform best.  The truth is that even the best predictors are often not that predictive.

This is especially the case in college admissions, where many steps (and much learning) lie between the admitted eighteen-year-old and her future career and life.  As William Bowen and Derek Bok found in their important book, The Shape of the River, even if you have a metric that can predict freshman-year grades fairly well (and a combination of high school grades and scores certainly has some predictive value), you still may know less than you think about how your graduates will fare in life and what they will contribute to society.  Of course, admitting students who are so unqualified for your college that they will fail to complete it does no one any favors.  But beyond that, things are not so clear-cut.  When liberal arts colleges today practice substantial gender affirmative action (to admit more men, who on paper are now systematically less qualified than the women in their applicant pools) it is hard for me to believe we can really be so sure those men will perform worse than the women over the very long run.  So it doesn’t seem to me such a problem for those colleges to seek a degree of gender integration, both for pedagogical reasons and for optics.

I think we may be better able to appreciate the contingency and tentativeness of our evaluations of merit in contexts like Romney’s cabinet choices or the hiring of high-level executives—where affirmative action, when it is practiced, looks more like the story Romney told at yesterday’s debate.  In those contexts, we all know that some people work out very well and others very badly—and that if one could really make a confident prediction in advance of which will be which, one could become very rich.  Performance prediction is hard.  If the standards we are using to predict performance (and thereby to select candidates) are selecting nothing but white men, it is of course possible that this is because our standards are perfect, and white men are the only people who can optimally perform the job.  But it is also possible, and probably far more likely, that something about our standards or our approach to recruiting is needlessly creating a narrow bottleneck* through which a lot of other people, who may be just as good, cannot pass.  And perhaps in such cases we should take a cue from Mitt Romney, and actually do something about it.


*You can expect to hear quite a bit more from me about this idea in a series of future posts.
Posted 11:30 PM by Joseph Fishkin [link]