Sunday, March 23, 2014
Nearly all law schools are nonprofits. I know, that isn’t the kind of hard-hitting economic insight you surfed here for. But nonprofits have a certain set of governance problems, by now pretty well-understood. A lot of what commentators have been pointing to as the problems of law schools are, in many ways, the problems of nonprofits generally.
A fundamental nonprofit problem is that customers have a tough time knowing whether they’re going to get a fair deal. If the quality of the experience they’re purchasing is difficult to measure, then opportunistic managers could divert some of the payment to profit, and cut corners on quality. As Henry Hansmann argues, the nonprofit form is supposed to be a way of efficiently contracting around this problem; by guaranteeing that managers cannot extract profit, customers are reassured that opportunism will be lessened.
So what happens at firms that conduct two different tasks---say, research and teaching---only one of which is the task customers are paying for? If monitoring allocation of labor between the two is difficult, managers again have an opportunity to “shirk” on the paid task to pursue the other. It may also be easier to measure scholarship than teaching, pushing faculty incentives towards emphasizing scholarship (the classic “two-task problem”). Rational consumers should be reluctant to pay the firm, given the large risk of opportunism available.
Although it’s not been put in quite these terms, a lot of the criticism of law schools fits this mold. “Scam” bloggers suggest that professors spend their time on irrelevant research instead of class preparation and time-intensive skills training. And they say that the high price of tuition is unjustified when much of it is spent in those unproductive ways. Customers, in other words, are starting to realize that the nonprofit form is no guarantee that their agents will be faithful ones.
Or that’s the perception, anyway. I believe it’s a false choice, but the data I’ve seen aren’t really conclusive. If good scholars make for better teachers, then the opportunism dilemma is no dilemma at all. I have lots of theoretical reasons for thinking that research and quality teaching are complements, but then it’s in my interest to think that. Let’s look at some evidence.
One data point is that major research universities have no lack of applicants. Students could choose to apply to equally selective schools with greater faculty emphasis on undergraduate education (Amherst, Wellesley), but they overwhelmingly apply to Harvard and CalTech instead. Maybe these are just unsophisticated consumers who don’t perceive the agency cost dilemma. More rigorous studies find some modest positive or zero correlation between research output at universities and student evaluations. Zero correlation is very good news for law school defenders of the status quo: it suggests shift of faculty energy to teaching costs students nothing.
Is law school different? There have been three major studies of the research-teaching correlation at law school. Merritt (1998), Barton (2008), and most recently, Ginsburg & Miles report little correlation in most of their specifications, which, as G&M say in their double-negative way, “[does] not support the existence of a tradeoff between scholarship and teaching.” (17). It’s possible this is too optimistic. For example, first-year courses are more demanding for instructors, and also tend to produce lower evaluations. That will tend to produce a spurious positive correlation between productivity and evaluations (upper-level teachers can publish more and get better reviews). But Ginsburg & Miles do control for first-year responsibilities in their regressions.
Another important finding G&M report is that the relationship between teaching and scholarship is not necessarily linear. (22-23). That is, there are some factors that align the two and some that push them apart. On balance, they find, the relationship is positive for all but the most devoted scholars. (23). They don’t “motivate” this test -- that is, we aren’t given any clear theory for why there might be two oppositely-signed effects. But it’s easy to think of some: maybe knowledge of the field and preference for work over leisure produce positive correlations, while time devoted to one reducing time for the other produces negative.
Should we believe any of these studies? I’m not sure. For example, Merritt uses a survey with self-reported outcomes. As she acknowledges, good teachers may have been more likely to respond; it would have been nice to see a regression design that accounts for selection bias, such as a “Tobit.” G&M’s data are all at Chicago, and as they admit it could just be that the Chicago environment is not representative of others.
More critically, as G&M hint at the very end of their paper, there is a big “endogeneity” problem for all these projects. That is, there are a lot of unobservable things happening in the background that could be driving the results. Maybe academic deans account for teaching quality and research productivity when they make course assignments. Maybe weak teachers select into courses that are more forgiving (although the counter-point there is that I’m voluntarily teaching tax classes, which get the worst evaluations of all in every study).
For my part, I continue to have good theoretical reasons to think that the agency problems of law schools are overblown. But I don’t think the data are there yet to show I’m right. What do you think?