« Let's take from the 1% what really matters. | Main | Errors In Government Spending »

December 10, 2011


TrackBack URL for this entry:

Listed below are links to weblogs that reference The Effect of Tuition Subsidies On College Tuition:



So should we have a graph of SAT score against poverty score and scale subsidies according to the matrix? Then work off your debt year by year practicing your profession among the needy, in the military, or some other socially constructive if less remunerative venue?


But Paul Krugman says that "no economic analysis" posits upward sloping supply curves. And he's a Nobel prize winner. So this analysis must be wrong.

More generally, it's clear that economists just make up whatever they want to suit their political predilections. There's no objective reality or agreed body of knowledge, like there would be in a real academic discipline.

Virginia Postrel

With all due respect to Alex, and I appreciate his link, the analysis is not his. It's a clearly labeled excerpt from my column here: http://www.bloomberg.com/news/2011-12-09/u-s-universities-feast-on-federal-student-aid-virginia-postrel.html

The demand curves in your second case should also be more elastic than those in the first case.


What evidence do we have that the marginal benefit to the individual and society of college attendance is lower at Fayetteville than Chapel Hill?

Sure, Chapel Hill students are going to do fancier things when they grow up, but it seems that the policy rationale should be based not on the total effectiveness of a college's grads but on the amount to which a degree at the college matters. Your assumption that UNC will make a larger marginal difference doesn't seem certain at all.

This doesn't mean that subsidies are entirely beneficial, but at least it might mean that more of the benefits are going where they're supposed to.



There is no evidence that, ceteris paribus, students coming out of top-tier schools do better in labor markets than students coming out of other schools. In fact, research shows no correlation after controlling for many other factors. http://www.nber.org/digest/dec99/w7322.html

But that is not the point I was making. My point is that the student most likely to improve her human capital by obtaining a college degree is more likely to attend a more selective school, while students who do not have the prior academic record, scores, and attitude to improve their human capital attend less selective schools.

There are plenty of outstanding students who will benefit greatly from a college degree attending the Fayetteville State Universities of the world; there is just a far greater concentration of them at the Chapel Hills of the world. On the other end of the spectrum, there are certainly students attending the Chapel Hills of the world who will not do much better in the labor market relative to having not gone on to college, but a far greater concentration of that type of student attends the Fayetteville State Universities of the world.


Thanks for the quick, thoughtful response, for the link to the NBER report and the email heads-up, Mark.

But either I still don't understand your argument, or you don't understand mine. Here's how I see the two:

Let x equal a future FSU student's human capital before entering college; let F equal the additional capital contributed (on average, etc.) by an FSU degree; let y equal a future UNC student's human capital before entering college; let N equal the additional capital contributed by a UNC degree.

Finally, let's say that for an FSU grad, x' = x + F. For a UNC grad, y' = y + N.

You write: "Consider the type of student we believe will realize positive returns, both to herself and to society, from a subsidy. ... The student most likely to improve her human capital by obtaining a college degree is more likely to attend a more selective school."

You seem to be saying we should assess a college's social benefit by comparing x' to y', or possibly x to y.

I'm saying that if we're trying to subsidize valuable educational products, we should compare N to F. That (regardless of the initial values of x or y) seems like a better measure of educational quality.


Or to put things another way: Given the choice I'd rather turn a bunch of 4 students into 6's than turn a bunch of 7 students into 8's.

I've got no evidence that FSU does this. But what if they do?

The comments to this entry are closed.