My friend Karl Scholz from Wisconsin was here on Friday presenting his paper on credit constraints (co-authored with the Meta Brown and Ananth Seshadri) in a joint session of the labor and public finance workshops.
The clever idea underlying the paper is to notice that the government's formula for the expected parental contribution does not vary with the number of college-age students a family has, conditional on their total number of children. This means that, for example, two families that are otherwise identical except that one has two children of college age and one younger while the other has one child of college age and two younger are eligible for different amounts of financial aid. Put another way, the second college-age child in the first family has greater access to credit. If families make (economically) efficient investments in college even in the absence of financial aid, then exogenous variation in the amount of financial aid of this sort should not affect years of schooling obtained. Comparing families like the two above then provides a test of credit constraints. Being cautious sorts, the authors go even farther, and look within families with three or more children at differences across children who do and do not have siblings close in age. Doing so, the authors find surprisingly strong evidence of credit constraints. As expected, these effects are concentrated among middle income families - poor families have very low expected family contributions regardless of the number of children so there is no variation for them, a similar point applies to high income families.
There are, as always, issues. Can one generalize from families with at least three children to families with just one or two? What about the wealth effects that arise from the fact that some financial aid is in the form of grants rather than loans and the fact that the loans are typically subsidized? How would things change, in either the estimation or the interpretation, if the paper worried about college quality rather than just quantity? Also, these are effects conditional on the current policy regime. In a world with no government loan and grant programs families would save more for college (and college prices might well be lower) which would alter the game.
Also, this is one of those literatures in economics that has to some extent gotten fixated on an uninteresting null hypothesis. There are lots of papers by top scholars that address the question of whether or not there are "no" credit constraints. It seems to me that the interesting question is how substantively important the credit constraints are rather than whether or not they exist at all as I do not think anyone writing in this area believes them to be literally zero. Put differently, when looking at the tables I would spend more time on the point estimates and less time on the *s.
Overall an excellent paper with which to end this semester's seminar season.
Who was my favorite student this term?
11 months ago