Formulated by former Secretary of Education (and now conservative commentator) William Bennett, the Bennett hypothesis holds that increases in student aid by the government (e.g., Pell grants) have the counterintuitive effect of increasing tuition.

The basis of the Bennett hypothesis is straightforward economic theory. If you give a subsidy to buyers of a good or service traded in a perfectly competitive market, you shift the market demand curve vertically by the amount of the subsidy, which in turn increases the market price. Of course, what is not often stated by those who believe the Bennett hypothesis is that as long as the demand and supply curves have some slope to them—as long as they are not strictly vertical or horizontal—the net price paid by buyers (market price less the subsidy) will still be lower than it would have been had there been no subsidy at all.

Interestingly, though, as Washington Post reporter Dylan Matthews points out, the empirical evidence for the Bennett hypothesis is mixed. To me this is not surprising. For one thing, even if you believed that the market for higher education was perfectly competitive, economic theory tells us that it’s still possible for government-provided financial assistance to have a minimal effect on tuition. If competition among schools was intense enough for the supply curve to be nearly horizontal, the presence of government-provided financial aid would shift the demand curve upward without increasing price very much. The same would be true if consumers were extremely insensitive to price, making the demand curve very steep. Both conditions—flat supply curve and steep demand curve—are plausible characteristics of the market for higher education.

But a better explanation for the mixed empirical support for the Bennett hypothesis is that the simple textbook model of a uniform subsidy received by all potential consumers in a perfectly competitive market does not capture the complex reality of the market for higher education. Most institutions have some degree of market power, which enables them to charge different students different levels of “effective” tuition—the list price minus any financial assistance received. Because most government-provided financial aid is means tested, those who receive it are probably not at the “top” of the demand curve, meaning they do not have the highest maximum willingness to pay for higher education.

Yet it is the top of the demand curve where those who pay list price reside. Indeed, list price is based on their high maximum willingness to pay and, one suspects, their relative insensitivity to price increases.  Increases in government-provided financial assistance may shift the mix of school-provided versus government-provided assistance received by those who pay less than list price, but it won’t change the list price. (Unless, that is, there is a significant risk that the students with the highest willingness to pay could, by hook or crook, qualify for means-tested government aid).

The Bennett hypothesis is a serious idea, but beware when you hear a pundit or politician talk about how federal assistance to colleges and universities has driven up costs. The empirical evidence on this point is mixed, as are the implications from economic theory.

Editor's Note. David Besanko is a professor of management and strategy at the Kellogg School of Management. Photo credit: Tax Credits.