Over at the Washington Post’s Wonkblog, Dylan Matthews has been tearing it up with a series of posts on increasing tuition he titled The Tuition Is Too Damn High. Thus far, the three must-reads are the following entries:
- Part 3: The three segments of public/non-profit higher ed, and why the tuition increased for different reasons in each. This entry also has a beautifully-designed set of charts.
- Part 5: Why Baumol’s cost disease does not explain increased tuition in public research universities.
- Part 6 (yesterday’s): Why (Howard) Bowen beats Baumol, or research universities as a vacuum with education as the loss-leader.
In his entry yesterday, Matthews follows Robert Martin and R. Carter Hill in arguing for a revenue theory of cost, or asserting that “universities will spend all the money they can possibly raise. If they raise more than they need for educational objectives, they will spend it on non-educational uses like climbing walls and nicer buildings.” Martin and Hill have accumulated sufficient evidence from about a quarter-century of financial data to make the Bowen hypothesis a much more likely explanation than the Baumol cost-disease argument.
And two days after my university’s football team was slaughtered on the field by McNeese State University, to whom the Athletic Department apparently paid $400,000 as an inducement to an expected massacre (only the other way around), the spending of that amount of money in this way certainly looks like a Bowen-like behavior.
But there are three problems with the naked Bowen argument: relative priorities, net assets, and the three-sector pattern Matthews explained before. First, the Bowen argument may explain total spending but not relative priorities among different spending categories. Last Thursday, several dozen USF students protested all night in front of the library, which was closing at midnight during the academic year for the first time in several years. The cost of keeping the library open 24 hours, 5 nights a week? $136,000. $400,000 payment to McNeese State for a single football game vs. $136,000 to keep the library open for students. As my university faces a lower long-term set of revenue streams for education, it continues putting some state/tuition dollars into athletics — not a lot, but many times more than the cost of keeping the library open, and I am aware of no mandate for the Athletic Department to give up any of its state/tuition dollars (the only VP area excused from state/tuition revenue cuts). Why does USF allow the Athletic Department to keep spending money while it reduces library hours? I have some guesses (and I know the students protesting the library hours cutback have some, too!), but the Bowen argument does not. Without an explanation of how a university sets internal priorities, the Bowen argument is incomplete.
The second problem is all about the money, or about the savings. While the strongest form of the Bowen hypothesis says that universities will just spend more if they raise more, that is not true in fact. Colleges and universities raise funds for endowments as well as current operations, they build up cash reserves at different rates, and they also go into debt in different patterns. If the strong Bowen hypothesis were true, then colleges and universities would have no cash reserves beyond what is legally required, would spend down their endowments, and would be very cautious about committing themselves to debt. But none of that is true, or that is frequently untrue when you look college-by-college. Again, this is an indication that the Bowen model is incomplete, not that it is deeply flawed.
The final problem with the Bowen argument is the failure to explain the different patterns of raising tuition among the various sectors of higher education. What about research universities give them the ability to increase raising revenues (and then keep increasing spending) in ways that other types of colleges and universities do not? And what explains the timing of that? While I suspect the pattern Martin and Hill identified stretched back a bit before the 1987 start of their data set, my guess is that it was different before the early 1980s. So what happened to public research universities that did not happen in other parts of higher education?
The Bowen hypothesis for research university spending patterns is plausible and has some empirical evidence but is not complete. I hope someone will pick up the baton and flesh it out, or at least begin to poke at the questions I have raised here.