Intemper, Intemper!

Yesterday, in conjunction with the Pope Center for Higher Education Policy, Cato held a debate on how to control spiraling college costs. It was a terrific discussion, and I encourage everyone to check it out. It also got some coverage in the Chronicle of Higher Education, though nothing beats taking in the whole thing for yourself!

Concerning the latter point, I’d like to respectfully suggest that at least one person who saw only the Chronicle's piece imbibe the whole event, as well as what follows in this blog entry. Unfortunately, it appears that the Chronicle article got her a tad apoplectic, and I think she might have written some things she didn’t really mean.

On her blog, Sara Goldrick-Rab calls me an “ideologue,” accuses me of being totally ignorant of empirical research on aid and college prices, and even calls me “unoriginal” because in 1987 then-U.S. Secretary of Education William Bennett asserted what I did yesterday: that government aid to students enables colleges to keep raising prices. Oh, and she matter-of-factly declares that “after more than 20 years of this nonsense it's time to call the idea what it is-- just plain stupid-- and stop giving ink to the people who repeat it.”

Hey, wait! I need that ink…

Now, I might very well be an ideologue (if by that you mean someone who doesn’t pretend to approach every issue as if I’ve never given it, or anything else, any previous thought), and I might even be a pretty ignorant guy – nobody’s perfect, right? – but that doesn’t change several, glaring problems with Goldrick-Rab’s assault on my character and on the Bennett Hypothesis.

Let’s start with Godlrick-Rab’s understanding simply of what, exactly, was discussed at our event. She says that “faced with a thoughtful, responsive piece of federal legislation to reform the financial aid system, some ideologue had to come forward with a proposal to end federal student aid entirely.”

Presumably, the legislation to which she is referring is the Student Aid and Fiscal Responsibility Act, which is working its way through Congress. Of course, we ideologues might disagree with the utterly unsubstantiated description of the legislation as “thoughtful,” but I nonetheless think I know which bill she’s referring to.

Anyway, I have, indeed, been critical of SAFRA, but not primarily on the grounds that more aid will continue to fuel college-price inflation. I think it will, but what troubles me most about SAFRA is that it is likely to cost taxpayers tens-of-billions of new dollars, while its defenders claim it's going to save us major dough. More to the point here, though, SAFRA never once came up in the debate, nor was the event in any way framed around it. Indeed, the debate stemmed from a paper by economist Robert Martin that examines numerous potential drivers of the college-cost problem – a very probing study – and says not a thing about SAFRA.

Now, on to the substance – such as it is – of Goldrick-Rab’s treatment of the Bennett Hypothesis.

First off, despite what Goldrick-Rab asserts, I am well aware of the Hypothesis. For proof of my not-total ignorance, check out a 2003 op-ed I wrote about aid fueling college-price increases. Moreover, I am quite familiar with the conflicting research on Bennett’s proposal, and what it enables us to definitively conclude: neither that the Hypothesis is right nor that it is wrong. Indeed, despite Goldrick-Rab's pronouncement that Bennett's suggestion is “nonsense,” the quote from Harvard professor Bridget Terry Long that Goldrick-Rab offers to defend her case-closer makes clear that the research doesn't offer any definitive answers:

While several studies do find a college price response, their overall results are mixed and often contradictory. In summary, none of the numerous studies on the subject have found a "smoking gun" in terms of college pricing behavior.

Why no “smoking gun”? Because as Long explains in the article to which Goldrick-Rab links (apparently in an effort to prove me and Bennett wrong):

One possible reason for these conflicting results is that it is difficult to isolate the effect of government aid on tuition pricing from other factors. It is unclear whether changes in tuition are due to changes in the Pell or other general trends in higher education.

In other words, as is so often the case in social science, it is very hard to isolate specific variables to reveal only their effects on outcomes. In 2001, Cunningham, et al., were clear about the same thing, stating that the sort of statistical modeling they could do was incapable of providing definitive answers on the effect of aid on price. Indeed, they reported that:

Finally, even with future improvements in definitions and prospective data collection, the technique of cost analysis will always provide only partial answers to questions about the reasons for price increases at colleges and universities.

Perhaps illustrating how muddled the research is, in 2004, Long – to whom Goldrick-Rab points to attack the Bennett Hypothesis – looked at the effect of Georgia's HOPE scholarships on college prices and found that schools both increased prices and decreased institutional aid in response to the new state scholarships. In other words, she found significant evidence that the Bennett Hypothesis is, in fact, accurate!

But there's a flip side: Cunnigham, et al., found no clear correlation between aid and prices, but did report a strong correlation between decreasing state support for institutions and rising public-college tuition. This finding supports a popular ivory-tower explanation for price inflation: Increasingly tight-fisted states force schools to make up lost revenue through tuition.

But here we see another common problem in extant research: The researchers examined only a small span of time – from 1988 to 1997 – not a long enough period to see medium or long-term effects of aid. As the chart below (which uses State Higher Education Executive Officers data and I employed yesterday) shows, the period studied was a trough in state funding, and per-pupil revenue through tuition did rise. What the analysis missed were the periods before and after the one explored, when both tuition and government appropriations to schools increased.

Of course, there is more research than this (you can read about some of it here) and it is admittedly mixed. (In fact, I said in my presentation that aid is certainly not the only driver of colleges prices.) But as should now be clear, the Bennett Hypothesis is far from disproven, and statistical research may well be incapable of definitely settling the question. This has at least two major implications: (1) Calling on broad, long-term data – as I did at our forum – is highly relevant to the college-cost debate, and (2) if we want to get to the truth about the effect of aid on prices, it will be absolutely necessary to continue “giving ink” to the Bennett Hypothesis.