The Bennett Hypothesis Redux

Research into effect of Grad PLUS loans on program costs

Was this forwarded to you by a friend? Sign up, and get your own copy of the news that mattered sent to your inbox every week. Sign up for the On EdTech newsletter. Interested in additional analysis? Try with our 30-day free trial and Upgrade to the On EdTech+ newsletter.

As a recovering social scientist, I get excited by good research. The best research is often counter-intuitive and challenges one’s deeply held assumptions and predilections. When it comes to higher education, I have, shall we say, more than a few of these.

Two of my assumptions, both of which are the subjects of vociferous debate in education policy circles, are the following:

  1. Increasing access to education and attainment has to have a positive impact on equality. This is the is the topic of a couple of books I have on my summer reading list: Jon Shelton’s just released The Education Myth and Cristina Groeger’s The Education Trap.

  2. That universities definitely increase costs in response to the government raising the amount that students can borrow (what is sometimes referred to as the Bennett Hypothesis, named after Ronald Reagan’s Education Secretary).

I want to direct my focus to the latter point. The Bennett Hypothesis is important and has captured the attention not just of policy experts but also a wider audience. It has become a significant part of the ongoing discussions regarding the factors driving student debt and potential solutions (as seen in sources such as here, here, and here). Unfortunately, some of the more popular literature that references the Bennett Hypothesis fails to properly engage with the available data, which is highly regrettable. To have honest and fruitful debates about the cost of higher education and student debt, it is essential not only to look at data but also to recognize its limitations and carefully interpret the findings. This post aims to make some modest progress in that direction.

New Paper on the Bennett Hypothesis

Sandra Black, Lesley Turner, and Jeffrey Denning just published a new paper on the impact of Grad PLUS on student persistence, degree completion, and program cost. In 2006 the government introduced Grad PLUS student loans allowing graduate students in graduate to borrow the full amount up to the total cost of attendance (tuition, fees, living expenses, books etc.). Prior to this, student borrowing at the graduate level had been capped at $18,500 per year. Essentially Grad PLUS vastly increased the amount of money students had available to pay for school.

There has been a lot of research on the Bennett Hypothesis, but in general terms the consensus seems to be that the impact of higher loan limits and Grad PLUS loans have had little to no effect on higher ed prices. In this latter category, the work with which I am most familiar is that of Robert Kelchen. He has focused on the impact of Grad PLUS on professional schools: medical, business, and law schools in particular. He finds little empirical support for the idea that giving students relatively unconstrained access to loans increases university prices.

The paper by Black et al was pretty exciting and a welcome addition to the literature, as Kelchen himself acknowledged.

In general terms, the authors find that the increase in borrowing limits introduced by Grad PLUS did not improve access to graduate programs overall or for underrepresented groups. Conversely, they find Grad PLUS led to a very small decreases in access for Black and Hispanic students and very small increases in access for White students. Grad Plus seemed to have no effect on persistence and degree completion.

However, they did find that in programs where there were larger numbers of students who needed Grad PLUS loans (because they had exhausted other federal options) tended to show a higher-than-average cost of attendance. More simply put, where students used Grad PLUS loans more, program costs went up more than at programs where they were using Grad PLUS less. The table below from their paper reflects this finding.

The effect of a 1 percentage point increase in baseline percent of students who are likely to be taking out Grad PLUS on program-level borrowing and prices.

As you can see in the chart, Grad PLUS loans and cost follow each other closely. The authors find:

prices increased by $0.75 per $1 increase in average per-student Grad PLUS loans and more than dollar for dollar with increases in total federal student loans.

This would seem to be proof of the Bennett Hypothesis – as the amount available for students to borrow increases, so universities will charge more. But as dug into the details I found their conclusions a bit underwhelming. They find that average Grad PLUS borrowing increased by $79 for each 1% increase in the number of students who needed Grad PLUS loans (because they had exhausted other federal sources). For each 1% increase in students needing Grad PLUS they see a $60 increase in the average cost of attendance. But overall annual federal loan aid per student increased by only $54 (pp 23-24).

While I started reading the paper hoping to have my implicit assumptions proven right, I increasingly found myself less persuaded. Like Kelchen, I like the paper, and I think it is great research. But even apart from the fact that it didn’t support the Bennet Hypothesis, it leaves me with a lot of unanswered questions. And maybe that is another hallmark of good research.

These are my quibbles and questions.

It’s a Difference of Degrees

I find Black et al’s breakdown of types of graduate degree to be problematic. They distinguish between professional degrees (such as law, business etc.) and more academic degrees such as PhD’s. They describe their logic thus:

if more than 85 percent of degrees within a (program) are terminal master’s degrees, we classify the program as a professional master’s degree program. If less than 85 percent of degrees are terminal master’s degrees, we classify it as an academic doctoral program

I would argue that this binary classification misses an important type of program, specifically the kind of masters that aren’t professional, but that aren’t part of a PhD program. These are important degrees because they are more likely to be the kind of “cash cows” that saddle students with a lot of debt but not a great promise of payoff (see these great pieces on this phenomenon). I would love to see analysis that captures the impact of Grad PLUS on these cash cow masters, or even whether the introduction of Grad PLUS availability is correlated with more of these types of programs being started at universities.

Impact of the 2008 Recession

Black et al discuss the impact of the 2008 recession on their results. They argue:

Our results suggest that Grad PLUS loans primarily benefited institutions and programs that were able to charge higher prices. However, the program was established in an era with robust private student loan offerings and lenient underwriting standards… As was the case with mortgage lending, the onset of the Great Recession led to substantial tightening of creditworthiness standards and contraction in private student loan availability. While our estimates of the effects of Grad PLUS do not appear to vary substantially before and after the start of the Great Recession, it is .. possible that…students substituted from private to federal loans.

The 2008 recession happened soon after the introduction of Grad PLUS loans. That seems to be an important fact that not many have explored. The recession had a big impact on higher ed beyond the shift from private to federal loans. Universities were under enormous financial strain and were doing all kinds of things to bring in revenue. This is something I think should be looked at in more detail.

The Mechanism of Raising Prices

A frustration I have with a lot of the argument’s pro and con on the Bennett Hypothesis are the way that they gloss over the way that this happens at universities, in the messy business of setting budgets and prices and making decisions. Black et al argue that:

Universities, recognizing that students have more ability to pay when loan limits are increased, may try to capture some of the additional funding through higher prices.

This is economist speak. I say this lovingly, as someone who has spent a lot of time in higher education, but universities are not terribly good at responding to things very quickly, if at all. I find it unlikely that folks in the Bursars Office said hey, the Grad PLUS loans will fund more so let’s raise those prices!

If it did happen, it is far more likely to have been a slower process. It would have been complicated by a lot of other things going on at the same time (state cuts to higher ed budgets, increased fee-paying international students etc.) and influenced by the extensive ways higher ed institutions watch, learn from, and copy their peers. Was there contagion, and if so, how did it spread?

Parting Thoughts

Student debt is a big problem in US and one that we need to understand and fix no matter what our incoming assumptions are. Having research like this most recent paper and good healthy debate is critical to making progress.

The main On EdTech newsletter is free to share in part or in whole. All we ask is attribution.

Thanks for being a subscriber.