It only now occurs to me that the rapid rise of massively open online courses (MOOCs) in the last year or so may allow us something of a natural experiment related to the incidence of the “profit motive” in education. Over recent decades (and particularly so over the past few years), there has been substantial discussion over the role of profit in higher education in the context of pitting for-profit colleges and universities against the non-profit sectors. In one sense, this really isn’t the appropriate context in which to make a cross-sector analysis. For one thing, the history and educational foci of the for-profit sector has been considerably different than that of the non-profit sectors. For another, as Cower and Papenfuss ably point out in their chapter in CCAP’s book, Doing More with Less, there are good reasons to suppose that the for-profit model may not, in fact, be as appropriate as the non-profit model in the context of providing quality general education, at least insofar as we have traditionally conceived of it. (On the other hand, they argue [pg. 186] that the “traditional efficiencies of for-profit firms dominate” when the “relevant [educational] output is easily defined and measured,” which explains why for-profits have tended to specialize in vocational training programs.) In other words, noting that Harvard, say, does a much better job at delivering a liberal arts education than does Kaplan University, for example, may not actually tell us anything meaningful about the efficacy of the profit motive in education (this effect is further compounded by the fact that Harvard has a much better established educational pedigree than does Kaplan).
In the case of the MOOCs, though, we may be able to get a much clearer picture of whether or not a profit-based education model works as well as a non-profit model in offering the exact same educational services. On the for-profit side, we have Coursera and Udacity (though even here there are notable differences between the two), and on the non-profit side we have EdX. Given that they all are operating on more or less the same educational model (offering courses to practically anyone who wants to take them, for no formal credit) and all arose at approximately the same point in time, we can actually now control for some of the factors which already may explain differences between for-profit and non-profit colleges. We can, in the course of this natural experiment, actually apply our “other things equal” assumption and know that it will hold. By doing so, we can then get a much clearer and better picture of the efficacy of the for-profit model in at least some aspect of higher education, though care must be taken not to overly generalize any results from this experiment.
According to data published by the Federal Reserve Bank of New York, in the fourth quarter of 2012, delinquency rates on outstanding household debt fell overall (from 8.9 percent in the third quarter to 8.6 percent). However, there was one debt category for which 90+ day delinquency rates actually increased: student loan debt (the delinquency rate is currently at 11.7 percent). The delinquency rates for various consumer debt categories is shown in the following chart. As the New York Fed itself points out, though, this may even understate the problem:
these delinquency rates for student loans are likely to understate actual delinquency rates because about 44 percent of the borrowers are estimated to be currently in deferment, in grace periods or in forbearance and therefore temporarily not in the repayment cycle. This implies that among loans in the repayment cycle delinquency rates are roughly twice as high.
Simply put, higher education is setting itself up for failure by making promises it will not be able to keep. Does anyone really believe that we can create a system where every student who enters college graduates four years later with a degree, debt-free? Or that we can have classrooms where all students learn the same amount and in the same way? Or that every college graduate will land the job of her dreams? Higher education has never, ever done that. Not in the 19th century or in the 20th. And it never will.
Occasionally the Bennett Hypothesis, which holds that federal financial-aid programs contribute to college-tuition increases, gains support from unlikely sources. A year ago, for example, Richard Vedder pointed out in this space that Vice President Joe Biden, contrary to the received wisdom in the White House, endorsed the basic tenet of the Bennett Hypothesis. Now we have Kevin Carey, in a somewhat confused but nevertheless thought-provoking essay, announcing his agreement with that hypothesis while trying to deny that he does:
The thesis advanced in some conservative and libertarian circles that federal aid causes increased college spending has tenuous empirical support. But federal aid has certainly abetted college spending, allowing institutions to avoid the kinds of difficult structural reforms experienced in other industries.
The first sentence is entirely defensible (even if it were to turn out to be incorrect), but I should point out that Carey is attacking the simplest (and laziest) version of the Bennett Hypothesis. Variations and improvements to the basic Bennett Hypothesis model (see this CCAP paper by Andrew Gillen) allow for the complexities of college pricing because those are necessary for us to explain why the empirical evidence on Bennett Hypothesis effects are often “mixed and contradictory.” It’s because of the mixed evidence that Carey makes a valid point when he criticizes the view that federal aid necessarily drives up college costs in all cases.
What is fascinating to me, though, is Carey’s next sentence in the above excerpt, which contains a surprising admission, given his previous sentence. His language (“federal aid has certainly abetted college spending”) is remarkably similar to former Secretary of Education William Bennett’s original formulation (“increases in financial aid in recent years have enabled colleges and universities blithely to raise their tuitions, confident that federal loan subsidies would help cushion the increase”). Despite his stated aversion to the notion, evidently Carey feels compelled to embrace the core of the Bennett Hypothesis. No wonder why it has stuck around with us for more than 25 years.
George Leef directed my attention to this tongue-in-cheek blog post by Don Boudreaux who imagines what it would look like if someone in Congress pushed for a “minimum grade” law. This may not be all that far-fetched in higher education. The University of Wisconsin at Oshkosh has already put in place a somewhat similar program (which will be expanded next year) in which the school basically increases pay for some faculty because others are paid more.
UCLA’s Higher Education Research Institute (HERI) recently released its Freshman Survey, the 2012 Freshman Norms, which reports responses from just under 200,000 first-time, full-time students entering 283 four-year colleges and universities. The survey questions cover why students want to attend college, which factors most significantly affect students’ matriculation decisions, and how confident the students are in being able to graduate in four years. Freshmen were also asked about their stance on common social issues and political leanings.
Among the highlights of the study, however, is the behavioral change in first-year students’ choice of housing. Overall, 79.3 percent of incoming students in 2011 reported planning on living in a dormitory, where as 76.1 percent report planning on doing so in 2012, a 3.2 percentage point drop. Meanwhile, there was a 2.2 percentage point increase between 2011 and 2012 among students who plan on living with family or other relatives.
Nevertheless, the HERI survey shows that there is a notable dichotomy between first-year students living on-campus versus those living with family. For example, 55.6 percent of those first-year students living with family report a “very good chance” of getting a job to help pay for college expenses. On the other hand, only 47.8 percent of those living in dorms report doing so. Also, 14 percent of commuter students report a strong likelihood of working full-time, compared to only 5.9 percent living in dorms.
Concerning financing the high costs of higher education, 48.7 percent of commuter students are likely to finance their first year through loans versus 62.3 percent of those living in dorms, a difference of 13.6 percentage points. Additionally, 21.1 percent of commuters are likely not to use family resources to pay for those first-year expenses against 12.7 percent of those living in dorms. In other words, there is a higher proportion of commuter students compared to those living on campus, who do not rely on loans or family to pay for their first year (of course, the fact that the students are living at home, we can view living expenses as an implicit family subsidy). As indicated earlier, the fact that they are working more than the students living on campus shows the measures that students have to resort to more and more as costs rise and, therefore, debt incurred from student loans would increase.
USA Today maintains a wonderful database on sports spending for more than 220 public colleges and universities across the country (they obtained the data through public records requests). This database, along with reporting total revenue and total expenses for athletic departments, also reports data on an institution’s “total subsidy;” that is, the amount of money athletic departments receive from students (in the form of fees as distinguished from tuition payments) or directly from the institutions. About three years ago, CCAP wrote a white paper that examined athletics subsidy patterns (for 2008) for a sample of 99 schools across the 11 FBS conferences. That report found significant differences across conferences in the proportional size of subsidies relative to total athletics revenues.
The chart below updates our 2010 analysis to include the latest data (for 2011) available from USA Today. The sample of schools included in this chart is the same sample (with the same conference affiliations) as in our 2010 paper. We kept the same conference affiliations to avoid confounding the effect of conference realignment (as opposed to the effect due solely to intra-institutional athletics-subsidy trends) on average conference subsidies (four schools in our sample—Utah, Colorado, Nebraska, and Boise State—were no longer affiliated with their 2008 conferences in 2011). The data still show enormous differences in subsidy patterns across conferences: whereas athletic subsidies accounted for three percent of total revenues in the SEC (which had more than $1 billion in revenues for 2011), athletic subsidies were more than 72 percent of athletic revenues in the MAC. Compared to the 2008 subsidy levels, the SEC, Big Ten, Big 12, ACC, Big East, and WAC saw modest decreases in subsidy levels (all less than 2.5 percentage points). The Pac-10 (now Pac-12), MWC, Conference-USA, and Sunbelt saw modest increases in the proportion of athletic revenues from subsidies (only the MWC saw an increase of more than three percentage points). The MAC, the conference with the highest level of subsidies, saw no change between 2008 and 2011.
Minnesota Public Radio’s “The Daily Circuit” featured CCAP’s new study on college graduate underemployment in a lengthy segment yesterday. Discussants included PayScale’s Katie Bardaro and Rutgers’ Carl Van Horn. Take a listen:
A segment for the Wall Street Journal‘s “MarketWatch” recent featured good paying jobs which don’t require workers to have a bachelor’s degree. For those students who may be worried about the cost of college (particularly given the risk of being underemployed after graduation), these types of jobs may be the occupational route for some students who are looking for alternatives to a job that needs a four-year degree.
In yet another dismal talking point about the job prospects for college graduates, they’ve been declining for six years.
As worrisome as the trend is, it’s probably understating the problem. It only counts full-time workers, not part-time workers, and doesn’t address the underemployment of graduates. Especially since 2008, I’d expect the trend to be more drastic.
The slide since 2000 could bolster Bryan Caplan’s claim that education works as a signaling mechanism; students are realizing smaller returns from higher education in the marketplace. Instead of gaining useful skills, the ever-increasing tuition teaches and prepares students for less.
Choosing to earn a graduate degree distorts the data as well; students who can’t find work after graduation decide to attend graduate school instead of joining the labor force in a low-paying, mismatched. If we look at graduate enrollment for 20-to-24-year-olds since 2006, it’s constantly increased:
I’m unsure how to explain the average wage increase during the mid-2000s; perhaps more graduates chose employment over graduate school (and more debt), whereas more graduates chose graduate school instead of low-paying employment. Thus, with fewer graduates in the job market who took low-paying jobs, the wage increase wasn’t weighed down. Or, the market simply had an increased demand for graduates; from 2003 to 2006, United States GDP grew by $2 trillion.
If universities are over-educating students for jobs that don’t exist (regardless of whether economic growth lags or increase), students become over-educated, underemployed, and debt-ridden. They successfully avoid some pain in the short run for greater pain in the long run.