I read the Education Sector (ES) report that inspired the news story and I have a few thoughts. First, I respect the analysis. Second, this is some drive-by thinking. I’m not purporting to be able to do a better job.
I’m motivated by two things:
1 Basic sociological principle: all social action will cause a reaction, usually not the one you wanted.
2. The never-ending trick of managing number 1 is to try to minimize the unavoidable unintended consequences of social action by focusing on mechanisms where you stand the greatest reward.
The actual report, in summation:
This report is a call for better, finer-grained data on student loan borrowing. It rightfully points out that federal data tracking is woefully unequipped to answer the kinds of questions many of us are most interested in. Those are questions like, do single parents who are poor enough to qualify for PELL have higher default rates than those who are just above the PELL threshold? Are there racial differences in default rates? Do those pursuing a certificate in cosmetology default more than those pursuing a master’s degree in engineering?
Second, this report is an argument for greater “accountability” in higher education. By the fifth page it’s clear that this analysis has staked out a neo-liberal, rationalized ideological position on accountability. That is important to remember. Just because the research takes for granted that accountability is defined as a discrete outcome variable and quantitative measurement of job rates is an objective measure, does not mean that we have to accept that. Neither does it invalidate the analysis. Just something to keep in mind.
Third, the report proposes a formula for the predicted defaulted rate of institutions based on the profile of the students it serves. If an institution with that profile has a higher default rate, it would be assumed that the institution is doing something janky. Janky. Yes, janky. I said it.
This report aims to clarify the limits of existing data federal data collection, to define accountability as predictable outcome measure, and give an example of how that might be used to help students make different decisions about where they will go to college. All admirable initiatives, not executed poorly, but with remaining questions.
First, recall my point about ideological assumptions embedded in analysis. This is a rational choice, econometric argument. As such, it has some common shortcoming of rational choice arguments. The first such shortcoming being that human beings are irrational as hell. We will mess up your model every damn time if you let us.
Individuals are rarely choosing from among the nation’s over 6,000 colleges. They are generally choosing from among the two or three in close geographic proximity. This is particularly true for lower-income and non-traditional students. They are also generally choosing from among schools that they perceive to be the “right” school for the type of student they perceive themselves to be. That’s basically a way of saying we interpret our social position and look at schools that fit within our perceived range of schools. That’s why poor single moms living in New Haven rarely apply to Yale, although it is geographically in the neighborhood. This is not a critique of the report so much as it is a limitation of its proposed accountability measures. Default rates don’t account for the aggregate risk factors of students that are in the geographical and social “sweet spot” of an institution. Those risk factors can be characteristically similar but not practically the same. For example, two institutions can serve the same percentage of PELL recipients and non-traditional students. But if one institution is located in an geographic area with an unemployment rate below the national average, while another location is in a city with very high unemployment we might deduce that graduates of the latter are entering a different job market than those in the former. And since loan repayment hinges greatly on a graduate’s ability to procure employment, that might matter to an institution’s loan default rates.
The report absolutely is clear that its proposed measure should be one of many measures to judge the quality of an institution. I trust their integrity on this. However, I should note that they say the same thing about end of year tests and SATs and ACTs and those are used as comprehensive measures to determine educational quality all the time. This makes me leery.
Finally, the report puts forth a list of “red flag” colleges. These schools have a higher than would be predicted default rate, using the accountability measure devised for the analysis. It’s a list designed to grab your attention (and headlines) and it does not disappoint.
I will use the figures from the report that USA Today used to compile its shame list although I prefer the adjusted default rate column. I think it is more interesting but the ES report doesn’t list the schools for that column. Of the 265 institutions with a default rate that is greater than it’s graduation rate and where 30 percent or more of the students had borrowed from financial aid, 117 or 44 percent of them were for-profit colleges. Whoa, right? You’re expecting me to go off here, right? Well, hold on.
19 of those 265 institutions are historically black colleges. That’s a mere 7 percent. However, keep in mind the sheer number of total for-profit colleges as opposed to the number of historically clack colleges. Of the 6,742 post secondary institutions in this country, 2,944 (44%) are for-profit and 106 (1.6%) are HBCUs. I do not point this out to start a conversation about HBCUs being in conversation with for-profit colleges. Trust me, I want no parts of that conversation ever. I do point it out because it points to the difficulty with universal accountability measures.
Anything we can currently devise to regulate the default rates of for-profit colleges will have a disparate effect on the small system of the nation’s historically black colleges. That may not be an issue for many people. However, it does get at issues of what a college is for and how we can compare institutional apples to oranges.
There’s an argument to be made that for-profit college default rates reflect the socioeconomic status of the students they are willing to accept. That argument is not much different than the argument that HBCUs make which can usually be boiled down thusly: “we take chances on students no one else will accept and we serve them with fewer material resources”. The reason for that statement is fundamentally different because of institutional differences. For-profits serve many for less because that’s how you extract profit. HBCUs do so because they have historically been deprived of equitable state appropriations, have low endowments and a mission to serve black students, who are still more statistically likely impacted by the list negative class based effects on k-12 education.
Frankly, this for-profit college versus historically black college narrative tension has discomfited me for some time.
The only way it works is to set up an explicit definition of college as an institutional social good. Real colleges would get the benefit of not being tied to job outcomes and less real colleges would not. Even then, well, as I said it is uncomfortable.
Accountability measures like this punt on precisely those kinds of larger issues. That is not a judgment on this fine analysis and what I am certain is earnest commitment from Education Sector.
As we are currently set up, poor students have to go somewhere. At least, we seem committed to a social policy and ideology that says that poor students must exist and that they must go to college to have a chance at not being poor. Socioeconomic risks make serving those students more expensive. The institution or the student will pay that expense, but it has to be paid. This measure assumes that the differences among institutions can be measurably mitigated by rational actors making better college choices and colleges doing whatever other institutions like them do to have lower default rates. Only, as I pointed out earlier, there is no a priori reason to assume that the differences are attributable to institutional behaviors. Theoretically, they could be attributed to differences in local labor markets or self-selecting behaviors among students.
So, I think it is a provocative list and a good argument for better data collection. The latter is something I am always fan of, the former not so much. The former puts together a list with lots of tantalizing treats for the rabid believers among us, self included. But, I can’t get away from that basic sociological principle of unintended consequences of well-meaning policy and the assumptions embedded in rational actor econometric theories such as these. Both portend something like this: an accountability measure will produce some kind of red flag warning to students who will choose to go to a school with better accountability ratings and this will either drive down tuition (more competition) or exert normative pressure on red flag schools to ship up or go out of business.
The alternative is that people won’t much care about red flags because they live in an area with two schools and one major employer. Comprehensive non-profit schools that offer better remedial education and services that address deficiencies in K-12 will be pulled under by attempts to legislate the predatory for-profits.
Enough sociological problematizing though, eh? I hear you. I grow weary of it myself but prognostications and prescriptions are tricky. But who reads this thing anyway, right?
There has to be some way of saying, hey, we take in students, we train them just fine, we charge them a fair price and some of them still don’t get jobs because the economy is pretty much trashed but we think they still deserved the education so let’s all eat that cost.
In lieu of that dream measure, I will also take better data collection for a thousand, Alex. But, also give me an accountability measure that transfers some of the empirical weighting of university outcome measures onto labor markets. It’s there implicitly anyway as jobs are the way most students repay loans. The economy keeps getting off here and that’s crazy to me. The structure of this accountability measure makes the same assumption most human capital theories make: that credentials are somehow operating both independently of the market and their value is a consequence of the market. They gotta choose one and stick with it. You can’t have both. Either the market matters or it doesn’t. I think the regionalism measure in the IHEP report has a way forward for all colleges, of all institutional types.
Barring all practical barriers, I’ve thought of a tiered system of accreditation for colleges that is based on some formula of who they serve, how they serve them. Tiered access to federal funds would be tied to this accreditation status. More apples and apples comparisons that way? It keeps the comparative nature of like students across institutional differences but minimizes non-comparable institutional differences (like region, status differences). Back to that sociological principle I opened with: the stratification is going to be somewhere. The trick to me seems to me can we put the bulk of that stratification in places that are the most sensitive to responsive regulatory mechanisms? Right now it’s all individual – get the student more information, more forms, more magic dust! – and I think that’s a disproportionate emphasis. It’s certainly seems to be the level at which social policy and regulation has the most indirect effect.
What we’re really going to have to do is decide once and for all if for-profit colleges should get the same access to the same level of funding as not-for-profits. Tuition is empirically shown to be pegged to maximum student loan limits and there is no way around the profit motive finding that attractive. If, as the president of ITT has the nerve to point out, default rates are tied to socioeconomic student factors we have but two choices. One, we have to accept either publicly subsidizing the higher cost and risk of serving poor students at public institutions or, two, we are going to have to accept transferring that cost and risk to already economically vulnerable individual students through high tuition/high loan models. No amount of accountability measures can address that basic tension: poor students exist, they have to go somewhere and it is going to cost somebody.