View profile

To be free or not to be free

Brendan Cantwell
Brendan Cantwell
Inspired by debates about free college, I take some time this issue to think about what kind of higher education system are possible, and what makes higher education possible. My overall argument is that societies need to determine what they want from higher education - and that’s a messy processes - before settling on a question like should it be free, which is really a question of who should pay and for how many people.

To be free or not to be free
This week the debate over free college in the United States was reignited in my corner of Twitter. I’m not going to get into that. Nor am I going to take a position of free college here. If you want to know my position, it will cost you $35. (I can’t tell if that’s an amusing or obnoxious joke.)
Rather than get into the question directly, I want to ramble some abut the underlying questions. Most of the discussion will be abstract. It might be frustrating for some people, and others might think I’m setting up a bunch of straw people to avoid actually taking a position. But I honest to goodness think this is a useful way to approach the problem.
A numbers game.
Let’s play a round numbers game to get us started. I wasn’t sure if the joke I made above was obnoxious, but I am positive the example I am about to give is.
Harvard’s 2020 operating budget was $5.4 billion, which paid for all kinds of things beyond education as we typically think about it. For the sake of an example, let’s divide that total operating budget by the number of students (undergraduate and graduate) enrolled, which is about 24,000. We are talking about a university that operates on a budget of about $220,800 per student. How does that figure play out on a national scale? With roughly 16.6 million higher education students in the United States, $220,800 x 16,600,000 = $3,665,280,000,000. Providing everyone currently enrolled in the US with a Harvard education would cost $3.37 trillion dollars. If US GDP is about $22.7 trillion, we are talking about 15% of the economy dedicated to Harvard for all. Of course, this is a silly example to make a dramatic point. No society is going to provide a $220,000 education per student at scale. Let’s look at more realistic but still big numbers.
In 2017-18 the US spent about $604 billion on higher education. According to Preston Cooper, a conservative education analyst, the US spent about $27,000 per higher education students, a figure that is the second highest in the world. Second only to Luxembourg which really doesn’t cont (sorry Luxembourg). Just about everyone already thinks the US higher education costs too much, so we are not going to get anywhere close to that number.
We face tradeoffs. Societies should decide what kind of higher education system they want, and how they want to pay for it. I think the question works in that order. First, what kind of system do you want? Second, how do you want to pay for it? You might have to revisit the first question once you start the work out the second question.
The Trilemma.
One way to think about the type of system a society wants to to lay out the tradeoffs as “trilemma” as Ben Ansel did in this 2008 article.
The basic idea is that policy goals related to higher education come with trade-offs and the result is that the most popular choice set is impossible. Assume three policy choices on a continuum:
  1. Enrollment, low - high (the share of the population with access to higher education).
  2. Subsidy, low - high (the share of the cost of education paid for using public funds).
  3. Public costs, contained - un-contained (limits on the total public obligations to higher education).
A quick note here on a topic that I know some people find obnoxious (obnoxious is the theme of the week): Price vs. cost. Price is what a student pays. Cost is how any resources it takes to provide an education. These differences matter when thinking about different types of systems.
Let’s consider some basic system types:
The Plutocratic system: Low enrollment, high subsidy, contained public costs: Possible.
In this sort of system social elites tend to dominate higher education, which has restrictive access. Public funding levels are high but also highly regressive.
The Luxary System: Low enrollment, low subsidy, contained public costs: Possible.
This system is for the few who pay. Demand for higher education might be low, as are public outlays.
The whatever it takes public system: The High enrollment, high subsidy, un-contained public costs: Possible, within limits.
In this sort of system, just about everyone has access to higher education, and the price to individuals is low because of generous public subsidies. Costs increase monotonically with enrollments, and higher education will make a larger and larger claim on the public budget over time.
The whatever it takes private system: High enrollment, low subsidy, contained public costs: Possible.
In this sort of system, access is abundant but subsidies are low. Students from high-income backgrounds pay out of pocket, those from low income backgrounds either borrow heavily to participate or seek out low-cost providers
The dream system: High enrollment, high subsidy, contained public costs: Impossible.
This is the dream, high enrollment, prices low for students, and it doesn’t burden the tax payers much. The only problem is that it’s not possible.
The trilemma is too simple.
Ansel’s trilemma model assumes clear tradeoffs made by government. I find it useful as a foundation for thinking about systems but too simple for the real world.
It is not 100% obvious that the tradeoffs are always zero-sum. Here are some other possibilities.
Targeted subsides: Subsidies don’t need to be universally high or universally low. Lots of systems try do make subsides targeted. The idea would be to direct aid (or a subsidy) to those who need it and not to those who don’t. Theoretically, this would allow systems to expand enrollment while at least partially containing public costs. The challenges of this approach are (a it may be politically unpopular among the upper and middle classes who often have lots of influence over government, (b it might be technically quite difficult to direct subsidies efficiently and effectively, and (c if the cost of providing education increases faster than incomes, either the subsidies will have to grow or process will grow, eroding subsidy value.
Lower the cost of providing an education: The trilemma idea assumes that costs are essentially static and grow only with enrollment growth. There are good reasons for this assumption. Major theories of college cost predict rising costs of production, and that seems to be the observed reality in most countries. But it could be possible to lower costs by efficiency gains (scaling cost saving technology, for example) or by reducing the scope of higher education (that is by doing less). The question, I suppose, is how bare-bones can higher education get before it is no longer higher education?
Stratifying the system into higher and lower cost segments: Another possibility that we see in the real world is system stratification. Planned stratification is pretty common. Consider the California Master Plan, that segmented public higher education into three systems: the high cost University of California, the middle-cost California State University System, and the lower-cost California Community College system. Initially, each of the tree systems were subsidized at a high rate, so that student prices were universally low. But not everyone count access the crown jewel University of California, which is of course stratified itself. A big question here is who gets access to the high resource strata of a system? Unplanned stratification is even tricker to deal with. Both competition and the self-organization of higher education result in a tendency towards system stratification even when it is not planned. The consequences can be pernicious.
The government doesn’t alone determine enrollment levels.
It is true that in some systems the government set’s the number of “places” available in a national higher education system. But that number often reflects social demand. And, anyway, in many systems the government does not set enrollment capacity. Since the 1970s, higher education enrollments have skyrocketed. Check out change in the ratio of the number of higher education students enrolled worldwide, over the size of the global population in the 18 - 24 age cohort. The image below is form the World Bank. I can’t get over it.
Global Gross Tertiary Education Enrollment, 1970 - 2020. Source: World Bank.
Global Gross Tertiary Education Enrollment, 1970 - 2020. Source: World Bank.
There are a lot of reasons why higher education is more and more common around the world. One is the transition from agricultural and manufacturing based economies to economies based on services. But the economic explanation probably can’t tell the story alone. A set of papers on the topic identify social forces and cultural change as the prime drives of enrollment. I was part of a project group that produced an extensive treatment on the topic. Anyway, the point here is that probably governments respond to demand for higher education at least as much as they set enrollment levels, as assumed by the trilemma model.
Why does this matter? I think it matters because it partly flips the assumption about why people participate in higher education, or in the American parlance, go to college. I have detected a general operating assumption that college is hoisted on people. The idea: people are forced to go to college because college degrees are required for jobs. I think that’s part of it. I also agree with the idea that in mass systems of higher eduction, going to college becomes normal and expected, maybe even a social obligation. And if you ask people why they go to college they will often say to get a good job. Still, I beleive individuals, families, and communities want college. They like it. Why they like is somewhat less important for this exercise than that they do. People want higher education. That’s my assumption.
Reframing the trilemma.
While I don’t think we need to think about a one to one tradeoff framework that is entirely the responsibility of government, there are clearly tricky questions and constraints when it comes to designing a higher educator system. Simultaneously (a satisfying demand for higher education (providing access), (b while controlling costs, and (c achieving an equitable system is a sticky problem.
I want to dwell for a moment on part (c, achieving an equatable system. What does this mean? I can’t answer the question difinitavely and I won’t even try. I think I am safe to say that the way people have generally thought about a system being equitable falls along two dimensions.
The dimension first is the access to what question. Let’s assume a system is segmented (or stratified) into cost strata. In this hypothetical example we have a high resource segment and a low resource segment. How do we decide which students get sorted into the high resource and which students get sorted into the low resource segment? People often like the idea of meritocracy. Those students with the most natural ability and who work the hardest are selected to the high resource segment, and the rest are absorbed by low resource segment. The problem with this idea is that academic merit - just about everywhere in the world - is tied up in social conditions. Privileged groups (by race, ethnicity, language, class, gender, etc) have access to more and often better educational resources and the measures of merit are normed to privileged groups. The result is social reproduction. Sociology 101.
In a market based system price could be the mechanism for sorting students. The price for students would be high in the resource intensive sector and less in the lower resource segment. A bare bones higher education could be offered at a low price (even at 0) to students, for example, where as a more elaborate higher came with a higher price-tag. Students with access to capital or who were willing to borrow to access cold chose spend their money (today or in the future) on the more expensive higher education. The problem for this, of curse, is that most people think it is unfair. Students who come from which families could easily afford the resource intensive experience, and those who cannot afford it: let them eat cake. That’s why targeted subsides are a popular way to deal with this challenge, even if it is not always an effective way to deal with it.
The second dimension is the ‘what is there to access’ question. That’s a mouthful. This question is about how providers (institutions of higher education) are segmented. It could be by resource intensity, it could be by mission, it could be by scope or scale, or by control (public vs. private). For example, in northern Europe a common divide is a binary system with academic universities making one segment and polytechnics or universiteis of applied science making the other. But the segmentation could also be about a teaching focus versus a research focus, like liberal arts colleges and research universities in the United States. There are innumerable permutations and nobody, including me, cares for me to list as many as I can think up. But you could play along: what ways could we segment a system of higher education that is not explicitly about resources?
Once a society settles on segmentation, that includes the sorts of things that higher education should do, then it can take up the question of who should pay. For example, maybe we want a lot of research from higher educaiotn. Should students pay for research? Maybe we want higher education to come bundled with consumption activities like pop-concerns and food-courts. Should the public pay for that? Who should pay for the education itself. Does it matter if the education is vocationally oriented or not?
So what do ya want?
I basically think the free college question - at least in the US context - is too often asked without first articulating what type of system we want to have. Of course, what I want, or what you want, doesn’t mean much. What matters is what society collectively wants, and determining that is political and will involve struggle.
Another thing. Most system designers don’t have the benefit of unlimited resources or a blank slate. The US higher education system, for example, is highly institutionalized, elaborate, and socially entrenched. It’s not easy to move it rapidly. Total re-design is probably not feasible. What if you didn’t want a private sector, for example. Just public higher education. There are good examples of this around the world. But it would be very difficult to achieve in the US without massive dislocation. It would be the kind of political power grab, at least in my estimation, that is so dangerous that the ends cannot possibly justify the means.
Not everything is really on the table because the table is already set. The most feasible goal is probably to re-arrange the table, given the current setting. The possible range of table settings is always wider the further out in time we think. But today is today. And most people want to eat sooner than later. (What an obnoxious metaphor.)
The conversation - and the concurrent political processes - about about what we want from higher eduction is best grounded, in my view, in a sense of human purpose. Rather than assuming that the goal is to maximize utility or to produce the most efficient system that is possible - we are not trying to hit the frontier of an indifference curve. I propose we ask twin questions:
  1. What about our society do we wish higher education to reflect?
  2. What do we hope higher education will bring about in our society?
As the political philosopher Danielle Allen has argued, social institutions like higher education may best be constructed through purposiveness. In a recent volume, Allen describes it this way:
Purposiveness incorporates rational self-interest, but it’s a bigger concept. Human purposiveness describes the effort of human beings to ascertain their own best path toward flourishing.
I am not a pollyanna. Determining the best path is not an easy to do. Who determines what, and for who? What do we do about conflicts? Power makes the whole thing tricky. Power and oppression. It’s hard to get things done in ways that move the needle and don’t leave people behind, or crush them all together.
Even so, I say we start by ascertaining what type of higher education system we want for the kind of society we have and the kind of society we want to have in the future. Then we get into the question of free college.
And now for something more empirical
This is pretty long already, but I want to turn our attention for a few moments to a little taste of empiricism. (And I am mixing metaphors again. Obnoxious is the theme of the issue.)
What drives costs in higher education, or, why does higher education cost so much? The basic answer is labor. It takes labor to do stuff. The more stuff you do in higher education, the more labor it takes. The more it’s gonna cost.
Let’s start by taking a look at OECD data on per student expenses. The below figure is for 2017 and it shows that the U.S. spends just over $32,000 (lower panel, tertiary educaiton) per student when you count all expenses, and something closer to $27,000 if you count just core expense, which is the figure I cite above. This is more than most countries. But other countries, including Austria, Belgium, Australia, the Netherlands, Norway, Canada, Sweden, the United Kingdom, USA, and Luxembourg, spend $20,000 or more per student per year. And remember, Luxembourg doesn’t count (sorry Luxembourg). Not all of these countries finance it the same way. Some, like Norway, achieve this spending nearly entirely through public funds. Some, like the UK, rely mostly on private funds (although that’s a bit complicated because it’s mostly government provided student loans, lots of which are never repaid). Some countries spend more on research relative to core services. These differences all reflect history, available resoruces, political choices, and the cultural preferences and expectations that people have for higher education.
Source: OECD Education at a Glance, 2020, Indicator C1.
Source: OECD Education at a Glance, 2020, Indicator C1.
Now let’s take a peek at how that money is spent. As you can see, most OECD countries devote well over half of all their higher education (tertiary, lower panel) on labor. The OECD average is nearly 70%. Among OECD countries, Chile, at just under 50%, spends the least on labor. Every other country shown directs at least half of all higher education expenses to compensate staff.
Source: OECD Education at a Glance, 2020, Indicator C6.
Source: OECD Education at a Glance, 2020, Indicator C6.
The share of expense spent on labor helps to clarify what choices system designers have and what choices they don’t have. Assumption alert: I don’t think we will see automation and technology reduce the need for labor in higher, at least not much. Other’s disagree, I know. But I just don’t think the evidence supports it. Higher educator is labor intensive. It will remain so. To do more things, you probably need more people, which cost more. Let me show this another way.
Here’s a simple analysis from data I collected for a different project. The sample I’m using includes 109 research universities form the US, UK, Canada, and Australia. I collected information on university’s total operating budget in 2013 and denominated it in thousands of 2013 US dollars. I also collected other variables, including enrollment, and the number of academic and non-academic staff (as well as other stuff, but that’s not for now). Take a look at the descriptive statistics. You will notice quite a lot of variation.
Brendan's sample of Anglophone research universities from 2013
Brendan's sample of Anglophone research universities from 2013
Now here are the results of an OLS estimation for total operating expenses. You see that both academic staff and non-academic staff were associated with a larger operating budget, but the number of students enrolled was not. An increase in one academic staff member was associated with an increase in operating expenses by about $225,000 and about $236,000 for non-academic staff. Of course, salaries were not that high, but staff need offices and equipment and are involved in activities with other expenses and so forth. The more people, the more activities, the more money it takes. But that’s not the case for students, at least not in this model. There are marginal costs to enrolling additional, but much of the costs are fixed and students come with revenue, through tuition fees and public finding. So the model found no significant relationship.
OLS estimation of Brendan's data
OLS estimation of Brendan's data
Finally, let’s have a look at the estimated margins means over a range of staffing levels. Increasing academic staff by 500 was associated with a $112.6 million increase in operating expenses. The slope was even steeper for non-academic staff; 500 additional non-academic staff was associated with an a $118.2 million increase in the operating budget.
This example is just for research universities, which are expensive, in high wage countries. But the point is illustrative of the general relationship. System designers don’t really have the choice of having to pay people. Who much people are paid is, of course, partly a discretionary question. You can also decide what sort of people are involved (by deciding what kind of activities you want). Teaching staff, research staff, administration or outreach staff are all necessary for different activities.
When taking about what higher education should do, we are taking about what the people who work in the sector do. How many of these people there are. How much money do they make? How many resources do they have to do their jobs?
Higher education is expensive to provide. And even when the individuals cover a large share of the costs through tuition payments, provision is a social activity and reflects collective choices. Who makes those choices and whether the choices are made actively, and consciously, or passively or by default is something to consider.
The people in higher education should help chart its future.
Given the link between people and activities, the fact that all system reforms more or less take place within existing systems, and following the principal of purposiveness, I think the people involved (those inside higher education already) should play a role in deterring the systems future. That sentence was a mouthful. By calling for a role for academic people in the discussion about what type of system of higher education a society wants I am not excluding others from the conversation. Government, civil society groups, families, and businesses all have an interest in the sector, and in society at large, and should all have a voice. The reform conversation often focuses on what outsiders should do and sees higher education faculty and staff as an obstical to change and improvement. I understand why that is. We can seem entrenched and self-interested. But the conversations will be more productive when they are more collaborative.
Ok, enough.
A closing line or two.
Well, I started this issue saying I would talk about free college. I lured you in with with talk about free college and then switched to some philosophical ramblings and then a little empirical analysis, closed out by a conclusion far too sweeping to be justified by the evidence alone. I hope you enjoyed it!
I am going to take a couple of weeks off to prepare for the academic year. I hope to be back in late August or early September.
In the meantime please consider subscribing so that you know when I put out another issue. Please share this post in your networks and do let me know if you have feedback.
Edited on August 2nd, 2021
Did you enjoy this issue? Yes No
Brendan Cantwell
Brendan Cantwell @@cant_b

Associate Professor @HALEatMSU and Joint Editor-in-Chief for Higher Education ( Speak only for my self.

In order to unsubscribe, click here.
If you were forwarded this newsletter and you like it, you can subscribe here.
Created with Revue by Twitter.