The Almost Ideal University, Part I

Tyler Cowen recently wrote about “his” university:

If you were to design a university from scratch, what might it look like? The idea isn’t necessarily to have a model for other schools to follow, but rather an experiment. Assume that various legal, contractual and accreditation constraints do not stand in your way.

https://www.bloombergquint.com/gadfly/what-would-your-fantasy-university-look-like

In this post, which may well end up being a fairly lengthy one, I will outline what “my” university will look like. This is a place I would want to work at, would want to be involved in the administration of, and would want many (but not all) students to study at. That’s what I mean by “my university”.


First things first: my idea is to have a model for other schools to follow. There are many interesting models of what an institution of learning should look like, but the problem with almost all of them is that they don’t scale. And the ones that do scale aren’t interesting models of what an institution of learning should look like. You want to be at the sweet spot between the two extremes.

It’s actually a little more complicated than that. You don’t just want to be at the sweet spot between these two dimensions, but you also want to have optimality along two other dimensions: affordability and replicability.

My ideal university, in other words, should optimize for a sweet spot that acknowledges the trade-offs between quality, scale, affordability and replicability.

All students who are in this university should get education of high quality at a price that is reasonable for the kind of education they are getting. All models of dispensing education in this university should be documented and freely available in the public domain.

When I say:

  • All students: I’m talking about scale. No student should be turned away because they were not able to solve x questions in y minutes on a particular day in the year. That’s how entrance examinations work, and the reason we need entrance examinations is because we can’t scale well enough. Patting ourselves on the back for getting the “best” students is our way of disguising the fact that we are unable to provide quality education to many more students.
    ..
    ..
  • High quality: I’m talking about quality. It is difficult to define and dangerous to measure, but you know it when you see it. Quality is not A+ on the NAAC, because nobody who has been through a NAAC process can put their hand on their heart and say that it is a good way to measure quality. Neither is quality the number of students who score more than x% in an examination, because ditto.
    Here’s a definition of quality that I am comfortable with: are you, at the end of your education, able to readily apply concepts you have learnt in order to be productive in your workplace?
    If you want to be educated in order to be a professor, did we teach you well enough for you to be able to teach right away? If you want to be educated in order to be a data analyst, did we teach you well enough for you to be able to work on a project right away? And so on, but that’s the point of a quality education – can you put what you claim to have learnt to good use?
    More: if your job requires you to acquire a skill we didn’t equip you with, did we teach you how to teach yourself? A quality education isn’t just about learning, it also ought to be about learning how to learn.
    Each of us has our own way of learning, of course. Some do better by listening, some by reading, some by visualization, some by introspection. A student who graduates from a university must have the ability to understand what she lacks in terms of skill sets, and have the ability to equip herself with that skill by using resources online.
    That’s quality.
    ..
    ..
  • A price that is reasonable: Tyler Cowen urges us to not worry about constraints, and I think I understand where he is coming from when he makes the request. Figure out the best you can build, and we’ll solve the constraints as we go along. But at least in India, one of the reasons higher education is in such a mess is precisely because we haven’t used the price mechanism effectively enough. Of the four, this dimension is perhaps the trickiest to think about.
    I’m a big believer in the fact that students should have skin in the game, and therefore I think that a price should be paid for acquiring an education. But I’m also all too aware of the fact that some students simply cannot pay, and therefore think that some amount of subsidization is inevitable.
    It gets trickier still, because you will almost certainly have to spend more resources on those students who will need subsidization. They are, other things held constant, likelier to need more intensive training in getting the quality of their writing up to the same level as that of other students, simply because they are likelier to not have had the same exposure to quality education in school. And this will apply to other dimensions as well: quantitative skills, the luxury of having time to practice their skills and so on.
    Cross-subsidization? Vouchers from the government? Income sharing agreements? My personal preference would be for the last of these, but I’ll happily admit to being uncertain about what the correct answer is.
    ..
    ..
  • Replicability: The last sentence in my statement above is about replicability. It doesn’t matter if your approach works or not, where replicability is concerned. Funding/regulatory approval for your university ought to be contingent on it being easily replicable. Your methods, your ideas and your processes must be open source. Why? Because an educated populace is the whole point of education! At a system-wide level, the opportunity cost of protecting the trade secret of an well-run educational institution is simply too high.
    Documenting the ideas, their implementation, the challenges encountered during implementation, the refinement, the impact evaluation and the evolution of the university – and that on an ongoing basis, needs to happen. And this should be publicly available, for reasons stated above.
    And it isn’t even that big a risk, because the secret to a well run university isn’t the ideas for it – it lies in their implementation, and therefore in human capital. Spread the knowledge of how to build good universities far and wide!

Now, about the trade-offs between these four dimensions. No university, anywhere in the world, can optimize all four (scale, quality, affordability and replicability) at the same time. The first two alone will inevitably involve trade-offs.

What would my ideal university optimize for, and what would I be ok sacrificing, at the margin?

The SQAR framework

I will sacrifice scale at the altar of quality, and I will also sacrifice affordability at the altar of quality. Quality and replicability are non-negotiable in my worldview.


Why sacrifice scale?

Because if I have to choose between scale and quality, I’ll choose quality every day of the week. A job well done is preferable to lots of jobs not-so-well done. Lot of jobs really well done is great in theory, but it tends to not work out in practice, and especially so when it comes to education.

Why sacrifice affordability?

Because if I define quality as the ability to apply what you’ve learnt, a graduate from my university stands a better than even chance of being productive, and therefore employable. And that means a better than even chance of income sharing agreements working out in practice. And so yes, education in my university may not be cheap, but you can always pay later, out of your future income streams. And my university has skin in the game too! No income stream, no income sharing, and my university has taught you for free. We have taught you badly, since you aren’t able to generate income, and so we don’t deserve to be paid. That’s it.


What might students actually do in my university, and how will it actually work? I’ll get to this in Monday’s post.