Massive online learning and the unbundling of undergraduate education

Is free online learning going to put traditional colleges out of business? For a college professor, it’s hard to read the news these days and not worry about this threat. For one thing, the businesses of journalism, publishing and music, all decimated by the internet, bear uncomfortable similarity to that of higher education. In each case, valuable intellectual property is threatened by free online copying and distribution.

For another, the industries of health care, higher education and (formerly) housing all share another uncomfortable similarity: in an era of stagnant household incomes, their costs are inexorably rising, leading to ever-increasing indebtedness. Simple arithmetic suggests that this can’t go on forever. As long as the cost of college keeps rising, the pressure to find a free, Napster-like alternative is going to keep building.

Of course, I’m convinced that higher education is a valuable product worth paying for. But I believe that these two forces—the internet and rising costs—are going to fundamentally change what most college professors provide their undergraduate students. (For now, I’m going to set aside the questions of research and graduate education.)

The basic way that these forces are going to change undergraduate education is by forcing an unbundling, or disaggregation, of the goods that colleges have traditionally provided their students in one big bundle. This unbundling will happen in three ways: for the whole college education, for the individual course, and for the way that college is paid for.

1. Unbundling the college education

A college education has traditionally bundled several different kinds of goods together:

  1. The curriculum: mastery of specific knowledge and development of more general reasoning, analytical, and communication skills.
  2. The extra-curriculum: a network of friends and contacts, and experience gained from clubs, sports, internships and other activities.
  3. The signaling process: validation of general talent or status by completing all of the above at a “better” or highly ranked college.
  4. The college experience: everything that is personally interesting, enjoyable or rewarding about living in a certain place with certain people, and having experiences that are personally valuable to the college student, regardless of their value to anyone else or to society at large—everything from late-night conversations about the meaning of life, to road trips, to pranks, sports rivalries, and “school spirit.”

Traditionally, colleges provided all of these goods in a bundle, simply because the best way to provide them was to expensively gather a lot of students, faculty and resources in one place for several years at a time. But now, with the internet, is the logic of bundling starting to break down?

I think it’s immediately apparent that the first type of goods—the curriculum—is by far the most vulnerable to disruption from the internet. Highly self-motivated students (i.e., Abraham Lincoln) have always been able to teach themselves, given the resources, and the internet is simply going to accelerate and expand this opportunity to anyone in the world who has an internet connection. This is where the disruption of higher education is going to parallel that of journalism, publishing and music.

My hunch, however, is that the second, third and fourth types of goods are going to be affected very differently. For these, there is simply no substitute for being in the right place with the right people. It’s the same reason why places like New York and Silicon Valley become ever more expensive to live: in certain cases, you just have to meet people in person. So, to the extent that colleges provide 2, 3, and 4, they are going to continue to thrive, regardless of what happens with 1. Ambitious students are always going to want to be in the right place.

The catch is that only a small minority of colleges—the “top” colleges—really do a good job at providing 2, 3, and 4. The vast majority of colleges mainly provide 1, and these colleges—i.e., most of them, the “typical” colleges—are most vulnerable to disruption and transformation. For a long time, the typical American college has been preoccupied with “rising in the ranks” by spending a lot on facilities and faculty, and trying to attract better students, in hopes of providing more of 2, 3, and 4 in the manner of the top colleges.

I think this pursuit is going to be abruptly curtailed by ever-rising costs and indebtedness, and by the finite nature of these resources. There are only so many top students and top faculty to go around. Only ten colleges can ever be in the top ten. For the typical college, a wiser strategy is to focus on providing 1. at a better value, embracing the disruptions and transformations of the internet. But how exactly is this disruption going to work?

2. Unbundling the invididual college course

A college course can be unbundled into two separate functions: first, there is the transfer of knowledge to the student, and second, there is the evaluation and feedback of student work. Knowledge transfer consists of all the information that a course provides to a student, which may be delivered in any medium: through assigned readings, lectures, websites, videos or otherwise. Evaluation and feedback consists of seeing what students can do, and telling them how they can do it better, whether through exams, papers, projects, or otherwise.

The internet’s transformation of knowledge transfer is rapid, vast, and well underway.  Its transformation of feedback and evaluation, however, is much less certain and proceeding more slowly.  The sudden, enormous and rapidly increasing popularity of massive open online courses, or MOOCs, has justifiably gotten a lot of attention.  Much of this attention, however, fails to distinguish between these two aspects of a college course.

With MOOCs, now anyone in the world with an internet connection can download and watch lectures from eminent experts at top universities, for free, and hundreds of thousands have done so. This is indeed a huge leap forward in the area of knowledge transfer.  But the equivalent leap in the area of evaluation and feedback has not yet taken place.

Professsors who teach MOOCs have experimented with different ways of evaluating the work of the hundreds of thousands of students enrolled in them. Peer evaluation, where students review each other’s work, and automated, computer evaluation of student work have both been tried. Both of these methods surely have some value. They will probably be most effective at evaluating lower-level work, such as checking grammar, syntax or arithmetic. They will be less and less effective for student work at more advanced levels.

For now, at least, attentive, expert evaluation of student work cannot be automated or scaled up using the internet. It is going to remain a somewhat expensive, one-on-one job, and it will be worth paying for because the results of these evaluations provide useful information about and for the students who go through them. Certification agencies, independent of universities, may emerge to examine and evaluate all comers, a la the AP exams for high school students. But the same logic applies. Relatively many people can evaluate a final exam for freshman art history 101; fewer people, mostly professors, can competently evaluate a senior thesis on Picasso and Braque’s development of cubism.

Actually, the wide availability of free MOOCs has every chance of changing professors’ teaching jobs for the better. If we professors can assign students to watch video lectures that cover the basics of the course material, that frees us from the burden of having to go over the same basic topics in every course, and allows us to spend precious class time in a way tailored to the needs of the particular students there—building up their weak points, or going farther with their strong points. Also, if automated programs get better and better at evaluating students’ low-level work (grammar and arithmetic), that leaves us more time to evaluate the higher levels of their work, which is a better use of our time anyway.

What’s more, this unbundling isn’t starting from zero. Most college courses are already partially unbundled in different ways. Every course that has a professor do the lecturing, while teaching assistants provide the evaluation and feedback, is already unbundled in that sense. The availability of MOOCs may transform the professor’s job, but not the TAs’. Likewise, every course where the professor assigns a textbook is already partially unbundled. That same professor can now assign his or her students to watch a MOOC video lecture along with reading the textbook. It is just one more resource available for the students. The professor’s job is still to bring home the course material to the specific group of students in the way that best suits them.

Once the hoopla dies down, it will become apparent that free online lectures provide basically the same thing that textbooks have always provided—focused, directed knowledge from a recognized expert—albeit in a cheaper, more accessible form. Much of the true value of a college course is in the opportunity for intensive tutoring and feedback (whose Platonic ideal might be the tutorial system at Oxford), which will remain scarce and valuable. College courses that mainly concern themselves with knowledge transfer will be disrupted, as they should be, but college courses that provide excellent one-on-one feedback and evaluation will retain their value.

3. The unbundling of college finance

The recent news that total student loan debt in America reached a trillion dollars—that’s $1,000,000,000,000—with a median balance of $12,800 per indebted graduate, focused attention on the issue of student loans. But most discussion of the issue has relentlessly conflated two separate questions: First, what is it worth to pay for college? And second, what is it worth to borrow for college?

There’s no limit at all to what it is worth to pay for college, if the money is available. If students, parents, college aid offices, and scholarship foundations think it’s worth paying a million dollars for a college degree, then nothing will stop them, and there’s nothing wrong with that. If we imagine that ten spots in the Harvard freshman class were to be auctioned off to the highest bidder, in secret, so that no one would know who had paid for their places, we can easily imagine those places going for $1 million or more. If we imagine that a certain course of study is going to lead its graduates to cure cancer, malaria, and tuberculosis, we can easily imagine the government paying millions to fund those degrees.

But what is it worth to borrow for college? That is a completely different question. A moment’s reflection should lead to the conclusion that the only justifiable amount to borrow for college, is the amount that you can expect to afford to repay from your future earnings, and not a penny more.

Recently, the typical starting salary of a college graduate was about $27,000. Depending on assumptions, one rule of thumb is that at the absolute maximum, one’s total loan debt should be no greater than one’s starting salary. Armed with only this information, it defies all reason to argue that the government or a bank should lend much more than this, or that a student should take on much more debt than this, to acquire a typical college degree. And yet, many thousands of students are in exactly this situation. This is the state of affairs that exactly resembles the recent housing bubble, and which can only end in misery.

But of course, we know that some degrees carry expectations of much higher earnings—for example, graduate degrees in medicine, nursing or engineering, not to mention certification as an expert machinist or plumber. Without question, students should be able to borrow much more to acquire these degrees and certificates, corresponding to the expected earnings of each field. The catch is that only some fields are in this situation.What percentage of all degrees and certifications carry expectations of higher-than-average earnings? I don’t know, but probably not the majority.

Given the high cost of college, given that college is subsidized by taxpayers, and given that student loans can become an immense burden on the graduates who may have signed on to an obligation of many thousands of dollars at age eighteen, with little adult experience to guide them in the wisdom of that decision, it is a scandal if the expectation of future earnings does not guide every single decision to borrow or lend money for college.

Don’t, however, lower-paying fields of study have intrinsic value not captured by the future earnings of their graduates? As an art historian, I wholeheartedly agree with this idea. But the only appropriate way to acknowledge this value is for the government, scholarship foundations, college aid offices, parents and students to write checks to pay for this study—not to mortgage the future of young students by making them borrow for it. If art history, religious studies, creative writing, etc., are valuable to individuals and society—and indeed they are—then individuals and society need to pay for them, not borrow for them.

The logic of expected earnings and indebtedness will also help drive the unbundling of college goods discussed in the first section above. If we think about what raises the expected earnings of a college graduate, it must be some combination of the goods 1, 2, and 3: the curriculum, the extracurriculum, and the signaling.

What about category 4, the college experience? It probably has little if any effect on future earnings; it is immensely appealing to students and their parents; and it is quite expensive to provide. Over the long term, this probably means that it will be the province of colleges that attract upper-income students whose parents will pay for the experience, while loans and government subsidies go to colleges that focus on categories 1, 2, and 3.

We could describe this process as the unbundling of college as consumption (category 4), from college as human capital investment (categories 1, 2, and 3).  College-as-consumption consists of all the experiences that are valuable and enjoyable to the college student, whereas college-as-investment consists of all the knowledge and skills that make the college student more valuable to society through his or her work. Of course, many times the same experiences that are valuable for a college student to consume are the same ones that build his or her human capital — but not always. Learning a foreign language in a study abroad program is probably both consumption and investment. Eating top-notch food in a top-notch dorm is probably just consumption.

There is one reason, however, why the unbundling of consumption and investment might proceed differently, or more slowly, than the unbundling of knowledge transfer and student evaluation. The reason is that whereas the internet is forcing the unbundling of the college course, it is not doing the same for consumption and investment. And whereas the combination of high costs and indebtedness might force typical colleges to devote more of their resources to providing value in the form of return on students’ human-capital investment, it might not affect top colleges in the same way.

On the contrary, top colleges might be able to continue to use their vast resources (in the words of Kevin Carey: wealth, prestige, and exclusivity) to provide an expensive, valuable college experience (category 4) that helps attract top students, high-paying students, top faculty, and donations, in a self-reinforcing cycle as the rich get richer. This self-reinforcing cycle might very well increase the stratification among colleges, as fewer and fewer colleges are able to “play the game” and attract scarce “stars” among students and faculty. Over recent years, one could argue that this has already been happening, as top public universities in the U.S. are less and less able to compete with top private universities.  These top universities provide their students with bundle that includes both a lot of “college experience” to consume, and a high return on their human-capital investment.

But for the vast majority of colleges and students, the unbundling process is probably going to lead to pressure to devote more resources to providing return on human-capital investment, rather than on consumption.


Thanks very much to Tyler Cowen, Reihan Salam, Arnold Kling, CAA NewsLarry GrossStephen Downes, Sui Fai Jon Mak, Dan Reed, and James Oswald for pointing their readers in this direction. On July 19th, I went on KPFA Truthdig Radio to discuss these issues (MP3 audio of the segment is here).