Bryan Caplan  

The Future of Online Education: Three Competing Perspectives

Up From Poverty... The Status-Good Model of Colle...
Amazon put Borders out of business.  Is online education going to do to the same to brick-and-mortar colleges?*  Reflecting on earlier conversations with Arnold, I've realized that there are three competing perspectives with three competing predictions.

Perspective #1: Human capital model. 

Analysis: The point of college is to teach marketable skills.  Online education will soon be able to teach marketable skills as effectively as brick-and-mortar schools at a tiny fraction of the cost. 

Prediction: Online education will soon have roughly the same wage premium as brick-and-mortar colleges, and rapidly drive these high-cost dinosaurs into bankruptcy. 

Perspective #2: Status good model. 

Analysis: Online education will soon be a great way to teach marketable skills.  But colleges are primarily places where young elites (and their tuition-paying parents!) bond.  In Arnold's words:
[G]oing to a top college today is like belonging to the right church in 1850 or the right country club in 1950. When you are supplying a status good, ostentatiously wasting money on buildings can increase demand.
Prediction: Brick-and-mortar colleges are here to stay.  However, online education will easily compete for the segment of students who only want to acquire marketable skills.  Students who opt for online education will earn a wage premium comparable to that of brick-and-mortar grads.

Perspective #3: Signaling model.

Analysis: Brick-and-mortar colleges are primarily places where students signal a combination of intelligence, conscientiousness, and conformity.  Online education suffers from a severe adverse selection problem, because the students most eager to avoid traditional education tend to be deficient in one or more of these traits - especially conformity to the established social norm that young people should go to a traditional college.

Prediction: Brick-and-mortar colleges are here to stay.  Online education may be a niche good, but the labor market will usually penalize its graduates with a low wage premium.

Summary table:


Fate of Brick & Mortar Colleges

Wage Premium for Online Grads

Human Capital



Status Good






I hasten to add that these are three polar cases; I'm happy to admit that each is partly true.  The real question is the weight each perspective deserves.  My rough guess is 20% for human capital, 10% for status good (though more at elite colleges), and 70% signaling.  I encourage critics to provide alternate breakdowns.

My implied prediction: brick-and-mortar colleges will probably experience a slight decline in coming years, and the wage premium for online grads will probably slightly rise.  In the absence of big changes in government policy, however, higher education isn't going to change much.  Old-fashioned colleges will stay in business, and the labor market will continue to heavily favor their graduates.

As always, I'm open to relevant bets.

* When I talk about "online education," I don't just mean students at existing brick-and-mortar colleges taking some classes from their dorm rooms.  I mean students enrolling in virtual colleges instead of physical colleges.

Comments and Sharing

COMMENTS (37 to date)
Ken Plahn writes:

I would go with 40-20-40.

I also think you are leaving out the employer side of the equation. By hiring employees with more targeted knowledge at a lower expense, employers hiring online educated students will have a competitive advantage over those hiring brick and mortar students.

Steve Sailer writes:

How superior is "on-line education" to the "correspondence courses" that have been around for generations? In Evelyn Waugh novels, naive lower middle class dweebs are always taking correspondence courses, while the rich go to Oxford and Cambridge.

Dave writes:

This seems to assume the typical 20-something student base, but I wonder how much that will continue to be true. Brick-and-mortar is very limiting for adults who already have jobs, families, houses, etc. And with a degree needed for practically any job worth having, online programs are ideal for mid-life career switchers. Granted, this falls within #1 which you discount, but I also think the negative signaling of online programs you cite above is less important when the student earns the degree at 40 something.

Curmudgeon writes:

In my own (fairly exotic) country, France, universities are weakened by the socalled "grandes écoles", which enjoy a high reputation under the Status and the Signaling models. Now the thing is that some positions in the public service are legal monopolies for people coming from the grandes écoles, and then under a system of mutual parasitism, some of the top civil servants become managers of big companies. So competition is considerably distorted. I predict the staus quo in favour of brick-and-mortar grandes écoles.

Peter St. Onge writes:

I think the proportion varies a lot by major. An IT or nursing degree seem heavy on marketable skills, a business or MBA seems heavy on the connections made, and humanities may be mostly signalling.

I'd be curious how "legacy" online universities (Phoenix, Houston) breakdown by major. Surely somebody's done that tabulation.

Technology gives us many more choices and alternatives that don't necessarily result in wholesale replacement of institutions, processes, traditions, methodologies, etc. The hand wringing over the "the future of A" and "the future of B" is misplaced energy. Most changes will be incremental and we will integrate those changes...and when an innovation is indeed revolutionary, it is obvious to all of us.

Isaac writes:

One twist to the status good model (and the others to some extent) is that in my experience most universities in the United States are basing thier plans for the future on the assumption that the number of students will continue to increase. If online schools eat away at their potential students, they may find themselves saddled with expensive investments (such as buildings) that don't recoup the cost.

Also, the current status quo of cheap student debt may not last, or alternately may not be extended to online schools. In the latter case, traditional schools would retain a massive edge, in the former case things could go either way.

Hugh writes:

At a certain point in the future, we will start to see a small number of managers and executives that have an "on-line" educational background: they will not be biased against job candidates with a similar education, and will hire them if they fit the job.

At that point the dam breaks, people do the math and go with the low-cost, high quality on-line solution.

Emily writes:

#1 and #2 both accept that online education is/soon will be great for teaching marketable skills. (It's not relevant or #3.) I don't think that's the case. "The point of college is to teach marketable skills. Online education has significant disadvantages in that area compared to brick-and-mortar schools, and will continue to for the near future" is another set of beliefs that would lead to "Endure" predictions.

Mike W writes:

Would the accelerated (one-course-a-month) format of the for-profit universities (National, Phoenix, etc.) be categorized closer to bricks-and-mortar or to on-line?

They would seem to have the main feature of the Human Capital Model in that they concentrate on marketable skills building in a classroom setting but without all the time (and money) spent on socializing as in a traditional campus university. They would also seem to meet the requirements of the Signaling Model in that completion of a degree program tells a potential employer something about the graduate's intelligence, conscientiousness, and conformity.

There would not seem to be any Status Good value from completion of a for-profit university degree but then there is probably not much from most public and private non-profit universities either.

Except for the protection and subsidies provided to traditional public and non-profit universities by governments it would seem the for-profit accelerated university would endure over either of the other formats.

Glen Smith writes:

You do realize that the phrase "marketable skills" would imply some amount of a signal of "intelligence, conscientiousness, and conformity"?

Matt C writes:

Somewhere in this analysis you should talk about brick and mortar colleges killing the golden goose.

Employers aren't actually getting the traits that are supposedly being signalled by brick and mortar graduates. Students, at least some of them, aren't going to get the wage premium they were expecting when they took on a six figure debt load.

Not only is there attractive competition on the online side, but the deal, on both sides, is getting steadily worse on the brick and mortar side.

I think Arnold is right about us hitting a tipping point . . . eventually.

Chris Koresko writes:

Based on these three perspectives I would predict that online colleges will win over the long term, provided that they really do produce as good an education as brick-and-mortar colleges at much lower cost.

The argument is that the human-capital effect is a stable one, in the sense that provided the requirements for human capital and the technologies that produce it don't change much, the relative value of online vs brick-and-mortar education will not change.

In contrast, the effects cited as protecting brick-and-mortar colleges in the status and signaling models are unstable: their strength today depends on their strength in the past. To the extent that any educational history conveys status, that status will be a reflection on the perception of past successes of people with similar history: if I went to Harvard, that associates me with the success of famous people who went there before. If a time comes when a critical number of successful people went to an online college with some well-known and identifiable brand, then presumably the status associated with that brand would also rise. That would help it attract more talented students, leading to a positive feedback.

Similarly for the signaling model: if getting your degree online today is a signal of non-conformity, there's no special reason that should be true tomorrow. If employers' experience over time tells them that they can identify a class of workers with solid skills and character who are carrying much less than the usual debt (and therefore may demand lower wages) then the signaling advantage of the brick-and-mortar colleges will be diluted. That process could feed on itself: the success of employers who hire the online workers will enhance the reputation (signaling value) of online education, again producing positive feedback.

Bill Nichols writes:

I think you are ignoring the importance of self-regulation and discipline in online learning. Provided the summative learning assessments are rigorous, online learning success should be an excellent signal for highly motivated, highly disciplined individuals who can excel in a 21st century knowledge economy. If you want someone who can take on a task and be trusted to plan it and carry through, pick people who have demonstrated the capability.

Dan Carroll writes:

I tend to side with the status/signaling model, especially (primarily) for elite schools. However, I'm not sure I agree with the assumption that online schools can't signal premium or superior traits. Self discipline, iniative, better learning materials, better lectures, all signal value vs partying, class skipping, and dumbed down curriculum at most traditional schools. Just because most online schools are replacements for AA degrees now doesn't mean that will continue; that just means that those degrees were the easiest early targets.

Indeed, traditional schools are vulnerable precisely because their value rests on a myth of perceived human capital. Online education is uniquely positioned to expose that myth.

The primary challenge they face is the networking value, though that is really only important at the elite schools. In addition, traditional education is state sponsored, and educational institutions will battle against any relevant measure of quality, such as standardized testing.

My prediction is that traditional schools will go virtual - leading to consolidation yet preserving their status as gate-keepers. However, it also will create price pressures due to significant excess capacity created by virtualization and the lack of a need for the fancy buildings. Low cost, high quality (or the most subsidized) producers will win. It will take about ten to twenty years.

MingoV writes:

Matt C. wrote:

Employers aren't actually getting the traits that are supposedly being signalled by brick and mortar graduates...

That is true, but unfortunately employers and professional organizations are not pushing for more rigor in undergraduate studies. Instead, they are pushing for more schooling. Certifications that required associate degrees now require bachelor degrees. Professions that required bachelor degrees now require master degrees. But, in some fields advanced degree coursework is no more rigorous than undergraduate coursework. The declining quality of higher education eventually will trigger a collapse, but that may be decades away (especially as more and more grads will be employed by governments that believe our higher education system works well).

Winton Bates writes:

One point that doesn't seem to have been covered is the quality of the degree. If I was setting up an online college I think I would be trying to create a credible signal that graduates would be required to meet higher standards.

Glen Smith writes:

I wonder about the quality of online classes that teach "useless knowledge" like most of my college level language classes were for me. To get decent grades in those classes, I actually needed external discipline and work ethic. For my STEM classes, I needed much less external discipline and work ethic since the opportunity costs of those classes were either very low or I could very easily relate what I was learning to my future paycheck.

Philo writes:

I do not see that being on line must prevent an institution from carrying out the signaling function. Perhaps no online institution currently serves this function, but eventually an online educational entrepreneur may figure out how to do it.

Tex Doc writes:

In response to Philo:
Online and signaling - cull the herd - provide honest and complete assessment and grading of students in the institution - no more grade inflation and external (parental - legal) interference. Period. If you are afraid of true assessment - and bad grades will "hurt the brand" - then your school fails to gain influence.
If the Ivies gain status by their acceptance of the best - the onlines must seek to produce the best by superior training.
Off topic, but: The Feds only demonstrate their idiocy by looking at passing rates in the JC world and elsewhere; let all in - and fail those who do not meet the standard. No harm, no foul - nice try, come again.

Ken Plahn writes:

@Steve Sailer:
The new course are not just better than correspondence, they are way better than brick and mortar. I've taken the new courses under Stanford and currently taking courses under Udacity.

Sebastian Thrun is one of the best professors in the world and 2/3 of his students stopped coming to his brick and mortar class because they could learn better through his online course. This is despite the fact they were paying lots of money to go to Stanford.

Winton Bates writes:

One point that doesn't seem to have been covered is the quality of the degree. If I was setting up an online college I think I would be trying to create a credible signal that graduates would be required to meet higher standards.

The hypotheses are not mutually exclusive, and all apply to each student to a degree. Further, they exclude a fourth possibility: self-taught, self-paced curricula. All that is necessary for a substantial shift to self-taught, self-paced curricula is for some publisher (e.g., Dover, Springer) to make a deal with some private, accredited university to construct exams around the publisher's texts. Make these exams available through licensed examination centers. Let students who desire a transcript from the cooperating university determine for themselves whether they need tutors, study groups, or time to themselves. Let competition between Sylvan Learning Centers, the Kumon Institute, and the University of Phoenix drive the cost of a college degree down to the cost of books and grading exams.

Konkvistador writes:

I don't think people realize that you can use online education that is selective, to signal intelligence and conscientiousness. Some research shows students learn on average about as much in online approaches as classroom approaches currently. The interesting part is that the same studies also show conscientiousness being a better predictor of outcomes in online approaches. Those who lack it do worse, those who don't do better than in the classroom.

But leaving that aside, surely they top 5% of a Googleversity would have more than nearly any average student of a regular universities and at the same time there are far too many in those 5% for the Ivy Leagues to absorb them. This is especially true since Googleversity would be global.

Unlike traditional education, the power of big data will be easy to apply to optimize the online experience. Imagine what data mining wizards or even automated systems could do to optimize the learning experience once they have a good data. What could you learn from every click, every solved exercise and every question of say 20 million students that have learned about Statistics. How much better could it get?

Udacity run by the Google researcher Sebastian Thrun seems rather close to that and there are signs they know this and are counting on it.

Bernie writes:

I don't know much about traditional college, so this could be wrong. The impression I have is that it's difficult to finish early (limited required course offerings and hard start dates). It seems like online college eliminates these problems.

Here's my guess, online college will move be where the really smart kids go to be able to tell employers they finished in 2 or 3 years. The online college will then be high status and the traditional system will collapse.

Slocum writes:

Online education suffers from a severe adverse selection problem, because the students most eager to avoid traditional education tend to be deficient in one or more of these traits - especially conformity to the established social norm that young people should go to a traditional college.

You're ignoring the possibility of different signals by those completing online educations. Online students will signal the ability for self-motivation & independent work -- and a lack of need for the spoon-feeding and hand-holding that happens in the traditional model. Also, online degrees will probably offer a much more objective measures of performance (vs the grade inflation endemic in traditional institutions). At least some employers in some fields may come to prefer these signals.

Paul writes:

Finally. Some rational thinking on this topic. The partisans on each side have dominated the debate till now. I was at a meeting of the Board of Regents of my public university and they grilled an education professor about online education. Visions of Khan Academy technology replacing all those expensive faculty were dancing in their heads. Their problem , in my opinion, is that they (and the State legislature) don't have a clear perspective of what they believe the university is supposed to be. So the Board will simultaneously approve an expensive large new sports arena while simultaneously pushing for more adjuncts to teach online courses...

Troy Camplin writes:

Given your model, online will completely replace community colleges. Which will be a pretty significant impact.

Mike W writes:

@Slocum, "Online students will signal the ability for self-motivation & independent work..."

Based on the student completion rates of for-profit on-line providers that is a pretty small population.

Ken Plahn writes:

Employers will 'get the signal' when one of their competitors begins running circles around them.

It's really hard to explain this to someone who hasn't taken an in depth class through Stanford (coursera) or Udacity (spinoff from Stanford). It's like trying to explain the internet to someone who's never used a PC with a mouse.

Meanwhile, MIT and Harvard are busy playing catch up.

mkt writes:

The "Human Capital Model" in Perspective 1 makes a huge and unwarranted assumption: that online schools can increase their students' human capital as effectively as bricks-and-mortar schools do. For certain topics, and certain motivated students, sure (online traffic schools are a good example). But for effectively teaching say introductory calculus or intro economics to millions of 18-year old freshmen? If we try to do that online, the dropout rates from those classes will make today's dropout rates look like chicken feed.

Online education is important, but via hybrid classes; it won't replace bricks-and-mortar education (except for those specialized niches).

tom writes:

Some commenters above suggest that it will depend on the type of studies. I think that's partly right. But the best way to consider the problem is: what powerful existing groups could push things online:

1. How much of online education will be driven by big state universities as a way to slow/stop the increase in tuition?

For example, say that Michigan State says you can either wait until next term to take [name an oversubscribed class] in person, or you can take the online version. Either way your transcript will show the same class taken. In very little time, the school can convert a big group of people to the largely-online option. The school can avoid increasing classroom space etc... from the first year, then in a few years it can start saving on needing to charge most students room and board. And the state universities could also say that you can get a degree from the main university even if you only 'attend' a local version of the school near your home town.

Within a few years, tuition can be spent almost solely on administrators for year-abroad programs!)

2. When do big reputable employers start to pay for their existing employees to do the _____ Academy or ____ University in [name your science/other technical field] online instead of night school or spending a year going back to school? Not secretarial or administrative school, but substantive work. When Intel and Amgen are vouching for an online program by 'sending' their graduates there, that will brand the program as a place for success.

3. How could a large enough number of highly capable and not-too-weird 17-18 year olds be gathered together to form some kind of class that could, if done ten years in a row, help to develop a positive reputation for a new branded online program? I'd guess that it would be by an already-prestigious university (maybe a state school system like UC, Wisc., Mich., SUNY...) creating a largely online degree program with exams still taken in person though largely graded by computer. The state buys participation by offering free education to top high-school graduates who accept the online school degree.

Forrest Higgs writes:


I suspect that only those universities and more specifically, those departments in universities that require large laboratory investments will remain in 25 years. The rest will evaporate and be replaced by on-line education.

essen writes:

Online education may not work for education dealing with the traditional engineering disciplines (read non-IT engineers) and medicine. You will have do the practicals with microprocessors, oscillators, furnaces and cadavers in brick and mortar labs.

David writes:

The breakdown between perspectives is probably dynamic with respect to time. Education now may be largely about signaling, but I can see it changing to be more about marketable skills in 10 years. By Bryan's logic, that should give online schools more of an advantage in the future than we might anticipate.

That said, I'm skeptical that in a "marketable skills" world, online education will dominate bricks-and-mortar education. Certainly, at current prices for college, online education has the potential for a big price advantage. But the price of highly effective online education may prove to be higher than we think it will be. And the cost of bricks-and-mortar education has a lot of room to decrease.

Brock writes:

I would not be sanguine about the prospects of the current institutional arrangement of higher education.

I think the most likely course is that higher Ed will be disaggregated. The educational parts will quickly be siphoned off by the likes of Udacity, Khan Academy or "competency based" models like Western Governors. This will significantly reduce the value proposition of standard University, as only signalling and status would remain. Those also could be achieved by other means, and without the "overhead" associated with education, such provision will be more cost efficient.

Heck, simple work experience is a good signal in my mind. Holding down a job for four years (with maybe a promotion thrown in during that time) is a very strong signal.

Geno.Pacioli writes:

I taught in a quantitative business discipline in Big 10 universities for about 12 years, and I've taught the same in a major online university (not UoP) for two years. I agree almost exactly with your weightings, but I would estimate 10-10-80. I can think of many examples of specific student attitudes, abilities and preferences I've seen over the years and your characterizations seem to fit the observations almost perfectly.

In particular, your statement "Online education suffers from a severe adverse selection problem, because the students most eager to avoid traditional education tend to be deficient in one or more of these traits - especially conformity to the established social norm that young people should go to a traditional college." is exceedingly accurate. About 80% of my students in the online university clearly fall into this category, though some didn't conform because they were young mothers and chose to defer their education.

Cheers, GP

Comments for this entry have been closed
Return to top