Quick, name a growing, well-hyped company that’s trying to Disrupt the Education Industry!
Lambda School is the obvious answer. I’ve written about them before, and I think they’re on to something big. Lambda is changing education as it’s popularly understood: you go to class to learn skills, then you use those skills in a job you couldn’t otherwise do. The popular narrative is that that’s what our education system is, and that that’s why we spend $1.1 trillion dollars per year on it.
But there are cynics who believe it’s something else entirely. In The Case Against Education, Bryan Caplan argues that the real value of education, for the people who get it, is signaling, not skills — memorably, he argues that a college is not so much a sculptor as an appraiser.
There’s good evidence for this, both quantitative and anecdotal. For example:
- The “Sheepskin effect,” where the single semester that accounts for the biggest share of the college earnings premium is the last one. Do schools save all their most valuable knowledge for the last few months, or do employers check a box that says “college graduate”?
- The relationship between income and years of education is much stronger for individuals than for countries; a college graduate is richer than a high school dropout, on average, but to a much greater degree than countries full of high school graduates are richer than countries where most people leave school earlier. This implies that people use education to compete for a relatively fixed pool of opportunities.
- Students feel bad about missing class, but feel happy when class is cancelled. Would you celebrate if you ordered a meal on Seamless, paid for it, then got an email saying “Good news everyone — the cook is sick. No food for anyone today”?
But Caplan is idealistic compared to the most cynical view, which is that American universities are just a nonstop party. To be fair to the colleges, under this scenario they’re at least a pretty good deal — on a dollars per night of partying basis, they beat Bonnaroo, 1 OAK, and other venues that sell the opportunity to get intoxicated while surrounded by attractive young people.
My view is that higher education is a bundle of all of these services. Like Netflix, it costs one fixed price but offers different attractive options to different subscribers. In this model, you might say college sells:
- Signaling value to parents, both through the hope that Microsoft and Accenture will think your child is as special and wonderful as you do, and through the fear that your kid will will be living with you at 35.
- Human capital to lenders and governments, who should only invest in schools if they produce either public goods like a well-trained populace or good collateral like a valuable employee.
- Fun fun fun to the students themselves.
Bundles work well when customers have heterogeneous preferences and the marginal cost of delivering one more feature is low. But bundled products tend to relentlessly increase in feature count and cost, which eventually makes them a bad deal for people who want exactly one thing.
Some schools respond by specializing. MIT, Olin, and St. John’s all focus on providing human capital. Harvard Extension and University of Phoenix are two ways to buy signaling value (one is, technically, Harvard, and one is, technically, a degree). But they’re still starting with a standard university model and subtracting features. The real question is what to build if you’re offering the same value to customers but you’re designing the means to provide it on a clean slate.
In Defense of Signaling
The thrust of Caplan’s book is that we, as a country, spend far too much on education. He makes a good case; signaling is pretty zero-sum, and if education is funded because it provides human capital, but it actually provides signaling value, that’s a loss for society.
But while signaling is pretty zero-sum, it’s not exactly zero-sum. Going back to the appraisers example, appraisers provide value, which is why they get paid.
It’s easy to misread the signaling critique to say that schools don’t provide any value, when the real critique is that elite schools provide value inefficiently. What Harvard really does is 1) tell 18-year-olds they have the Harvard Stamp of Approval, and then 2) Make them wait in line for four years to collect it.
No wonder college students spend so much time partying and protesting. They’re bored.
A more efficient approach might be to staple a diploma to the admissions letter, and make all courses extra-credit. Given that the standard deviation for Ivy League GPAs appears to be roughly 0.05, it wouldn’t make a huge difference.
But that’s a tad extreme. It’s a bit of a waste, really. There are two problems:
- One way schools add value is by introducing talented people at a formative age. Even a stripped-down signaling-only college should at least host a few mixers so future elites can get to know each other.
- If a college provides a good signal, this means its admissions department can accurately judge what a young person will accomplish in their life. But a closely-related skill to this is telling them what they should do a bit differently. Bill Gates dropped out of Harvard, but one of the things he learned there was that he wasn’t going to succeed as a mathematician.
If you were reinventing the Ivy League as a signaling-focused product, your stripped-down version might look like this: you invite a small cohort of talented people to move to a city for about three months, you host some social events so they get to know each other, you have them work on projects and you advise them on those, and afterwards you introduce them to a bunch of savvy rich people.
In other words, you’d invent Y Combinator.
Tuition and Endowments
Let’s keep going. We’ve streamlined the product, now let’s streamline the revenue model. The way elite colleges work is that they charge a variable tuition, where the sticker price is designed so non-rich students feel like they got a great deal. Many schools make good money from research grants, but we’ve dropped that whole business line. However, many of the best schools make money from student gifts, and from preferential access to good investments for their endowment funds.
We can combine all of these features. What’s more attractive than low tuition? Negative tuition. What’s better than ad-hoc fundraising and after-the-fact investing? Just taking equity upfront.
Harvard owned a sort of hazy equity position in John Paulson’s future earnings; he got an MBA there, which got him a job, which led him to start a hedge fund, which made him billions of dollars. Decades later, $400 million of that came right back to Harvard in the form of a donation.
The transaction costs would have been lower if Harvard had just gotten carry in Paulson’s career. So that’s how YC does it. Of course, a YC alum can “default” on their implicit obligation to provide returns for YC, but this model switches the donation system from opt-in to opt-out. Unless you either fail or actively attempt to screw them over, you’re donating some of your future earnings to your beloved alma mater.
This raises an interesting comparison with (YC graduate) Lambda School. Famously, Lambda’s tuition is in the form of a capped income-share agreement; you pay a fraction of your increased earnings, up to a set maximum. This implies that Lambda is helping students earn more, but with a pretty normal distribution; it’s mezzanine debt. YC charges in the form of equity, and their returns are dominated by a few extreme successes.
This says something about the returns on signaling compared to skill-building: it’s hard to learn a million-dollar-a-year skill from school (perhaps the people who could make a million dollars a year are so insatiably obsessed by their chosen career path that they’ll learn it fastest on their own). But as YC’s success indicates, there have been many hypothetical billion-dollar companies that didn’t come into existence because the founders never got the right introductions. Your takeaway from this should be that if you’re very ambitious, the biggest risk you can take is to be a loner.
American universities are some of the most powerful institutions on earth. They provide a valuable brand name to talented people, they create incestuous networks that persist for decades, and their in-house hedge funds are all on the AUM leaderboard. But they’re not the best opportunity for everyone, and in particular they’re not the best opportunity for some of their best students.
As Paul Graham once pointed out, both Facebook and Microsoft were not just founded at Harvard, but specifically founded during the January “reading period,” when students are supposed to be studying for exams. The most lucrative months in Harvard history are the months where class is not in session!
YC is creating a lot of wealth by providing a more focused version of the elite undergrad experience in about 95% less time, but if they keep expanding in this niche, they’ll eventually lose out. Already, there are a few companies that have applied to YC, gotten accepted, and ostentatiously turned it down.
If they succeed in building a better Harvard, they’ll do well for themselves and do good for the world, but, like Harvard, their most successful alumni will be the ones who turned them down.
 I always assumed Steve Ballmer was the emotionally volatile buffoon he was portrayed as in the 90s, but it turns out he took the Putnam, and outscored Gates.