Bryan Caplan  

Caplan-Glaeser Debate: My Opening Statement

PRINT
Futures Trading Brought Down B... A Parable About ZTE...
Yesterday Ed Glaeser and I debated "All government support of higher education should be abolished" for the Soho Forum.  Many thanks to Gene Epstein for organizing, and Ed for being a great sport, great scholar, and great guy.  Here's my opening statement.  I'm happy, of course, to post anything Ed would like me to run.

Why should higher education receive government support?  There are two main arguments.

The first is the economic argument.  Government support is allegedly economically beneficial not merely for individual students, but for society as a whole. 

The second is the humanistic argument. Economic effects aside, government support is vital for the promotion of intrinsically valuable ideas, culture, and values. 

If I merely supported spending cuts, I'd only need to argue that both the economic and humanistic arguments are overrated.  Since I advocate full separation of college and state, however, I've got to go further.  And I do. 

My book, The Case Against Education, maintains that both the economic and humanistic arguments are deeply wrong.  Economically speaking, subsidizing higher education is like subsidizing polluting industries.  It's probably good these industries exist, but the free market tends to produce too much, not too little.  The humanistic argument is similarly flimsy; while I share the humanists' ideals, higher education simply isn't very persuasive or transformative.  The vast majority of college students arrive as philistines and leave as philistines. 

Since Ed and I are both economists, I'll focus on the economic argument.  The standard story says that college is great place to "build human capital."  Professors supposedly spend four years pouring useful job skills into their students.  Why should we believe this?  Because college graduates outearn high school grads by over 70%.  Employers aren't stupid.  If college didn't build tons of human capital, why would the labor market shower rewards on college grads?

This is a convincing story... until you remember what college professors actually teach.  Sure, there are a few majors that regularly prepare their students for the world of work, like engineering and computer science.  But most of what college students study is simply irrelevant in the labor market.  In real life, how often do you use history, government, literature, foreign languages, psychology, philosophy, or higher mathematics?  By and large, students can safely forget such material after the final exam, because it never comes up again as long as they live.   And forget it they do. 

These observations are so obvious you might wonder how anyone can deny them.  Think about your own educational experience: How many thousands of hours did you spend studying foreseeably useless material?  Economists' standard response is simply to double back to the labor market.  If college coursework were largely irrelevant in real life, real life employers wouldn't pay college graduates a handsome premium. 

Strangely, though, there's a Nobel prize-winning economic model that explains why even the most irrelevant coursework and silliest majors can be financially rewarding.  It's called signaling.  Basic idea: Academic success is a great way to convince employers that you've got the Right Stuff - to show off your brains, work ethic, and sheep-like conformity.  Since people with these traits are productive workers, employers happily reward people who display them - even if the display itself has nothing to do with the job.

Think about it like this: There are two distinct ways to raise the value of a diamond.  The first is give it to a gemsmith to cut the diamond to perfection.  The second is to give it to an appraiser to attest to its flawlessness.  The first story is like human capital.  The second is like signaling. 

What difference does the mechanism make?  For the individual student, not much.  For society, however, it makes all the difference in the world.  Insofar as the human capital model is right, government support for college enriches society as a whole by upgrading the skills of the workforce.  Insofar as the signaling model is right, however, government support for college impoverishes society by sparking a credentialist arms race.

So which model is right - human capital or signaling?  The truth is obviously a mix of both.  In The Case Against Education, however, I argue that signaling's share of the mix exceeds half - and probably a lot more. 

Why should you agree with me?  For starters, look at the curriculum.  Most of what we teach in college is so otherworldly that you're only likely to use it on the job if you become a college professor yourself.  No sane person with a non-academic job panics because they've forgotten everything their professors taught them about history, literature, or philosophy.

Curriculum aside, you probably already tacitly agree with me.  Did you bother to enroll in college or pay tuition?  If all you wanted was the learning, this was a total waste, because you can unofficially take classes at virtually any college in America for free.  There's just one little problem.  At the end of four years of guerrilla education, you won't have the crucial signal: the diploma.  Hence, unofficial education barely exists.

Suppose you could have a Princeton education without the diploma or a Princeton diploma without the education.  Which would you choose?  If you have to ponder, you already believe in the power of signaling.  By contrast, if you were stranded on a island and had to choose between knowledge of boat-building and a boat-building degree, you wouldn't ponder.  When you face the labor market, it's important to be impressive.   When you face the ocean, all that matters is skill.

Signaling dominates if you look at the way college students approach their studies.  They routinely seek out "easy A's" - professors who dole out strong signals in exchange for little work.  Of course you don't learn much from such professors, but who cares?  After the final exam, you'll never need to know it again.  Signaling likewise explains why academic cheating isn't just "cheating yourself."  When you impersonate a good student, you hurt employers who hire you - and the honest students whose merits you indirectly call into question.

Academic research reinforces common-sense.  While economists typically measure education's annual return, scholars who look find enormous diploma or "sheepskin effects."  Senior year pays far more than the earlier three years combined!  This is very hard for human capital theory to explain.  It makes perfect sense if college students are trying to signal their conformity by completing their degree. 

Macroeconomists, similarly, have found that while individual education has a big effect on individual income, national education has only a small effect on national income.  To be fair, they rarely embrace the signaling explanation; instead they cry for better data so "we can get the right answer."  But signaling cleanly explains their results: If one laborer gets more irrelevant education, he outshines the competition; but if a whole labor force gets more irrelevant education, society's time and money go down the drain.  Given the small effect of education on GDP, it's hardly surprising that few researchers find that education leads to higher GDP growth.  If you're still trying to figure out if your machine moves at all, you can safely conclude it's not a perpetual motion machine.

The most striking academic evidence for signaling, though, comes from the literature on "credential inflation."  The average worker is years more educated than he used to be.  How much of this is because jobs are more cognitively demanding?  How much of this is because workers need more education to get - though not to do - a given job?  In The Case Against Education, I examine all the main studies.  Punchline: The evolving labor market explains only about 20% of the rise in education.  The remaining 80% is credential inflation: You need college to convince today's employers to give you the same jobs your parents or grandparents got right out of high school.  This is puzzling for human capital theory.  Why should employers pay for B.A.s when you need only a high school education to do the job?  Signaling explains it elegantly: The more degrees proliferate, the more you need to stand out.

When I present these arguments, economists rarely deny that signaling seems like a persuasive story.  Instead, they usually retreat to apriori objections: Appearances notwithstanding, signaling can't be right.  The most popular objection: College "passes the market test."  If it were mostly signaling, someone would have figured out a cheaper signal long ago. 

But this this crazy.  Higher education receives hundreds of billions of dollars of taxpayer support every year - a classic sign that it fails the market test.  There are probably plenty of socially cheaper forms of labor market signaling.  But as long as the massive subsidies continue, the substitutes will remain on the fringes.  The easiest way to discover good alternatives is to end government support for higher ed - and see what comes next.

At this point, you could respond, "Sure, education is mostly signaling.  But the economic rewards are so great that it's still worth subsidizing."  But signaling aside, there's far less to education's economic rewards than meets the eye.

Why?  First, college graduates aren't randomly selected.  Most were already high performers back in high school; if they hadn't gone to college, they probably would have been fairly successful anyway.  When researchers statistically compare high school graduates to college graduates with equal pre-college ability, they almost always find the true effect of college on personal success is smaller than it seems.

Second, standard comparisons focus on people who finish college.  But this is cheating, because the college graduation rate for full-time students is about 50%.  When you weigh college as an investment, this slashes the expected return.

A debate is admittedly not a great place to do arithmetic.  But in The Case Against Education, I snap all these pieces - and many others - together.  Along the way, I seriously study potentially neglected benefits of college: health, crime, you name it.  Because confirmation bias is bad.  Final result: From a social point of view, investments in college aren't just overrated; they're ruinous.  Subsidizing this rat race is as economically foolish as handing out big cash prizes to the world's dirtiest polluters.

When the economic case for tax-subsidized college crumbles, even many economists suddenly discover the "finer things in life."   What about the humanistic argument that college inspires love of ideas and culture - that it refines and elevates us?  My quick response: Refinement and elevation would be great... if it really happened.  But the actual data say that it's mostly wishful thinking.  Only a miniscule fraction of college grads take a meaningful interest in ideas or culture after graduation.  People who attend events like the Soho Forum for fun are really weird.  That's why I love you guys! 

You could agree with every word I've said so far, but respond, "Instead of abolishing government support for higher education, government should use the power of the purse to fix higher education."  We can certainly imagine a world where colleges fill every student's mind with human capital and every student's heart with Shakespeare.  Why not do that?

Simple: Defunding dysfunctional systems is almost fool-proof.  Fixing dysfunctional systems, in contrast, is horribly hard.  As a professor, I assure you that the entire system bitterly resists even mild reforms.  Most professors detest the very idea of objectively measuring the value of anything they do.  They're artists!  You can't deal with these people - and it's foolish to try.  If someone says, "Sorry for wasting trillions of tax dollars.  But we did a few good things, and we'll spend your money wisely from now on," the prudent reaction is to draw a line in the sand and say, "You're fired." 

But why should we be so extreme?  Why not just cut half and see if that does the trick?  Pragmatically speaking, abolition is far more transparent.  The scope of partial reforms is always confusing and debatable.  When you separate college and state, it's clear-cut.

Since this is the Soho Forum, let me end with my principled argument for full abolition: the presumption of liberty.  I know there's a wide range of libertarian views.  In fact, there's one libertarian view per libertarian.  But we should all be able to agree that the burden of proof rests on the advocates of government intervention.  If politicians are going to take our money without our consent, they should at least have solid proof that the money is very well-spent.  Government support for colleges does not meet that bar.  Not even close.

Update: Gene Epstein sent me the vote tally.


PRE

POST

CHANGE

YES

50.49%

63.11%

12.62%

NO

25.24%

25.24%

0

UNDECIDED

24.27%

11.65%

-12.62%





COMMENTS (9 to date)
BC writes:

"If all you wanted was the learning, this was a total waste, because you can unofficially take classes at virtually any college in America for free."

Here's an idea: suppose we only offer college financial aid for students that *don't* receive degrees. Caplan says most universities will let students audit classes for free but, if a lot of students did that, then universities might start requiring those students to pay tuition. Offering aid to only non-degree students should satisfy *both* signalling and human capital proponents. If education is mostly signalling, then few people will choose non-degree programs, so the financial aid will go mostly unused. If education builds lots of human capital, then non-degree financial aid will enable students to build that human capital. So, if financial aid is limited to non-degree programs, then we don't even need to agree on whether college is mostly about signalling or human capital.

Dave Smith writes:

Half of the audience started out agreeing with you? Did I read that right?

Bryan Pick writes:

Dave, the audience's sympathy with Caplan's argument is less surprising when you consider that the Soho Forum is an explicitly libertarian event.

Ahmed Fares writes:

The fact that college graduates do better in life is simply due to selection bias.

It has been said that if you took the college-bound and locked them in a closet for four years, they would still do better in life than those who didn't qualify to get into college.

Students come out of college smart because they were smart going in. Students that graduate from Princeton are smarter because Princeton selects at a higher level. There's nothing special about Princeton. Princeton teaches from the same textbooks. It's not the college that makes the student, it's the student that makes the college.

So yes, of course college graduates do better in life. But you could tell that just by looking at their high school marks.

Alex Douglas writes:

Any chance at a full video or transcript of the debate? Would love to see the back-and-forth.

ASG writes:

My impression of computer science courses is that they are only loosely related to what happens in industry. Other opinions are available.

Seth writes:

"Sure, education is mostly signaling. But the economic rewards are so great that it's still worth subsidizing."

If it's such a good investment it doesn't need to be subsidized.

Michael Rulle writes:

Every time I read one of Bryan's essays on this subject, my opinion on what is written swings wildly back and forth between agreeing and disagreeing. Because I agree with some things and disagree with others. A few random or ad hoc thoughts.

It is easy to agree that the naturally brilliant will succeed regardless of signaling or substantive learning. These are also not a mutually exclusive phenomenon.

I believe Bryan leans too negatively on the ability of people to learn in college in such a way that it is beneficial for the long run. We all know one can skate through even the best colleges. But my experience with young people is most have a true desire to learn and also to learn what will lead them toward their desired professions.

My own "common sense", for what that's worth, tells me the cost of college is far too high for what it delivers. Much of the increased cost of schooling appears to be a function of providing a modern and wealthy environment to live in, the increasing modern-day perks of a private school are large and unrelated to education. Like a socialist paradise.

I do agree that push financing by schools to encourage students to take on debt (with the government's assistance) borders on, or maybe is well within the borders, of immoral behavior.

Without a doubt, parents and students do believe to the point of desperation, in the importance of getting into the best name school possible.

Community colleges, even if subsidized, clearly cost significantly less. Many people spend two years in CC and then two years in a traditional college. But kids feel insecure and succumb to peer pressure and feel embarrassed to do that unless they have to.

We should not subsidize colleges with federal loans pushed down to students. A more free market will correct some of these issues by necessity. This is the most practical thing we can do to improve the issues Bryan raises.

Companies also are part of this phenomenon. They feel most comfortable with a certain status or class type which is part of what drives the process.

But it is hard to believe that in the end, free markets will not prevail. Companies should want the best people for the job, regardless.

But it really is true that the most natively intelligent people, tend to go to the best schools---although this is certainly not universal. The opposite is not true of course. Just because you do not attend a "top school" does not mean you do not have high native intelligence. I would wager that at least the top 10%-20% of huge major state schools match the quality of students at schools ranked in the top 25.

Michael Skolnik writes:

Bryan’s arguments — and the evidence that he uses to support them — are quite compelling, and if forced to choose Yes or No to the proposition posed for the Soho Forum, I suppose I would choose to stand on the Yes side. But it would be an uncomfortable choice because of the heterogeneity of the enterprise called “higher education”.

One set of activities about which I would have particular concern is that of postsecondary vocational programs, such as the applied associate degree programs offered by community colleges in fields of high employment demand. In contrast to many traditional Bachelor of Arts programs, the value of these 2-year Associate of Applied Science (AAS) degrees lies mainly in the job ready occupational skills that they produce in graduates, though for astute employers they may also signal some desirable personal traits the most important of which may be a preference for substance over symbol. If Bryan’s principal argument for not giving public funding to higher education is that it does not embody a sufficiently high proportion of human capital creation, the prescription risks throwing out some human capital laden babies with the signaling bathwater.

Of course opening the door to exceptions — and I can think of a few other meritorious ones — is to embark on the slippery slope that leads us back to where we are now, as the experience of trying to provide special funding for STEM (and earlier attempts to target programs deemed critical to national defense) attests. I don’t have a solution to that problem, but since it is ideal states that we are lining up for rather than policies that have much chance of actual adoption, another attractive option is for a substantial reduction in public subsidization of higher education, along with retaining its use selectively in cases where the public benefit can be readily established.

POST A COMMENT




Return to top