Bryan Caplan  

Reply to Tyler on The Case Against Education

Bad news from China... Caplan versus Hanushek...
Tyler has two recent posts referencing my new book.  Here are my responses.  Tyler is in blockquotes; I'm not.

Response to Tyler's first post:

I'll be doing a Conversation with Bryan, but for the time being I'll say this: everyone obsesses over the mood-affiliated "I'm going to lower the status of education signaling argument."  Hardly anyone has discussed what to me is Bryan's strangest assumption, namely a sociologically-rooted, actually anti-economics "conformity is stronger than you think" argument,
Coordination games have long been standard in economics: If everyone else is driving on the left-hand side of the road, conformity is prudent, regardless of what's globally optimal.  Indeed, branded as "path dependence," empirical economists have used this approach to study not only technology, but location, language, and much more.  Yes, many of the specific tech path-dependence stories have been overblown.  But the general insight remains profound: our current system of English spelling is godawful, but who expects a decentralized switch to a sensible phonetic system?

Thus, there's nothing "anti-economics" about my approach - unless it's "anti-economics" to take psychology and sociology seriously.

That said, conformity does play a somewhat novel role in my analysis.  To be specific, I argue that:

1. Conformity is one of the main traits that employers want. 

2. Education signals conformity. 

3. Trying to signal conformity in unconventional ways generally signals non-conformity, leading to serious lock-in of existing conformity signals.

Again, how are any of these claims "anti-economics"?
...which Bryan uses to assert the status quo will continue more or less indefinitely.  It won't.  To the extent Bryan is correct (and that you can debate, but at least he is more correct than most people in the educational establishment will let on), competency-based learning and changes in employer behavior will in fact bring about a new equilibrium...not quickly, but certainly in well under two decades.
As you you might guess, I'm happy to bet Tyler on this.  Tyler didn't bite when I offered terms on this seven years ago (and wisely so).

And what about on-line education?  Well, a lot of students don't like it because they have to actually work on their own and pay attention.  To the extent education really is just signaling, that should give on-line options a brighter future all the more.  But not in the Caplanian world view, as conformity serves once again as an intervening factor.

Fans of online education initially made (fairly reasonable) arguments about how it's better for human capital creation.  Lately, they've switched to (pretty unreasonable) arguments about how online education is going to provide great signals too.  My story explains why online education won't dominate despite its obvious advantages for learning.  What other explanation is there?   

For better or worse, Bryan's book subverts economics as a science at least as much as it does education. 

Again, no.  My book shows the power of economics, though of course economics works best when married to psychology, sociology, and beyond.

Bryan of course is smart enough to see the trade-offs here, and he knows if the standard model of economic competition were allowed to reign supreme, we would (even with subsidies, relative to those subsidies) tend to see strong moves toward relatively efficient means of signaling, if only through changes in the relative sizes of institutions.

When government gives the status quo a trillion dollars of subsidies per year, efficiency "relative to those subsidies" is nothing to crow over.  And I see little sign that we're approaching even that low bar.  Some small schools are going bankrupt, but that's just one pro-efficiency force in a deeply dysfunctional system.

Response to Tyler's second post:

My view is this:

1. Learning at least one language is of high value for America's elite.  It helps them see different points of view,

Seems like a small payoff for 2-4 years of high school coursework.

and prepares a small number for careers in the foreign service or in other international capacities. 

Only slightly, given the low quality of instruction.

It makes intellectuals deeper and improves their scholarship.

Frankly, it's not even clear that the correlation is positive.  Lots of Ph.D. programs have dropped their foreign language requirements.  Should we bring them back?!


2. If we could target foreign language acquisition to this future elite, I would gladly let the vast majority of the student population off the hook.  One move toward this end would be to use foreign language "tiebreakers" for those wishing to finish in the top quarter of their high school class. 

An improvement.  But we've got the standard inefficiency of signaling: If you need a foreign language to be the in the top quarter, far more than 25% of students will study one in the hope of occupying that top quarter.


3. Here is an estimate that knowing a foreign language brings a wage premium of about 2%; I have not read the paper, but I do not wish to overclaim on the causality front.  It still is measuring something about quality.

That seems like a pretty low estimate for actually knowing a foreign language.  But as I've shown, near-zero Americans even claim to have learned to speak a foreign language very well in school.

4. Many European countries teach their citizens English (above all), French, and German with reasonable success and high returns, especially for English.  So it is possible to succeed with this endeavor within a public education system.

Yes, it's much easier to teach students subjects they expect to use one day - and frequently practice outside of school regardless of the curriculum.


6. I do not think opportunity costs during high school are especially high.

This is probably my biggest disagreement with Tyler.  If the government required him to study a blatantly useless subject for years, I think he'd consider the opportunity cost very high indeed - even if it just crowded out his leisure time.  I say that imposing such requirements on kids is an outrageous form of bullying.  If any government did this to adults, who would defend it?

Addendum: Disagreement aside, I am so pleased that this year the two big exciting economics books so far are by Bryan and Robin Hanson.  Do buy and read both!  As for Bryan, maybe some of you are thinking you just can't accept his argument about education being so wasteful.  But I'll say this: Bryan always defends his "absurd" views much better than you think he is going to be able to.

Thanks!  I would add, though, that most of what I say seems like common sense to readers who focus on their first-hand educational experiences instead of ubiquitous pro-education propaganda.

COMMENTS (17 to date)
DWAnderson writes:

Re the utility of foreign languages: my wife is a former US Foreign Service officer who speaks Italian and Spanish. Despite her years of Spanish study in high school and college, she believe that most language instruction in school is a waste of time. Unless you use the language, you will not retain the vast majority of what you learn.

FWIW, she found the best mode of foreign language instruction to be the multi-month program all-day instruction the Foreign Service used (uses?) in preparation for overseas assignments. That is obviously far afield from what any school can (or should) do.

JH writes:

Hi Bryan, I have a follow up question about the role that conformity plays in your argument. I mentioned something similar on another post, but I want to raise the question again.

Can you say why the equilibrium is that going to college signals conformity? Why did we converge on that equilibrium? More broadly, why are many countries converging on this equilibrium as well? For example, college enrollments are up in other high-income countries too.

Part of my puzzlement about this issue is that college otherwise seems like a relatively bad way to signal the qualities that employers want. You present evidence in your book that most college students don't work hard or learn marketable skills. So why would employers prefer college students over people who skipped college, actually worked for four years, and have high IQs? It seems like this would signal what employers want more efficiently than college (intelligence, diligence, etc).

So, suppose that Marvin and Ed have equal IQs, are equally hard working, and so on. Ed goes to college and gets a degree in English. Marvin works for four years instead--perhaps he starts in low-wage jobs and then gradually moves up. Why would employers prefer Ed over Marvin?

Your response is that skipping college doesn't signal conformity and employers care a great deal about conformity. But this raises the question: why did we converge on the "college signals conformity" equilibrium to begin with? The other way I suggested (actual work experience) seems more efficient. What's especially peculiar is that many societies have converged on the "college signals conformity" equilibrium.

Perhaps your answer is that government puts its thumb on the scales with subsidies for education and that's why we prefer it. Maybe. But I'm just not sure that's plausible because so much money is on the table. If there is a better way to signal, I'm skeptical government subsidies would do much to stop this other way from winning out. And why would many countries be moving in the same direction? There is considerable variation in how much governments support higher education.

BTW, I really loved the book and I definitely find the signaling model more plausible now. Still not entirely sold though because I can't think of good answers to the above questions.

Alberto Zaragoza Comendador writes:

Walk around the streets of any decent-sized Spanish city and it won't take you three minutes to stumble on an English academy. Not a "language" academy; there are some of those, but most are run by the government (Escuela Oficial de Idiomas). And academies that teach exclusively a non-English language? Literally the only ones I can think of are those sponsored by foreign governments, e.g. Goethe Institute.

Really, the supply of (and demand for) English education must be about 10 times as big as that of all other languages combined. Part of this is due to academic requirements (many schools are supposedly "bilingual" now), but a lot is also coming from adults in their free time, people who take in-company classes, etc.

Cowen's point is that learning a foreign language can be a worthwhile investment; as an English language teacher I do agree! But for someone who already lives in an English-speaking country, learning something else is usually not worth the time. Lots of Americans take Spanish as a second language in high school, but walking around American cities how often do you see Spanish academies?

Ricardo writes:

"and prepares a small number for careers in the foreign service"

Number of foreign service officers, for whom foreign language proficiency is required: roughly 8,000 (State Department website).

Number of professional athletes in the US: roughly 11,800 (BLS).

Jay writes:

At least one study of online-only education so far says it's pretty much worthless:

@JH: Obedience to arbitrary rules is one of the main things employers want because their employees are expected to adhere to arbitrary rules. It's impossible to manage thousands of people if they won't conform to expectations, even when the reasons for those expectations aren't obvious (or are ridiculous, but aren't worth the company's effort to fight).

john hare writes:

I did a years worth of credits in engineering at a community college that I have found quite useful over the last few decades. In my business though, not to signal employers. I also dropped out of the sixth grade and don't think I missed much learning in skipping six years of junior and high school. In my company they call an elementary school dropout "boss".

darf ferrara writes:

To lower the costs of learning a language it should be done earlier. Ideally when children are very young (preschool and younger). Between the ages of 8 and 12 the language acquisition hardware in the brain loses the ability to easily learn language.

The benefits of learning second languages isn't limited to on-the-job functions either. Knowing a second language may help prevent Alzheimer's, and it seems to help patients who have had a stroke.

Ted Craig writes:

I think the weakest of Tyler'so arguments is that learning another language brings increased perspective.

Dan writes:

Bryan, your conformity argument is misguided. Social psychology does support the notion that people conform to the norms of the group...but signaling that one is a conformist is a negative signal...especially in an independent culture such as our own.

Joe munson writes:

This is only slightly related It's weird to me that people often use "brain is developing" and or low executive function to justify legislating teenage lives and binding them to their often terrible parents, but never extend it to the intellectually disabled and or people with more minor learning disorders.

Daniel Jelski writes:

I listened to the debate between Bryan and Mark Hanushek--really interesting. (My quotes are paraphrases.)

Mark asked "what costs would Bryan actually cut from K-12 education?" Teachers aren't hugely overpaid (though pensions are too generous), facilities are shabby, etc.

I have an acquaintance--a self-described "master violinist" who teaches middle school orchestra to earn a living. He hates the job (though he takes pride in the orchestra sounding "really good!"). He hates it for all the obvious reasons: the students have no talent, aren't interested, are disruptive, etc. Even in the best of circumstances none of them will learn how to play the violin (absent private lessons). It's extraordinarily frustrating.

So Bryan will suggest that we get rid of orchestra class and simply shorten the school day. But that would be a disaster.

I think society has three choices:

1) Get teenagers jobs. (That's an option for the minority who can hold down a job.)
2) Put them in jail. (Unpleasant, unjust, and expensive.)
3) Stick them in orchestra class.
4) Let them run wild in the streets!

Putting large numbers of teenagers out on the street is not a good idea. Somehow we have to keep them busy doing something at least nominally useful. Orchestra class is as good as any. And perhaps 1% or so might actually become musicians as a result.

So what would Bryan cut?

john hare writes:

5) Let them start dealing with the real world.

I'm not sure if it is semantic differences or a real disagreement that I don't agree with your #1. Let them find a job I would agree. Finding them jobs is (to me) a variation on orchestra or other time filler class.

2 & 3 are just different levels of unjust.

4 might be better if they are allowed to experience the consequences of their own actions. Hungry if unemployed, punished if lawbreaking, etc.

Niko Davor writes:

"My story explains why online education won't dominate despite its obvious advantages for learning. What other explanation is there?"


If an online class offered by Coursera/edX/udacity fared much worse than a similar online class offered by traditional university, that would be explained by the signaling theory and the power of the credentials and the institution.

If an online class offered by a traditional university fares much worse than the same class in traditional in-person format, offered by the same university, that actually contradicts the signaling theory and speaks to human needs for physical proximity to human peers and human teachers and coaches.

mkt42 writes:
If an online class offered by a traditional university fares much worse than the same class in traditional in-person format, offered by the same university, that actually contradicts the signaling theory and speaks to human needs for physical proximity to human peers and human teachers and coaches.

San Jose State University has been there, done that, and had high failure rates to show for it. Their use of online math classes in parallel with traditional classes ended after just one semester because the students in the online classes performed so miserably.

The notion that online classes can replace brick-and-mortar classes at the undergraduate level is a pipe dream. Online classes require highly motivated students and thus do not work for most college students.

Niko Davor writes:
San Jose State University has been there, done that, and had high failure rates to show for it.

Sure. Other universities had similar experiences.

The notion that online classes can replace brick-and-mortar classes at the undergraduate level is a pipe dream.


Normal students need the physical proximity to human peers going through the same class and physical proximity human teachers or coaches. Stripping those human elements away and dumping class content online doesn't work.

This doesn't mean you can't get the benefits that online education aims for. It means that the strictly online format doesn't work and you need brick + mortar presence.

This isn't explained by Caplan's signaling theory nor is it supported or refuted.

Nick Lilovich writes:

I would have challenged Brian differently - how does he respond to the apparently huge returns to women's education (especially in developing countries)? If those K-12 type years were also just signaling, why do women benefit so much?

If we agree with his premise, then this would imply that the main sexism in developing countries comes from the family - which won't let a woman work if the wage is low, but WILL let a woman work if the wage is high.

If employers were the sexist barrier, that would imply that signalling won't change anything, which it clearly does (in this signalling world Brian has argued).

Grahame Booker writes:

Although Bryan's approach is anti-economics it is apparently one of the two exciting economics books of the year!


Return to top