Bryan Caplan  

Education and Signaling: Rejoinder to Bill Dickens

Government in the Mortgage Bus... The Medicare Fraud in Our Futu...

Today I'm celebrating Labor Day by continuing my exchange with Bill Dickens on the signaling model of education.  (Previous rounds here, here, and here).  What's the Labor Day connection?  Simple: If I'm right, we'd be collectively better off if we spent fewer years in school and more years in the labor force.

One caveat: While I think that signaling is the main reason why our education system is so inefficient, it's hardly the only reason.  In my view, our education system is a Frankenstein's monster - a grab bag of inefficiencies hideously stitched together.  For example, K-12 schools' tolerance for disruptive behavior seems highly inefficient, but neither the human capital nor the signaling model explains this dysfunction.  In my reply, I try to stay focused on signaling, though I do occasionally reply to Bill's broader comments.

To keep the debate from getting stale, this is my last installment.  If Bill wants to reply, I'm happy to give him the last word.

Now onto the celebration.  My reply, point-by-point:

As I understand your position it is that private returns to education mainly reflect inefficient signaling. That is that employers are willing to pay for schooling not because students learn very much, but because completing schooling signals some traits (intelligence, perseverance) that are hard to observe otherwise.

Correct.  While I think that intelligence is fairly easy to observe (even in a regime where IQ tests are only semi-legal), conscientiousness and conformity are often hard to spot - especially when people have a strong incentive to fake them.  Even worse, low educational attainment relative to IQ is a strong signal of low conscientiousness and conformity.  So when employers interview a smart person with little education, they infer that the person is well below-average in other productive traits.

Further, you must believe that there is a more efficient way to solve the problem of figuring out who is fit for what type of job than running them through the educational gauntlet. I take it from your reply you think less schooling and more work is the way to accomplish this.

Right.  Some signaling is socially productive because it improves the match between workers and jobs.  But it's privately optimal for students to far exceed the social optimum. 

My preferred policy is simply to end government subsidies for education.  It might be even more efficient to go further and impose a Pigovian tax on education.  But since education is a mix of human capital creation and signaling, and government has already made a mess of things, I think the all-things-considered best approach is separation of school and state.  But someone could accept my views on signaling without going that far (or by going even further!).

...In addition, I believe that increased wages are only part of the benefit of education to the individual (the private returns to education) AND that there are social returns to education above the private returns so that the total value of education to society exceeds the individual returns and considerably exceeds the increase in wages we observe.

I agree that education has some positive externalities that partially balance out the negative externalities of signaling.  But I think that these positive externalities are overrated, and in any case stem from a tiny subset of coursework.  Positive political externalities of economic education?  I'll buy that.  Positive political externalities of history?  Probably not - only a handful of people can competently reason from historical examples to current issues.  Political externalities of English or sociology?  If anything, they're negative.

There is a third position, and I will claim it as my fall back position. That most of the return to education is due to it signaling desirable characteristics, but that there is no more efficient way to sort the capable from the incapable...

Even on your fall-back position, don't government subsidies for education make a bad situation worse?

In your reply you stated that education's value added was maybe 20% of the private returns to education. Does this mean that you think fully 80% of the value of time and other resources spent on education are wasted? Wow! Given how much we spend on education and how much time we spend in school that is an enormous inefficiency!  How is that possible?

Suppose you told me that only 20% of oil production was socially efficient, because the other 80% had large negative externalities - pollution, global warming, whatever.  Would "How is that possible?" be a reasonable reply?  I don't think so.  In both cases, the point is that individuals and firms are advancing their own interests by ignoring the large negative side effects of their behavior.

In the case of the education system: Students advance their interests by looking good compared to other people in the job market, employers advance their interests by hiring smart, determined candidates, and schools advance their interests by making their students look good to employers.  If this leads to a world where schools design meaningless hoops, students jump through these hoops, and employers financially reward the winners, why should students, employers, or schools worry if it's globally efficient?

The strongest argument for either of the two positions I'm defending is that if there was a much more efficient way to do whatever it is that education does, somewhere some society would have found it and we would all be trying to copy them.

Before you enlightened me about the manifest inefficiencies of mass unemployment, I might have said the same thing!  Unless you've had a major change of heart, you too admit that individual self-interest plus human psychology creates some awfully inefficient outcomes.

Here in the US we've now had at least 50 to 100+  years of experience with mass education and we haven't been able to find a better way? (I say 50 to 100+ since universal grammar school has been around much longer than cheap large scale higher education.)

At first glance, you've pointed out an important distinction between mass unemployment and signaling: mass unemployment doesn't last decades.  I'd reply, though, that the key inefficiency - nominal wage rigidity - has persisted for not just decades, but centuries. 

In any case, then kind of innovation you get depends on incentives.  If pollution is free, there's not much reason to think that innovation will solve the problem over any time horizon.  The same goes for signaling.  Students, firms, and schools get better at "playing the game," not reaching a global optimum.

...Bryan, I hope you appreciate the irony here. In our first public debate you are taking the side of widespread market failure while I'm defending the ability of private ingenuity to work its way around whatever problems the world throws in its way - the reverse of our usual roles. I must say I'm enjoying being in this position because I normally grant that it is the person arguing against efficiency who has the burden of proof.

It's not an ideologically comfortable position for me.  But first-hand observation of the education industry, plus the psychological evidence against the empirical importance of "learning how to learn," "training the mind," etc., leave me little choice.  On a deeper level, admittedly, I am taking a characteristically Caplanian position: Government is making a genuine market failure worse because voters systematically overestimate the social benefits of education.

Let me take a guess how you would shoulder this burden based on your response. Misguided government policies of subsidizing the sort of education that teachers like locks in a major competitive edge for the public education model. It's not worthwhile for businesses to set up their own screening mechanisms as long as the state is doing it for them and private education has to mimic public education for credibility. Further, to the extent that education is supported by charitable giving there still is little accountability for outcomes.

I agree that government subsidies are an important part of the problem.  I'd like to be able to say that government is the whole problem.  But that just seems false.  Long before government subsidized higher education significantly, schools designed meaningless hoops, students jumped through them, and employers rewarded them for it.

Finally, any attempt by the private sector to find a more efficient way of screening would be subject to adverse selection (you said only losers would apply to your "quickie tech"), and besides, anyone who went to an alternative school would be branded a weirdo.  I'm not buying it.

First off, if inefficiency is of the magnitude I believe you are claiming then public support of the system wouldn't be near enough to keep it in place.

Agreed.  Government makes a bad situation worse, but it's not the root of the whole evil.

So why doesn't the private sector respond? You seemed to imply that any attempt to do the screening cheaper would be subject to adverse selection (your indictment of Quickie Tech) and would also have reputation problems (graduates are weirdoes). I don't think it takes much imagination to get around the first problem and I don't think the second problem is real.

To begin with, you don't have to jump all the way to the most efficient possible system in one leap. A small innovation that gives you a slight advantage should be enough to allow a private institution to prosper and shouldn't cause a big adverse selection problem as so many factors go into college choice.

This sounds very reasonable.  But reread your last sentence carefully.  If the advantage is only "slight," then a slight adverse selection problem is probably all it takes to cancel out the benefit.  So there's not much reason to be optimistic about stepwise progress to a low-signaling world.


And if experimental schools are ruled out because they signal that the graduates are weirdoes you would hardly know it from the very large number of experimental schools of all sorts - private and state funded - at all levels of education... BTW, I'm one of those weirdoes - I got my BA at Bard College at a time when it was known as an "alternative" college. For every person who thinks you are weird for going to an alternative school there is someone who thinks you are interesting.

"Experimental" covers a wide range of experiments.  Bard College is experimental in some sense, but it still makes students jump through a bunch of meaningless hoops.  If you can handle four years of that, you'll probably make a good worker.  But experimental schools where you simply cut out hoops so their students can graduate sooner?  Employers might find their graduates "interesting," but they're more likely to dismiss them as slackers who took the easy way out.

But why not an instant jump to a much more efficient format? Instead of "Quickie Tech" why not "Elite Accelerated Tech."  "We teach you all the material in half the time. We hold our students to the highest standards, but if you've got what it takes you can make it through and demonstrate to the world that you are one of the elite." Since you think that most schools are way too soft on students shouldn't this work?

Actually, most colleges already offer this.  You're free to take heavy courseloads in order to finish your education more quickly.  It's a strong signal, too.  Unfortunately, you have to be super smart and determined to send it.

...Some of the very highest reputation private schools are known for being very forgiving of less than optimal performance and aren't very demanding in terms of the amount of work they expect students to do. That is a terrible way to run a screening system, but exactly how you want to run an education system! As you know, Bryan, I moonlighted as a flight instructor for several years. People most certainly don't know how to fly airplanes when they start flight training, but a few months to a year later they do. If you work for an airline you're going to be bounced out of service if you get an airplane into an unusual attitude. But if you do it to your flight instructor during training you will likely get a laugh and be told to try again (maybe with a little less left rudder). When you are teaching people you give them a lot of slack and then give them feedback on how well they are doing so that they can find out what works for them.

You're conflating two very different kinds of forgiveness.  Forgiveness toward students who take risks with big upsides and small downsides often leads to valuable results.  But forgiveness toward students who are disruptive or lazy doesn't.  (This point is largely orthogonal to signaling, but since you raised it and it's interesting, I wanted to respond).

As you know from my original post, what I think people are mainly learning in high school and college is work habits that allow them to function well in the sorts of jobs they will get. I think schools are more forgiving of less than optimal performance than work because they are teaching these characteristics. If they were sorting to the degree that you believe they are then they would be more demanding.

Why?  Schools let students send a wide range of signals because students have a wide range of signals to send.  Many students have the brains and determination to finish an English degree, but not a math degree.  Schools serve both, and employers reward them accordingly.

Are they effective? Robin Hanson has noted how businesses in countries with less well developed school systems complain about how undisciplined the workers are. This is often cited as one of the major reasons why labor in the developing world doesn't attract more investment despite wages that are a fraction of the developed world...

These countries have lots of other problems, too, Bill - poverty, pre-existing culture, etc.  Why assume that the school system is the weak link?

Also, why was the business community so interested in establishing the system of public education in the form it was set up (see Bowles and Gintis, Schooling in Capitalist America p36-44)?

When firms lobby for detailed policies that affect their specific industries, I'll accept the presumption that their means are well-tailored to their ends.  When firms lobby for broad social policy, though, there's no such presumption.  Policies that are good for business in general are a public good for any particular business.  So I figure the "business community" was largely engaged in public relations - trying to look good by voicing support for whatever people at the time saw as "right-thinking" school reform.


Your main indictment of my argument that education produces value was that work could better teach the sorts of skills that I was claiming schools taught. Back at you brother! The same could be said for screening. I would argue that screening could be done at much less cost, relative to education, in businesses... If the problem is signaling - getting people to reveal their true type - then businesses could offer deals where people start off receiving very low wages (or even paying the business tuition for training) while demonstrating their ability... Your argument that innovation is impossible because of adverse selection is untenable in the face of massive amounts of innovation in the educational system everywhere.

I never said that all educational innovation generates adverse selection.  You get adverse selection when you try to make education easier - when you reduce the number and difficulty of meaningless hoops students have to jump through.   Co-ed dorms and high-tech classrooms don't attract losers.  Cutting math and foreign language requirements do.  And as I said earlier, it only takes a slight adverse selection problem to cancel the benefits of a slightly more efficient curriculum.

By the way, if you believe that offering to work for free during a trial period is such a great way to overcome signaling, why don't you think it's a great way to overcome labor market rigidities?

...I don't have the time right now, but one of us should go through the courses taken by students in different majors and figure what fraction directly relevant to their work.

Yes, I've been interested in interviewing people about their transcripts for a few years.  Know of a massive grant I could get with minimal effort? :-)

...Personally I would say that close to 100% of the courses I took in college get used in my professional work (the research, not the teaching part).

For my education, I wouldn't go higher than 30%.  Your "100%" figure puzzles me greatly.  You told me that you used to believe in the signaling model.  If your first-hand experience didn't deceive you, what did?

So Bryan, ready to do some back tracking? I think you've got to at least admit that the educational system is a lot more efficient than you originally claimed. Otherwise I think you have gotten yourself out on a limb that won't support the weight.
I'm standing my ground.  I'm happy to admit that I've got a 10 percentage-point standard error around my "80% signaling" estimate.  But a priori arguments about how easily market forces would undermine signaling don't sway me.  I've spent too many years learning - and teaching - material that's all-but-useless in the real world.

Comments and Sharing

COMMENTS (22 to date)
Hyena writes:

If your narrative is true, why hasn't the market developed to pick up people who forgo school and head directly into employment. According to your narrative, there are thousand dollar bills laying on the ground everywhere thanks to the education system. I would think these would have been picked up by now.

Instead, we actually have precisely the opposite: firms led the charge in universalizing college education, they were even the drivers of the business school.

Josh Weil writes:


Reread the post. An individual's decision to attend university does not imply overall efficiency.

[Broken link fixed.--Econlib Ed.]

z writes:
Students advance their interests by looking good compared to other people in the job market...
This reminds me of an argument Tom Sowell made. Namely that requiring a college degree is a simple and reasonably efficient way for businesses to select the top applicants. As a result, the benefit to college grads is that the degree signals that their ability is likely higher than non-college grads.

If that model is true, then the global inefficiencies may be even worse than you describe because, as the government subsidizes education and more people get college degrees, those at the top have to get even more education in order to produce the desired signal.

Troy Camplin writes:

You might want to interview me after I get a job other than hotel night auditor. I'm pretty sure that my Ph.D. isn't doing a whole lot of signaling at this point. So far the best it has done me is get fellow Ph.D.'s to repond to my e-mails and a few invitations to write papers on spontaneous order theory (did I mention that the Ph.D. is in the humanities?)

Steve Sailer writes:

The concept of diminishing marginal returns is useful in thinking about the benefits of education. Offering universal free public elementary school helped make Massachusetts a better place to live in the 19th Century than, say, Mississippi. A society of general literacy is better for most people than a society without general literacy.

Extending free public schooling out through high school also had a return for society in general, but a diminishing one.

The same is true for subsidized mass college education, only even more so.

David Friedman writes:

My extended exchange with Robert Frank on my blog, some time back, touched (among other things) on negative externalities from schooling, and so is to some degree relevant to this discussion:

ajb writes:

I think that Bryan underestimates the most important role of education -- indoctrination. It works even when it fails. That is, it makes schools with successful indoctrination high status and paves the groundwork for our cosmopolitan elites.

Our elites are generally capitalistic but also hostile to traditional norms, especially with regards the family, religion, and patriotism. In that sense, Bryan is part of the elite. But the elite also wants PC, environmentalism, etc and respect for cognition and book learning over street smarts and physical prowess. They prefer bureaucracy to unfettered markets. They dislike the military. They despise traditional American notions of country and social solidarity.

All these things are promoted by schools and the current hierarchy glorifies these notions. At the same time, the low functioning schools give an outlet for the lower classes while imposing their costs on the lower middle classes and the struggling upwardly mobile. Moreover, the school system simultaneously gets to change the social norms while nominally claiming to be the inheritors of American tradition. They get to own the Constitution as they debase it.

For those at the top, what's not to like?

Roger Sweeny writes:

Government doesn't just subsidize schooling; it strongly penalizes lack of schooling.

Back in the 60s, the Duke Power Company used intelligence tests in hiring and promotion. Whites did better on the tests and so got hired and promoted more than blacks. In March, 1971, in Griggs v. Duke Power, the Supreme Court ruled that this was a violation of the Civil Rights Act of 1964. If the tests had a "disparate impact," they were presumed to be discriminatory. The only way to overcome the presumption was to "validate" the tests, to show to a very high degree of certainty that they closely predicted success on the job. Almost no test has been able to meet the Griggs test.

Even though degrees have a very strongly disparate impact, the Court did not find them illegal as a job qualification, and the courts since then have specifically allowed them. They are legally and socially acceptable ways of telling people "you aren't good enough."

Consider two employers with 20 identical positions to fill:

The first gives IQ tests to all applicants and takes the people with the 20 highest scores. They turn out to be 19 east asians and 1 hispanic. She is in a world of legal trouble.

The second advertises that a college degree is required. He ends up hiring 15 east asians and 5 whites; no blacks or hispanics with college degrees applied. He has no legal troubles.

Charles Peters, the founder and long-time editor of the Washington Monthly, was not a fan of educational credentials. He was also not a fan of pencil and paper tests for job applicants. In the late 60s and early 70s he tried to nurse an emerging “alternative testing” industry. Among other things, he ran a number of articles on people and companies who had developed different kinds of tests for various jobs.

But after Griggs came out, the infant industry quickly starved to death. “Validating” the tests to the Griggs standard would have been much too expensive and taken much too long.

Roger Sweeny writes:

[Much] shorter version of the previous post: there are thousand dollar bills on the sidewalk, but if you pick one up, you will be arrested.

Roger Sweeny writes:

Re: how much of a college education is used later. I recall reading recently that only half of college graduates are hired into jobs that have anything to do with their major. I suspect many of that half are like a friend of my daughter's. She graduated with a degree in drama and now works for a company that rents video equipment.

Rebecca Burlingame writes:

" too admit that individual self-interest plus human psychology creates some awfuly inefficient outcomes."

Only because too few have seriously considered individual self-interest and human psychology as having vital roles for reconsidering social/organizational structures in the first place.

Hyena writes:

@Josh Weil

I've read it, but every time this argument comes up, whoever makes it needs to address this issue better than Caplan has. His argument is essentially that much of the population--between 35% to 100%, depending on where this model of education ends--has been involved in a market failure.

Further, we have to believe that profit-seeking industries have not found a way to usefully exploit this fact and are actually part of a market failure whose mechanism, at the firm level, is difficult to discern. We have to believe that, in theory, there are vast arbitrage opportunities along with relatively simple, known ways of discovering and exploiting them yet (almost) no one has.

Because this entire line of reasoning is extremely salutary to anyone believing they have an above-average IQ, it is unusually subject to cognitive bias among that group. Since that overlaps with the group of people who go to college, that danger seems even greater. Finally, since most of us have forgotten what it was like to think before the present moment, we are likely more subject to availability bias; we're incredibly unlikely to properly estimate the cognitive impact of education because we're likely assuming our current cognitive condition is representative of our past.

At all times we have to be wary of any theory that is fodder for our baser nature as egocentric, short-sighted status-monkeys.

In short, we have very large, basic issues with the "mostly signaling" idea and very large biases which loom over the theory. That's not a good position in which to find oneself and it should make people incredibly wary of it. I'd argue that the presence of such massive bias-makers alone should start our default position very deep in the "education is not mostly signaling" camp. Possibly "burn the heretics" deep.

Hyena writes:

@Roger Sweeney

Because engineers are rarely hired in their field and business school graduates have a hard time finding entry level positions in the broadest possible category of the economy....

Peter Gerdes writes:

As a postdoc who spends a fair bit of time teaching calculus I have no doubt that most college coursework is pure signalling and perhaps economically inefficent but I think you massively underestimate the huge gains in utility provided by college.

Most significantly I would cite the social benefits. For instance the opportunity to make friends and explore romantic relationships that are not availible at home nor would be at short trade schools. Once out in the buisness world matters of propriety, lack of sorting by age/class and the superior efficency of small groups collude to prevent us from forming the kind of personal attachments that we know are one of the few things to boost long term happniess.

Also college performs an essential societal/cultural function for us. In a traditional society children largely grow up experiencing the same kind of social enviornments they are expected to navigate as adults. Yes, you grow up in a house filled with extended family but in traditional societies that doesn't change with adulthood or marriage. In modern western society the opposite is true.

Most parents work to ensure their children don't have the option of sleeping over at a romantic partner's place, going out drinking, staying out partying until the wee hours of the morning, or living on their own yet these are all activities we expect adults to be at least familiar with. College provides the protected enviornment for learning about the social norms and prohibitions in the arena of adult life and socialization without the harsh consequences of say ineptly hitting on a coworker. Moreover, LIKE MOST COMING OF AGES RITUALS COLLEGE MUST PROVIDE THE ILLUSION OF HARDSHIP AND ACCOMPLISHMENT. That alone almost requires all the make work and the myths about it making you a better person, it's our coming of age ritual.

Finally I would argue that while quite inefficient it's possible there is no more efficient alternative (and indeed government support is beneficial). The value society recieves from the small minority who make serious contributions to mathematics and the hard sciences is enormous and for this tiny fraction of individuals a substantial amount of training is required. Ideally we might only offer the training to those individuals but it's in everyone else's interest to pretend to be such a person.

Indeed, I would argue that it is the substantial dependence of research institutions on private donations and tuition payments that exacerbates the problem. This creates incentives for the college to collude with parents and students who aren't truly gifted in a useful area to pass themselves off as if they were. This then is what dillutes the useful work of research and educating future researchers at the university with all the hoops that exist solely for signaling.

Roger Sweeny writes:

"Further, we have to believe that profit-seeking industries have not found a way to usefully exploit [the fact that most college expenditure only buys inefficient signalling]. We have to believe that, in theory, there are vast arbitrage opportunities along with relatively simple, known ways of discovering and exploiting them yet (almost) no one has.

NO, no, no, no, no.

There are no vast arbitrage opportunities for firms because they don't retroactively pay their employees' tuitions. The potential employees do.

There may be other ways for firms to try to figure out who will be a good employee but those ways are generally unknown and/or illegal.

Speaking of biases, most everyone reading and commenting here has a degree. One would expect them to be too harsh on the theory that degrees are mostly inefficient signaling. That goes double for the people who work in the industry that confers those degrees.

Hyena writes:

@Roger Sweeney

First, those ways are not illegal, very much the opposite: the Federal government itself gives a civil service exam which is nothing more than a basic IQ test.

Second: your assertion is not enough. The obvious arbitrage opportunity is in identifying bright people who did not go to college or in offering lower wages in exchange for a better resume (Google does this, though tellingly, not for non-graduates). If education is mostly signaling, then both the potential student and employers can come to an agreement that replaces the education signal with an experience-based one and capture the savings.

Third: I would expect graduates to bemoan their credentials, not non-graduates. Once you have an education and the assets it provides, it is worth your time to consistently downplay the institution's role. It simply makes you look better in the same way as saying "oh, I don't practice much" makes you look like a musical genius.

ekon4341 writes:

Don´t forget the power of ideas over people. This applies both to the question of why alternative means of signaling aren´t used more widely, and also to the question regarding K-12 education that you bring up.

With K-12 education, stopping disruptive behavior requires the exercise of power over children, backed by the potential use of some form of force. This is viewed as authoritarian, and hence not acceptable, by the current educational establishment. Hence, a higher level of disruption / hazing, etc. is tolerated, as the alternative (authority / force) is seen as immoral and/or destructive.

About the use of alternatives to higher education, the notion that college provides substantial value is far more widespread among the HR departments of the US than the signaling hypothesis.

Lo Statuz writes:

Suppose a college degree signals some trait x that's hard to test
for quickly and cheaply, like conscientiousness. Students with
low x would tend not to graduate, which leads to a testable
prediction. If graduates and non-graduates look about the same
on tests that are easy and cheap to give, that's evidence for the
signaling hypothesis. But if they look different (for example,
if graduates have markedly higher IQs), that would be evidence

Roger Sweeny writes:

Lo Statuz,

College graduates will in general have markedly higher IQs than non-graduates. However, it is illegal--repeat, completely, totally, utterly illegal--to give an IQ test to job applicants and use that as one basis for deciding who to hire.

However, it is perfectly legal to use the possession of a college degree as a basis for deciding who to hire. The degree signals, not perfectly but fairly well, a higher IQ.

Anthony writes:
About the use of alternatives to higher education, the notion that college provides substantial value is far more widespread among the HR departments of the US than the signaling hypothesis.

You're misunderstanding the nature of signaling. Signaling may be wasteful, but it may also be simultaneously accurate. College degrees *do* signal something, with some acceptably high degree of accuracy. The problem is that the signal has gotten more and more expensive, without actually changing what the signal is, or its accuracy.

Lo Statuz writes:

@Roger Sweeny

Consider all freshmen who started at Stanford in 2000. If the
ones who graduated differ markedly from those who didn't graduate
on something that's easy and cheap to test for (like IQ), that
would be a bit of evidence against the signaling hypothesis.

In countries where it's legal for employers to IQ-test job
applicants, do they?

Roger Sweeny writes:

Lo Stasuz,

As Anthony writes, getting a degree may be inefficient and expensive ($200,000 for 4 years at Stanford!) as signalling but it may be fairly accurate. I would bet a significant amount that it is a fairly accurate signal. It signals IQ but it also signals goal direction, time management, an at least minimal abilty to read and write, etc.

I wish I knew what the situation is in other countries.

Comments for this entry have been closed
Return to top