Bryan Caplan  

A Brief Letter on Signaling

How Deserving Are the Poor? D... The Career Consequences of Fai...
I heard a rumor that a famous economist was asking about my book in progress, The Case Against Education.  So I sent him the following email:

I heard you were asking about me at the GMU dinner earlier this week.  I am indeed working on a book defending the empirical importance of the signaling model of education.  I'm happy to discuss my project at length, but here's the short version:

1. The vast majority of research on the return to education - including IVs, RTCs, etc. - does not empirically distinguish between human capital and signaling.  The better papers explicitly admit this.

2. Students spend a lot of time learning subjects irrelevant to almost all occupations (except, of course, teaching those very same irrelevant subjects).

3. Teachers often claim that they're "teaching their students how to think," but this goes against a hundred years of educational psychology's Transfer of Learning literature.

4. When education researchers measure actual learning, it's modest on average, and often zero.  And yet employers still pay a big premium to e.g. college students who've learned little or nothing.  The same goes for the return to college quality.  It doesn't seem to improve learning, but it substantially improves income.

5. There is a growing empirical literature using the El-SD (employer learning - statistical discrimination) approach to measure the effect of signaling.  It usually finds moderate signaling, at least for non-college grads.  It looks like you have to finish college to quickly get employers to reward you for measurable pre-existing skills.

6. The sheepskin literature finds large effects of merely finishing degrees.  They eventually fade out, but it takes 15-25 years.  This isn't iron-clad evidence for signaling (what would be?), but it's strongly supportive.

My book will also argue that ability bias is a much bigger problem than the David Card consensus will admit, and that the positive externalities of education are overrated.  So the social return to education turns out to be quite low.  In terms of policy implications, I'm going to argue for large cuts in government spending on education, and a lot more vocational education on the German model.

Comments and Sharing

COMMENTS (9 to date)
Milton Recht writes:

There are several employers who will hire high school grads, promote them up to a point and then require a college degree (or a 2 year associate degree) for further promotions. I believe Target for example follows this rule. If I am correct about their employment rules, then it is probably not the only company following a limited promotion policy for its employed non-college grads.

Since the company has direct observation of promoted HS grad employees, what further signal does a college degree have at that point beyond what was observed that motivated the lower level promotions?

David Youngberg writes:

I often encourage my students to do basic math without a calculator. The first reason is human capital (doing the math in your head is good for developing an intuition on if an answer makes sense; it combats calculator typos); the second reason is signaling (smart people can do math in their heads; great to show off at parties).

Someone familiar with some of the psychological literature on intelligence once told me that the brain is like a muscle: the harder you think now, the smarter you become later. So one could argue that "math without calculators" also makes students smarter in the long run.

Is this a signaling story (it's now easier to send signals that you're smart because you are smart) or a human capital story (it's easier to master both general and job-specific skills because you are smarter)? Can a line even be drawn between them?

Glen Smith writes:


A major factor to making the college signal model work is having a large population that disagrees with the signal model. If I can convince you I'm smart, you are a lot less likely to take advantage of me by playing on my stupidity. In other roles, you might grant me credibility that I otherwise would have to spend time proving.

GIVCO writes:

I agree with your policy prescriptions and have, through experience, determined that US News college rankings are poor proxies for many occupational requirements. But that's the government and hiring POV, what about an 18 year old? Will college provide hi useful capital?

How would you have researchers empirically measure human capital? Does human capital just mean the taxable income generated by an individual? That seems limited.

An average 18 year old is impulsive and lacks a clearly understood occupational objective. What else should an 18 year old do with their time? At college, 18 year olds spend more time doing other stuff than studying irrelevant subjects; is there any value to that other stuff?

Mark Brophy writes:

Can anyone point me to educational psychology's Transfer of Learning literature that shows that learning chess or piano well fails to teach students how to learn?

Mark Brophy writes:

GIVCO claims, "An average 18 year old is impulsive and lacks a clearly understood occupational objective." He asks, "What else should an 18 year old do with their time?"

I was 18 years old 30 years ago. I wanted to study computer hardware and software design during all my waking hours that I wasn't eating, and wanted to study no other subjects until I found a job programming or designing computers, and saved some of my earnings. This was impossible because neither my parents nor the colleges could foresee the immense value of computers.

If I were 18 years old today, I'd still want to spend all my time studying software, although hardware design is not as important as it was 30 years ago.

Never underestimate the brain of an 18 year old. Your son is probably smarter than you. If he is not, you may have failed in your parental duties.

Jayson Virissimo writes:

Bryan, it isn't too late to change the name of your upcoming book to the more precise The Case Against Schooling.

Bernie writes:

Have you thought about comparing the outcomes of the population who graduate college to the population of people who enlist in the military?

This would compare two groups who are both motivated and capable. The difference would be the degree.

Please ignore the "some college" grouping since spending any time in the military almost guarantees some college course completion.

My guess is that you will not find any difference between the two groups.

Jacob Felson writes:

I'm very interested to learn more about #3.

"Teachers often claim that they're "teaching their students how to think," but this goes against a hundred years of educational psychology's Transfer of Learning literature. "

What's the best / most efficient summary of this literature?

Comments for this entry have been closed
Return to top