Bryan Caplan  

Mixed Signals: Why Becker, Cowen, and Kling Should Reconsider the Signaling Model of Education

PRINT
Punk Songs for Classical Liber... College and Summer Camp...

Gary Becker, Tyler Cowen, and Arnold Kling have all recently criticized the signaling theory of education. If you haven't heard, the signaling theory says that to a significant extent, education does not increase workers' productivity. Instead, the fact that you obtain an education shows that you were more productive all along, which makes employers want to hire you.

Here's a simple thought experiment to illustrate the distinction. Which would do more for your career: A Princeton education, but no diploma, or a Princeton diploma, but no education?

All my personal experience (30 years in school!) tells me that there is a lot of truth in the signaling model. I don't use most of my schooling in my job, and I'm a professor, for God's sake! But would I have my current position if I had failed high school Spanish? No way.

I must admit, however, that Becker, Cowen, and Kling have even more personal experience with education than I do. What are they missing?

First, Becker:

I believe it [the signaling model] declined because economists began to realize that companies rather quickly discover the productivity of employees who went to college, whether a Harvard or a University of Phoenix. Before long, their pay adjusts to their productivity rather than to their education credentials.

There is a big equivocation here. Sure, employers eventually figure out how productive a worker is IF they hire him. But interviewing is expensive, and so is getting rid of disappointing workers. So it still makes sense to use credentials to make interviewing and hiring decisions: You save valuable time, and reduce the chance of hiring unproductive workers.

Second, Cowen:

If education is pure signaling, just give everyone a standardized test in seventh grade and then close up the schools.

This is a straw man. Even firm believers in the signaling model like myself grant that schools teach some useful skills. But more importantly, this objection only works against specific kinds of signaling. Yes, if all that school signals is IQ, then a test is a cheap substitute. But what if school signals conscientiousness and/or conformism? Think about it this way: Would you want to hire a high school drop-out with a 150 IQ? Probably not, because you'd immediately think "This guy had the brains to do anything. Why didn't he finish high school? What's wrong with him?!"

Tyler's defends an alternative, "self-image formation" model of education:

Men are born beasts. But education gives you a peer group, a self-image, and some skills as well. Getting an education is like becoming a Marine. Men need to be made into Marines. By choosing many years of education, you are telling yourself that you stand on one side of the social divide.

This is interesting as far as it goes. But jobs create self-image too - and probably self-images that are more conducive to being a productive worker. Schools instill the self-image of the speculative day-dreamer; jobs instill the self-image of the practical go-getter. Tyler anticipates this:

Of course apprenticeships can turn beasts into men, but apprenticeships also turn them into working-class men. You spend your childhood hanging out with other laborers. As society becomes wealthier, more parents are willing to spend on education rather than apprenticeships.

But this is mostly semantic. Working-class jobs have "apprenticeships" which promote a working-class self-image; middle-class jobs have "internships" which promote a middle-class self-image. Ending government subsidies for education wouldn't create a new working-class generation; it would lead businesses to massively expand the employment of interns to take advantage of the large pool of talented, young people who can't afford tuition.

Last, Kling:

As long as we're just speculating, let me suggest another hypothesis. Education is supposed to increase our ability to learn. It is not that we accumulate useful knowledge in school, but we build up the mental equivalent of muscles. But many people stop learning at some point in life, resulting in mental atrophy.

Teachers of useless information often seek refuge in the "learning how to learn" story. But the facts are not on their side. The psychological literature on "fade out" essentially finds that mental atrophy is the rule. You can temporarily raise IQ, but it "fades out" in a few years. Furthermore, the psychological literature on "transference" of learning from one area to another finds surprisingly weak evidence of it. Since education has a long-run effect on earnings, but only a short-run effect on learning ability, we've got to look elsewhere for answers.

To conclude:

Becker claims that the signaling model used to be popular: "The signaling interpretation of the benefits of going to college originated in the 1970's and had a run of a couple of decades, but is seldom mentioned any longer." Cowen claims that it's still popular: "Nerds will hate education and tend to embrace the signaling model... This is one reason why the signaling model is so popular in economics."

I wish! I've been defending the signaling model to other economists for 15 years, and always met fierce resistance. Frankly, I think that this resistance mostly stems from a failure of introspection. If you really compare what you learned in school to what you actually use in your job, the disconnect is too large to ignore.



TRACKBACKS (12 to date)
TrackBack URL: http://econlog.econlib.org/mt/mt-tb.cgi/452
The author at The Club for Growth Blog in a related article titled Monday's Daily News writes:
    The Silence of the Good News - Larry Kudlow, NRO Why Blunt Lost - Robert Novak, RealClearPolitics Income Relativism - Bill Steigerwald, Pittsburgh Tribune-Review Tax Rates and Revenues - Alan Reynolds, Cato Institute Bush’s ‘07 Budget Calls... [Tracked on February 6, 2006 10:12 AM]
The author at Innovation Online in a related article titled More on Signals writes:
    The nerd element of the blogosphere is abuzz with the controversy over whether education is anything more than signaling. Some economists argue that when you get education you don’t actually become a better worker, you simply signal to prospective em... [Tracked on February 10, 2006 11:02 AM]
The author at Newmark's Door in a related article titled Signaling, summer camp, and Alchian's model writes:
    Bryan Caplan thinks the signaling model of education is underrated. Arnold Kling and Robert Lawson prefer the summer camp model. Two years ago I briefly noted Armen Alchian's model of higher education. I found it intriguing then, but if it's [Tracked on February 14, 2006 5:55 AM]
The author at Ben Casnocha: The Blog in a related article titled Would You Rather Have a UC Berkeley Diploma and No Education, or Education and No Diploma? writes:
    There's a very interesting debate going on among economist bloggers on the signaling theory of education. Would you rather have a [insert prestigious college name here] diploma and no education from there, or the [prestigious college name] education wi... [Tracked on February 14, 2006 4:08 PM]
COMMENTS (16 to date)
Bill Millan writes:

I spent ten years as a recruiter and I agree that a degree is used as a screening device. Yes, I would rather start out with the Degree and no college than the college with no degree if I were looking for a entry level job. But I see this as "false alternatives."

The other thing the degree shows is character. Did you have the guts to stick to the program and complete it? Did you coast through with a C or did you work hard and get an A?

The key thing to hire is good character.

Silas Barta writes:

Why bother debating it with others? If you really think the signaling theory has more truth (as I think it does), you could follow the advice I gave the last time you brough up this topic:

Start up a school that screens for IQ, and basically puts students through hell -- difficult topics involving critical thinking, research, teamwork, and long hours, all without any grade inflation. It would be cheaper because you wouldn't necessarily need well-educated professors. That would be extremely powerful signaling. You might not even have to charge tuition -- if you could run it cheaply enough, just charge some fraction of the next 10 years of their income. (Please don't make insightful comments like "but someone could stay out of the labor force for ten years!".) Also, you might be able to pack four years' worth of standard signaling into two years, and people who only made it through one year would still prove something. If it failed, so does the signaling thesis. This is actually something I've wanted to try out and would if I could get investment.

There is, of course, the risk that it would go unaccredited and, if not shut down by the education cartel, would be vilified by a compliant media demonizing the commercialization and "unprofessionalism" of this cheaper alternative.

Ann R. writes:

I don't see the signaling theory practiced in the military medical system. Just because you have the certain diploma does not mean you are going to be any more productive. Granted, doctors in the military must have a certain level of training like all doctors to "signal" there ability to practice medicine. However, once in the military system the need to stay educationally competative ceases to exist; it doesn't matter if you were top of your class or if you barely made it through (which is a scary thought). Also, this example shows that pay does not always reflect productivity. Given the example of military doctors it doesn't matter how productive you may be, as long as there is a budget to pay your salary you have a job. Instead of excessive medical training one of the most important credentials military medical staff can gain from any sort of education is conformity, because at the end of the day that is what gets you a raise in rank. I have known some smart, amazing, productive doctors in the military medical system who have not risen in rank or pay grade over their inept, lazy peers because they practice medicine out of the bureaucratic box. So, unlike the business world the military system can get away with unproductive workers, so credentials, to a degree, are not important.

MBA student writes:

Education is only a signaling device. I have had this conversation with my mother (who is a theoretical mathematician and university professor at a top school) multiple times.

Two years ago, I decided to go back and pursue an MBA. I had the choice between top 10 schools with no financial aid and attending a ‘second tier’ school (but still top 25) with a full ride and a substantial stipend. I very studiously researched the quality of the second tier school’s faculty, and I read all of their research papers on NBER and JSTOR. I was very impressed by their caliber, and I thought that I was discovering a diamond in the rough.

My mother, however, warned me that ‘name’ trumps almost everything, especially when pursuing a degree as vocational as an MBA. Thinking of myself as a savvy businessman (and that I know everything), I chose the second tier school due purely to the economics.

And I regret it to this day. I have a 4.0 in a difficult double concentration. Recently, I had a Goldman Sachs recruiter tell me, “You obviously know your stuff, and you seem like you have the right personality for the job. Normally, I would push you along to the next round, but there is just no way I am going to vouch for somebody who goes to XYZ School of Business. There’s just no way.” Lehman Brothers gave me essentially the same story. The issue is not just with me, as I have heard this same story repeated over and over again from my business school colleagues.

The majority of top firms are staffed and managed by good ‘ole boy networks of Ivy alums, who have no vested interest in seeing that network change. It is not that Ivy League graduates are that smart; it is that they are smart enough to prevent the status quo from being changed once they graduate. They look out for other alums, who are deemed competent not by their capability, but by the fact that they went to a certain school. How can an education from a specific school, in this very common instance, be anything but a signaling device?

To put it another way, if education was not a signaling device, why would schools at all levels fret so much over their USA Today, WSJ, and US News rankings? All of the decent schools are already turning away more applicants than they admit. They have to be signaling to the people that demand their graduates that their product is better. Again, this is signaling.

David Thomson writes:

“The majority of top firms are staffed and managed by good ‘ole boy networks of Ivy alums, who have no vested interest in seeing that network change”

Are these bigoted Ivy League alums likely to be politically liberal? Is this a factor, or am I imagining things? Are there any studies to determine if my theory is possibly correct?

It is also viciously unfair to describe all soft science Ph.D. as intellectual sluts. However, this is often the case. In many of the so-called elite universities, one has to be a leftist whore to “earn” an advanced liberal arts degree. This is especially true if the individual hopes to become a tenured professor.

rakehell writes:

Good post, Bryan. I was really surprised that Becker was so dismissive of signaling and told him so on his blog.

A follow-up to MBA Student's comment: there's only so much that can be assessed in a few interviews, and even in a battery of tests that an employer might try to administer. It's much more cost effective to look at grades and school reputation. It's not really an old boys' network, it's just a sorting mechanism.

Silas: You're really just attempting to create a top notch private university, of which there are already plenty in the United States. Standardized tests screen for IQ (and they are generally the prime factor in admissions). The student loan system creates the financing. For some reason you want the degree to be shorter. Why just one or two years instead of four? And why wouldn't the professors need to be well-educated? There are many schools in this country where the students spend four years in a pressure cooker environment, or create that environment themselves by taking rigorous courses and getting good grades. Savvy employers know what to look for.

John P. writes:

MBA student, I think you're basically correct, and I would add another layer of interpretation to your view of fancy professional firms. I'm a partner in a law firm and am heavily involved in our recruiting efforts. Part of the reason (in my experience, the biggest reason) why the status of the credentialing institution is so important to the firm is that it is also used by clients as a signaling device. Potential clients look at the schools from which the firm's professionals graduated and use that as a proxy for the quality of the firm's work.

Also, in response to Bryan's question about why the signaling model meets with so much resistance: I suspect that most of the people with whom Bryan discusses this issue are fellow academics. Of course they're going to resist a theory that says that what they do doesn't matter, that all that matters is the school's seal on the degree.

Phil writes:

I have been a believer in the signalling model even before I knew there was a name for it.

As a university co-op student, I worked for a software firm that ignored education almost completely when hiring or interviewing. They simply looked for "signs of genius" on a resume, or gave a standardized test.

It worked very well.

Among the people I worked with were a genius university dropout and a genius high-school dropout.

My question: why don't more firms do this? Do young people have so rosy a view of education that they wouldn't go for it?

Silas Barta writes:

rakehell: I think you missed the point of my proposal. I *know* there already exist "powerful signalers". But these signalers (Harvard, etc.) are *expensive*. My proposal is to provide the same benefit but strip out all the unnecessary stuff so that it just provides signaling, and not the other stuff educators (wrongly) deem important. This would allow Caplan (or whoever did it) to run laps around traditional universities, since they'd be cheaper and get you better jobs. As for why to pack it into two years, if you studied economics you'd understand that this reduces the opportunity cost of going to college.

David Thomson writes:

“As a university co-op student, I worked for a software firm that ignored education almost completely when hiring or interviewing. They simply looked for "signs of genius" on a resume, or gave a standardized test.”

This software company may have been vulnerable to an affirmative action lawsuit. The ugly truth is that many companies don’t dare hire a lesser credentialled white man over a so-called minority candidate possessing a more advance degree regardless of how brilliant he might be.

“But these signalers (Harvard, etc.) are *expensive*”

Harvard is grossly overrated. Have we already forgotten that the mediocre John Kenneth Galbraith taught at this institution? A rational person should suspect a Harvard economic degree to be possibly fraudulent. It definitely not something to be automatically proud of.

daveg writes:

“The majority of top firms are staffed and managed by good ‘ole boy networks of Ivy alums, who have no vested interest in seeing that network change”

Are these bigoted Ivy League alums likely to be politically liberal? Is this a factor, or am I imagining things? Are there any studies to determine if my theory is possibly correct?

If this is the business school I doubt liberal politics is an issue. Other graduate schools perhaps, but not the business school.

rakehell writes:

Silas: I was being impish, because I've found that people who dislike school tend to dislike signaling, and thus often try to construct alternatives that are shorter, or less brutally sorting. For instance, they are big on apprenticeships, without proposing what mechanism will determine who will get which internships.

If you'd studied economics, you'd know that highly talented people can acquire signals more easily. That's a key insight of the signaling literature and it does indeed flow from the concept of opportunity cost.

"[Spence] argues that an individual of higher ability is able to accumulate educational credentials at lower cost. Education thus not only enhances human capital but also has a valuable information role for higher ability workers." --The New Palgrave

College is not at all expensive for people who have good grades and can get good scholarship offers. It's not expensive for people who can work part time and still ace their exams. It's not expensive for people who are willing to take out loans, but have a skill (such as mathematical ability) that means they'll be able to readily find work in high paying industries after they graduate. It's not expensive for people who consistently get good grades and test well, and thus in all likelihood get admitted to professional schools. It's not expensive for people who get admitted to top schools whose graduates typically earn top salaries.

And why wouldn't the professors need to be well-educated?

I suggest Personnel Economics for Managers, by Lazear; Modern Labor Economics, by Ehrenberg and Smith; and Economics, Organization, and Management, by Milgrom and Roberts.

Silas Barta writes:

rakehell: I'm sorry you were being impish. When you're ready to take this topic seriously, let me know.

Xellos writes:

--"College is not at all expensive for people who have good grades and can get good scholarship offers."

The key qualifer there is "and can get good scholarship offers." For the (ostensibly) middle-class white male with exceptional grades going into engineering, my experience was that scholarships were rather thin on the ground. Certainly not enough to take the really expensive down lower than just plain expensive.

As far as loans go, borrowing against potential future earnings can only take you so far. Especially when universities take the amount of loans you're eligble for into account when setting their tuition... At least, the impression I got from the process was more of a "our cost is the most we can get from you" rather than a "here's our price, can you meet it" feeling.

Rob writes:

I failed high school spanish and still got my PhD in economics.

David Foster writes:

MBA student...from the firms you mention, you are pursuing jobs in one of the most credential-conscious industries that exists. Why not try companies that actually make and sell stuff other than money? I think you will find them less credential-oriented. And, if you work in early stage-companies and do well, you could position yourself for a later job in venture capital.

Comments for this entry have been closed
Return to top