Arnold Kling  

Returns to Education, Again

PRINT
Amar Bhide on Bank Regulation... On Profits vs. Non-profits...

Tyler Cowen writes,


If you believe in the signaling theory, however, his marginal product is fairly low, much lower than the wage he will be paid. They will fire him. He'll come out a bit ahead, if he is not too demoralized, but within a few years he will be paid his marginal product.

The central problem with the signaling theory is that a worker's productivity is observable. Why pay a large wage premium to somebody indefinitely regardless of their actual productivity?

My views about how education raises wages:

1. Certainly for some people, education adds to skills. However, I believe that for many other people, it does not. My main complaint with those who favor expanding college education is that they do not measure carefully enough along the right margin.

2. Education adds some of it value due to credentialism. The hypothetical fake college degree will get you a better salary in government pretty much automatically. And, no, you won't get fired when it turns out that your productivity fails to match your credentials.


Comments and Sharing





COMMENTS (9 to date)
Hugh writes:

The productivity of a single worker is not really so very observable. A company may decide it needs an extra head in Marketing/HR and that person will stay there until the political winds change direction and someone decides to outsource or to cut back on marketing/HR.

OT (but only at the margin) I thought that Charles Murray made a good observation on the whole education debate: http://blog.american.com/author/cmurray/

steve writes:

I went to four years of school as an engineer but the classes pertaining to the actual work I do amounted to about a single semester. I was trained as a general engineer. But, after three or four years of work thats over and your pigeon holed as that specialty forever more. Would have been less trouble just to take the one semester.

My dad was also an engineer (in the 50s through 70s) and he tells me they used to use technicians with two year degrees to do engineering work after a few years on the job. Nobody does that anymore. What happened?

There is a lot I don't understand about college degrees.

Anonymous writes:
The central problem with the signaling theory is that a worker's productivity is observable. Why pay a large wage premium to somebody indefinitely regardless of their actual productivity?

The company will need to spend time training new employees. Hiring is a type of investment, and it takes time for the company to realize whether the new worker is good or bad employee. If the company realizes this too late, the employee will already have gained a lot of training and will have a higher MPL that justifies his wage-rate. Of course, had the company been able to observe characteristics from the beginning, the company might have seen that this employee would be a bad investment - but most likely they will realize too late, and then the training is a sunk cost.

Hugh writes:

Here's Charles Murray's comment in full:

The New York Times’s David Leonhardt has weighed in with a defense of college-for-all. Arnold Kling dissects some of its obvious flaws on econlog. But the larger problem is that Leonhardt misses the point. The choice should not be framed in terms of college or no college. Almost everyone needs more education after high school. The problem is a piece of paper called the bachelor’s degree that has become both the requirement for first class citizenship in this country (being “just a high school graduate” makes you distinctly second class) and at the same time has become meaningless as an indication of what you have learned. End the BA, stop requiring four years worth of courses, stop glorifying the residential campus, and create a post-high-school educational system that takes advantage of all the ways that technology offers to let high school graduates tailor their post-secondary education to what they need to realize their abilities. Forget about the percentage of people going to college and focus instead on how antiquated, inefficient, and punitive the BA system is.
Marc A Cohen writes:

The fact that education is almost pure signaling does not necessarily imply that it is sending an inaccurate signal. To complete the college degree, you really do have to have combined intelligence and diligence sufficient to do the work. The point is that you could get almost the same result by just giving everyone an IQ test and a personality test, for a lot less cost. Education has huge negative externalities because individuals spend enormous sums competing with each other to show who is smarter, demonstrating their relative ability-capital without actually increasing their absolute level.

Kevin Dick writes:

I agree w/Hugh. Anyone who has ever spent any time in large corporation knows that productivity for a lot of workers is highly _unobservable_.

I would add two points:

(1) if you believe in the Garrett Jones organizational capital hypothesis, you should be more likely to believe that productivity is less observable. There's no direct measurement of organizational capital.

(2) credentials create confirmation bias under uncertainty. If a Harvard grad gets a bad outcome, it's because the task was incredibly difficult. If a dropout gets a bad outcome, it's because he's an idiot.

Mike writes:

I'm with Hugh and Kevin.

I will also make this claim: In the right line of work (sales and marketing, anyone?) the skills necessary to convince a hacker to produce fake Harvard credentials are a strong signal of the possession of the qualities necessary for success in the job. Also, person smart enough to know how to produce the fake credentials should be smart enough to know where to put the fake credentials to their greatest advantage.

Chris writes:

There seems to be an assumption that only formal education is really education.

I don't think that anyone would really claim that education broadly defined doesn't add value - but you don't have to search very hard to find a slew of people that will tell you that they didn't really learn anything in college.

A college degree can help you get that first job - after that I don't see how to matters all that much unless your chosen career demands credentials.

Peter writes:

I would even go farther than that Chris. The assumption is that only formal education and certification in a academic degree program is really education. Auditing an entire bachelors doesn't provide you much in the intelligentsia bubble.

Comments for this entry have been closed
Return to top