Saturday, March 24, 2012

Strawmen are not very good sparring partners

I've been fascinated by some of the responses to my recent post on Caplan and education. Some of them are general comments, not directed at me. I'd still say they're strawmen of some sort because I don't think you see many people at all in the "economics of education" community saying them. But several commenters seem seriously confused about my point.

Caplan argues that the value of college education is substantially overstated, and that a lot of what we think of as the returns to education is signaling. I'm not aware of anyone that rejects the signaling model - they just might not think it's as important as Bryan does. And my point was this: Caplan suggests that a lot of education has no bearing on jobs, and that what a degree does is signal intelligence, conformity, conscientiousness, etc. We can add professionalism, self-motivation, forming and defending arguments, etc. My point is just that the curriculum that Caplan suggests has nothing to do with jobs provides or strongly reinforces these traits. I took one English class in college. The substance of the course has had no bearing on my career and it's not going to have any bearing on my career. But I did have to write a lot and form arguments in the class. Who cares if those arguments were about Hawthorne, Melville, and Faulkner? The point is I was writing a lot more in that class than in any economics class I took and I got a lot out of it. Yes, degrees signal skills. But they also provide skills, and simply asking "does this English class contribute to his job as an economist?" seems to me to be a terrible way of assessing whether college courses provide job skills.

The whole college experience provides skills too.

So my point is simply (1.) Caplan is not very convincing on course content, and (2.) he's taking the signaling model - which is unambiguously true in my opinion - too far. I think the signaling model primarily tells us why you see a lot more people getting four years of college education than three years. If it was simply a human capital story, you wouldn't see a big discontinuity at four years. But we do, because degrees as signals matter.

Somehow, a lot of people seemed to think that I was saying you can only get these skills through college. Of course not. Saying "X, therefore Y" in no way implies "Z, therefore not Y". Come on - you guys can do better than that.

Working right out of school or vocational/technical education can absolutely provide this as well (along with a lot of job-relevant skills that aren't taught in college). Why do you think I always share apprenticeship stuff on here and talk about Bob Lerman's research on it?

Here's my view of the economics of American education in a nutshell: The biggest problem we have with the distribution of human capital investment in the educational system is that it is bifurcated, and that has a big impact on the economic prospects of students. There are a lot of people that drop out or get sub-standard high school education, and there are a lot of people that go to college, but we have a gap in the formal provision mid-level skills (some of this is filled in on the job, of course). The biggest problem I see for higher educated workers is a demand-side problem, particularly for very highly educated workers. Jonathan Catalan provides a great summary of the argument for caring about demand-side rather than supply-side issues here.

At the lower end, though, I think we have supply-side problems in addition to any demand-side problems. We simply don't produce enough workers with vocational skills in this country. There are some positive trends. You see more career and technical education than you used to, the "college for all" mantra is slowly chipping away, and community college attendance is growing. These are all very good trends, but there's a lot left to be done. I think community college is especially important because of its versatility. It can provide mid-level skills that are closely connected to the skills the labor market is demanding, but it also provids an entry point to four year college for those students who decide that's right for them. So it's flexible in that sense. Career and technical education is important because it usually happens at the high school level, and it's very good at keeping students from dropping out and providing skills that are marketable. Finally, apprenticeship is important because it can take up drop outs later in their lives, and it provides earnings while making a human capital investment.

UPDATE: Oh - also worth mentioning - by mid-April Hal Salzman (Rutgers), Lindsay Lowell (Georgetown) and I are going to submit a proposal to the Sloan Foundation that's related to this - it should support one of my dissertation essays. There are several components, but the title of the proposal is "Science and Engineering Pathways in a Loosely Coupled System", with "loosely couple" referring to this Caplan point (and it's a point a lot of people make) about how people don't do work in their degree fields. We have specific focus on S&E people. My chunk of it is going to be looking at wage expectations and wage responsiveness of early career scientists and engineers. It's going to be a basic occupational choice approach, with somewhat more flexible specification of wage expectations than you often see to try to estimate discount factors that people use. I think it's also unique in that it's going to try to get at not just (for example) "life sciences occupation vs. non-life sciencs occupation wage differentials", but "life science major in a life sciences occupation vs. life sciences major in a non-life sciences occupation wage differentials" which is presumably the more relevant story. There's not a lot of data on both majors and occupations out there, so you don't see this as much - but the two most recent years of the ACS (and I assume future years too) does have this, which is great. The only crappy thing is that 2009/2010 is a horrible study period for doing a labor market analysis, but we'll see how it works out.


  1. Was my comment a strawman, Daniel?

    Cos' I hope it wasn't.

    But to be on topic, this reminds me of a discussion an English professor of mine had in the spring term of 2011 at Trinity College. (I was in a survey of English literature course from Alexander Pope to Samuel Beckett.) He was discussing how education is moving in the way that "practical" people wanted it to. In other words, these people, a la Jeremy Bentham, wanted something like a vocational education, and scorned the humanities.

    I'm firmly on your side of the argument, and not Caplan's. College educations do provide excellent skills for life, and writing essays about Faulkner, Hawthorne, and Melville, or Blake, Conrad, and Shelley, can make you a better person.

  2. "I think the signaling model primarily tells us why you see a lot more people getting four years of college education than three years. If it was simply a human capital story, you wouldn't see a big discontinuity at four years."

    That's interesting isn't it. In Britain University degrees are 3 years long for a Bachelors degree, so our discontinuity is at three years. But, American employers often accept UK degrees.


All anonymous comments will be deleted. Consistent pseudonyms are fine.