correlation, causation, and "gains" to education
Arnold Kling, on big structural shifts in the economy:

We are seeing fewer jobs where there is the external discipline of the time clock and the assembly line. The human robot that was once needed to lift and pound in a factory continues to be replaced by non-human robots.

Instead, we are seeing more jobs where internal discipline is required. I suspect that this explains some of the wage differential that shows up for college graduates. Graduating high school shows that you can submit to external discipline. Graduating college shows that you can operate under internal discipline.

There has been some theorizing — almost entirely by people with PhD's — that education, especially in college, doesn't produce more productive workers, but merely filters for them. Usually I've seen this presented in terms of "signaling"; employers prefer to hire the people who have been filtered, even though those same employees would be just as valuable if they hadn't gone to college. (If this is true, then you can think of college as a four-year job entrance exam.) What Kling suggests here is very closely related, but slightly different; he is referring not to the ability of these people to get hired by someone, but to the higher marginal productivity that employers are seeking in the first place.

More specifically, though, Kling is referring to a particular trait to which he argues the economic return has been increasing. This is new to me; the arguments I've seen before have tended not to address the growing spread in wages, while people arguing for a more fundamental benefit to education have tended to argue that those skills acquired in college are more valuable in an increasingly technological workplace. It seems just as reasonable — and perhaps more so — to suggest that the latent characteristics for which college filters might have that property as well.
- posted by dWj @ 11:45 AM