Skip to content

Estimating the college wealth premium: Not so easy

Dale Lehman writes:

Emmons_Kent_Ricketts_College_Still_Worth_ItHere’s the article referenced on Marginal Revolution today. I thought it might be of interest and worth blogging about. It is quite thorough and fairly complex. The results are quite striking – and important. My big concern relates to a critical variable – financial literacy. On page 14 they claim that it is an exogenous variable to education. It is defined in Appendix A. I don’t quite see how it is clearly exogenous to education. I would think that college educated (and postgraduate educated) people are more likely to feel confident in their financial literacy. So, if including this “exogenous” variable eliminates most or all of the wealth effect of education, isn’t it possible that it is just measuring that effect and not controlling (I know you don’t like that word – perhaps we could say adjusting for) it?

My quick thought was that it’s a political thing. The linked post highlights the following claim:

Among non-Hispanic white family heads born in the 1980s, the college wealth premium is at a historic low; among all other races and ethnicities, it is statistically indistinguishable from zero [boldface added by Cowen].

Cowen is conservative and education is a liberal institution, hence he’s supportive of studies that disparage education. This is not to say that the claim is wrong, just that I’m guessing that the conclusion is motivating the post.

But let’s get to the statistical issue that Lehman raises. I agree that adjusting for “financial acumen” seems a bit of an all-else-equal fallacy.

A related question is whether we would expect the college wealth premium to be positive.

Just thinking off the top of my head—the last economics class I took was in 11th grade—I’d think that, if you adjust for enough things, the college wealth premium should be negative. To put it another way, if you decide not to go to college, you should get some money in return. It won’t work this way in practice, because not everyone has the opportunity to go to college—like so many other institutions on society’s escalator, college is often an institution allowing the rich to get richer—but if you adjust for enough things, presumably this would include adjusting for whatever it is that gives you the opportunity and inclination to go to college.

Ummm . . . let me step back for a moment. What’s the causal effect of college attendance on future wealth? Of course we’d expect it to be positive. You learn stuff, you make connections, you get a credential, this turns into $. On the other hand, supply-and-demand and all that: if college is such a great deal, colleges will start to increase the price until on the margin its expected value in terms of future wealth will be zero. Or even negative, if college is actually fun and thus a “consumption good” like bullfight tickets etc.

We could also think about this from a causal-inference perspective. Let’s compare potential outcomes: future wealth of a person who is assigned to go to college, compared to future wealth of a person who is assigned not go to college. But then it depends on the person. The causal effect is hard for me to think about, even in the abstract, in that the effect of the treatment would depend on how it is applied.

Leave a Reply