Caplan versus Hanushek
I watched the whole of the debate at the American Enterprise Institute between Bryan Caplan and Eric Hanushek and found it fascinating. It’s one of the best debates I’ve seen in a while: neither of the debaters used debating tricks, they focused on the issue they were supposed to be debating, and they brought a lot of relevant evidence and arguments to the debate. I add that both Bryan and Rick are friends. I’ve known Rick since we were colleagues in different parts of the campus at the University of Rochester in the late 1970s, and we are now colleagues at the Hoover Institution, and I’ve known Bryan since I was on sabbatical at George Mason University in 2007.
What follows is not a comprehensive comment but it does hit the things I found most interesting and/or compelling.
First, the debate topic was the following: “Is Education Worth It?” I thought the moderator, Nat Malkus, did an excellent job with his questions afterward.
Bryan leads and does quite a good job, hitting the parts of his book that are relevant to the topic, which means, essentially most of the parts his book. An excellent summary if you don’t have time to read the book.
Then Rick follows with his critique and some interesting data that contradict Bryan’s findings about the sheepskin effect.
I won’t post highlights from Bryan’s talk because he has virtually a verbatim transcript here. There is one highlight I want to note because I missed it the first time. I thought Bryan was saying that Rick would be a good czar of the education system, but I missed Bryan’s very important “If.” Bryan said, at about the 17:30 point, “In sum, if I had to hand over a trillion dollars of taxpayer money to one stellar researcher, I’d be sorely tempted to hand it to Eric Hanushek.” Also, of course, Bryan then goes on to say that even that would be a colossal waste of money.
Times below are approximate.
19:10: Rick argues that Bryan is surfing on two waves, the backlash against the idea that everyone should go to college and the “Trump wave” of spending less on academic education and more on vocational education. I think that’s true but the implication that I feared the audience would come to is that Bryan is saying these things because they help him surf these two waves. Anyone who knows Bryan knows that he steps to the beat of his own drum.
20:30: Rick says that he agrees that the subsidies to education should be cut substantially but that his grounds are distributional. In context, that means that subsidies to higher education are a net transfer from the poor to the rich. At least that’s what I took him to mean.
22:00: Here’s the first point that I thought Rick scored. Bryan’s proposal for early tracking and vocational training, to make sense, would require that we know where the jobs will be some years in the future, and we don’t. Ooh, I thought, good point. But wait until you see how Bryan answers in the Q&A.
23:00: Rick shows, using a 2.85 million-size data set from the U.S. Census, American Community Survey, that a year of high school gets a wage increase of 17% and that the sheepskin effect is only an additional 5%. This is quite at odds with Bryan’s data, but I read a large part of his book on an airplane last week and found his data impeccable. (A lot of them are in the footnotes and I never miss the footnotes.) But study and after study finds a much larger sheepskin effect than Rick finds. See how Bryan handles that in Q&A.
24:30: Rick shows a graph comparing his results to Bryan’s. Rick points out that Bryan’s data are from the General Social Survey.
26:00: Rick notes the overlap in wages between high-school graduates and college graduates. He attributes this to skills. Interesting graph.
27:00: Note Rick’s comments on the effect of cognitive skills on economic growth rates.
28:00: Note Rick’s mention of Raj Chetty’s work. That will come up in Q&A.
29:00: Rick notes that if you think that IQ differences account for skill differences, you would have to believe the following: If we compare Massachusetts to Mississippi, an 8th grader in Mississippi would be equivalent to a 5th grader in Massachusetts. We on the East Coast are tempted to believe that (he’s referring to the ugly stereotype about Southerners here) but he’s obviously saying that he doesn’t believe it. He makes the same point about people in 9th grade in the United States vs. people in the 6th grade in Singapore.
35:30: In Q&A, Rick mentions Griggs v. Duke Power as an important explanation for why employers can’t legally test employees to find out their cognitive skills. I had always thought this was a good criticism of Bryan’s work, but, as I learned in reading his book this weekend, Bryan has a slam dunk argument against this view on pages 88-90. He walks the reader through a beautiful numerate cost/benefit analysis.
36:00: Rick has what seems like a good argument about why signaling is less important than Bryan says–employers learn over time about who is good and who is not and act accordingly by promoting or firing–but Bryan makes the point that if it’s true that employers will easily fire people who don’t work out, then they would take a chance by hiring people without credentials.
38:00: Malkus gives Bryan a chance to respond to the point highlighted at 23:00 above.
39:00: Bryan gives his defense of the sheepskin effect but notes that if Rick’s numbers are good, he has a publishable paper because his result on the sheepskin effect is way smaller than that of anyone who has published on it in the last 25 years.
41:00: Nat asks a good question of Bryan: can we get kids to have the minimum skills while still cutting spending? Bryan answers that the two thirds who are not functionally illiterate and innumerate should be able to get those skills by the end of the 8th grade. I found Rick’s answer unpersuasive. See what you think.
43:30: Rick says that what stuns him is that Bryan seems to think central planning works: schools can decide early who gets voc ed and what kind of voc ed. This elaborates on his point in 22:00 above.
44:00: Bryan knocks it out of the park with his answer. He advocates exposing people to 20 different careers that they might reasonably have versus teaching them poetry or history or sports because they might be poets or historians or professional athletes. “It’s great to give people variety but give them variety that might work out for them rather than a bunch of pipe dreams.”
47:00: Bryan discusses Chetty’s work that shows that students who get better teachers get better cognitive skills. But the research is weird, says Bryan, because the skills fade within a few years. I didn’t quite understand his point about why Chetty’s own numbers show modest effects of good teachers on earnings. It had to do with multiplying some number by 1,200.
50:00: Bryan points out beautifully the opportunity cost for the society of more years of education: less time in which to earn income and huge spending, something that will become increasingly an issue as governments face huge deficits.
56:00: In his summation, which follows Bryan’s, Rick asks what does Bryan’s proposal–spend less–mean. Does it mean cutting salaries of teachers? Bryan doesn’t get to answer because he has already spoken. I know my answer. It means doing two things: quit subsidizing anything after 8th grade and reduce the compulsory schooling age to 14.
57:00: Rick says that even though there is much waste, spending on schools is a good investment.