About 75% of Americans favor more spending on education; about 5% favor less. In this thorough, hard-hitting book, Bryan Caplan, a professor of economics at George Mason University, makes the case that if we care about the social good, we should try to grow the 5%. The reason is simple: there is now a wide variety of evidence that the American education system fails to substantially build human capital. 

For many readers, the claim that our education system fails in this way, and Caplan’s evidence for it, will stand out the most on first read. It is a subsidiary claim though in the book’s overall argument. Caplan’s main thesis is a claim in applied labor economics: as he puts it, education’s substantial wage premium is mainly due to the signaling function of education. He devotes an opening chapter to the claim that even after adjusting for people’s abilities, education carries substantial increases in people’s wages (what economists call ‘wage premiums’). Caplan then argues that educational achievements produce these wage premiums mostly because they signal to employers that graduates are far more likely to be intelligent, conscientious conformists than non-graduates are. Educational achievements signal Caplan’s trifecta of intelligence, conscientiousness, and conformity – they certify who has these traits. But education does almost nothing to improve these traits. The education system, according to Caplan, is like a jeweler who accurately grades rocks and stones, but does very little to improve rocks and stones. This view is pitted against the human capital theory of education. It says that educational achievements carry a wage premium because education adds useful skills. Through various calculations, Caplan arrives at an 80/20 split: education’s value is 80% signaling and 20% skill improvement.  

Regarding the wage premium, Caplan estimates that high school graduates are paid on average 23% more than high school dropouts, college graduates are paid 73% more than high school graduates, and those who go on to earn master’s degrees are paid 122% more than high school graduates. However, these wage differences might be due to differing abilities between those who pursue, are accepted, and complete their diplomas and degrees, and those who don’t do these things, differences which are already in place before the relevant years of schooling even begin. In other words, since college selects for students with greater abilities, and college students are hardly a random sample of high school graduates, it is crucial to correct for these differences in pre-existing abilities. Here, and elsewhere in the book, Caplan clearly justifies what calculations need to be done and then looks for the relevant data. The estimate he lands on is that 45% of the education wage premium is due to differences in pre-existing abilities. 

The remaining 55% is what is split between signaling and human capital. Caplan offers a battery of arguments for assigning such a small share of this to improvements in human capital that are caused by education. One of these arguments, to get the flavor, concerns the so-called sheepskin effect. Take a student who doesn’t quite complete all of the requirements for a bachelor’s degree. Suppose she successfully completes 95% of them. The human capital theory predicts that she will get close to the full wage premium that the bachelor’s degree carries. But after correcting for differences in preexisting abilities, Caplan estimates that for each school year that a student successfully completes which is not a graduating year, she will earn an extra two to three thousand dollars per year. But, successfully completing the graduating year of high school pays about an extra $9,000 per year, the graduating year of a bachelor’s degree another extra $20,000 per year, and the graduating year of a master’s degree another extra $13,000. Our hypothetical student is going to earn an extra $6,000-$9,000 per year, but miss out on that extra $20,000 per year. This counts strongly against the human capital theory. 

The sheepskin effect supports Caplan’s favored version of the signaling theory. It says that possessing the degree is what sends the signal that someone is highly likely to have the trifecta of intelligence, conscientiousness, and conformity. These traits, according to Caplan, underlie success in both education and the workplace. So while our hypothetical student sends a good signal regarding her intelligence, she sends bad signals regarding conscientiousness (conscientious people follow through on big projects, something our student didn’t do) and conformity (conformists recognize and obey social norms about completing big projects, something our student didn’t do). 

Proponents of human capital theory who say that the human capital that education adds consists in mastery of content face trouble with the research which shows that students forget an enormous amount of what they learn in college. Nor does it help to repeat the mantra that education teaches students how to think, rather than what to think; nor does it help to claim that it teaches critical thinking skills. Studies on the transfer of knowledge show that very few people are good at transferring reasoning strategies to new domains without being explicitly told to do so. And other studies show that students’ critical thinking skills improve (at most) very slightly, even over several years of schooling. The results of these studies are dismal and depressing. Higher education doesn’t promote love of high culture like the music and literature that are the focus of some college classes, either. The actual consumption of Bach, Austin, and the rest set the upper limit on how much higher education instills a love of their works. As we know, actual consumption of their work is very low, and almost all college graduates prefer spectator sports and low brow pop culture. Caplan charts just how impressively and broadly ineffective higher education is. Again after correcting for preexisting traits, earning a bachelor’s degree has very little effect on people’s dispositions concerning voting, political affiliation, marriage, divorce, and religiosity.

As Caplan points out, all of this adds up to a nasty collective action problem. Education is “smart for one, dumb for all”. What is selfishly best for a given individual is to acquire diplomas and degrees, since doing so will vastly increase their lifetime earnings and give them a much bigger piece of the collective pie. But since education only barely builds human capital, it does very little to grow that pie. The $1 trillion dollars that Americans spend on education, roughly half of it public tax dollars and half of it private dollars, is an enormous waste. Education is mainly a hugely inefficient status game. Caplan argues that this result is quite robust. He helpfully adjusts the signaling to human capital split from the 80/20 split that he favors to 66/33, 50/50, and 33/66 splits, and then looks at what follows about the social (not the selfish) returns on education. Once signaling’s share reaches even half of education’s return, the social returns on excellent and good students sinks drastically and high school and higher education emerge as poor social investments. The social returns are even less for fair and poor students, though high school comes out as a good investment for male students who are academically weak since it significantly helps to curb socially expensive criminal behavior.  

Caplan offers two main remedies. First, we need to drastically cut the tax dollars that we spend on education, especially higher education and high school. This will reign in credential inflation and we can use this money for more socially valuable projects. And, second, we need to increase support for vocational training, both financially and with greater social esteem. 

If you’ve thought of a counterpoint or two, or maybe a dozen, while reading this, it is likely that Caplan recognizes and addresses those points. This is a thorough book that synthesizes an impressive amount of research in education, psychology, sociology, and economics. It is also a model of clear social science: Caplan is almost always clear about his assumptions and methods. It also bristles with humor. I highly recommend it to anyone who cares about education and is willing to move beyond our current underinformed positive thinking. You might even join and grow the 5%.

© 2019 Peter Murphy

Peter Murphy, Department of Philosophy and Religion, University of Indianapolis, murphyp@uindy.edu