The Volokh Conspiracy
Mostly law professors | Sometimes contrarian | Often libertarian | Always independent
Bryan Caplan's Case Against Education
A new book raises fundamental questions about the value of education spending.
In his important new book The Case Against Education, economist Bryan Caplan argues that a high proportion of our massive spending on education is a waste of time and money. Since government spending on education amounts to some $1 trillion per year, this suggests we are gravely misallocating resources that could be better used elsewhere.
I. Caplan's Theory of Wasteful Educational Signaling.
Caplan doesn't deny the extensive evidence indicating that education increases income. High school graduates earn far more than dropouts, and college graduates far more than high school graduates, for example. Caplan contends that most of this income gain (perhaps as much as 80 percent) is not the result of improved skills (investments in human capital, in economic terminology), but rather "signaling." The big reason why employers prefer college graduates is not any skill those people acquired in college, but rather that getting a diploma is a signal of intelligence, conscientiousness, and conformity - characteristics that are valuable in most jobs.
Caplan provides a great deal of data and other evidence supporting his position. One of the more compelling is the "sheepskin effect": income gains from getting a diploma are far larger than those from earlier years of education that do not result in a diploma. If increases in human capital from learning course material were the main reason why education leads to higher higher income, the piece of paper you get at the end would not matter so much. Another notable line of evidence is the extensive data indicating that most students forget a high percentage of what they learn soon after the end of the class in question. Students generally cannot make use of knowledge acquired through education if they don't even remember it. From a signaling point of view, however, passing a course is still a useful indicator of intelligence and conscientiousness, even if the student promptly forgets what she learned, soon after the final exam. Caplan concedes that education in basic "reading, writing, and 'rithmetic" has a big payoff, and that the same is true for some types of vocational and professional education. But he concludes that signaling accounts for the lion's share of income gains from most other education.
If "signaling," not human capital, is the main explanation for the correlation between education and income, then much education spending turns out be a wasteful arms race. A high school diploma was an effective signal of worker quality in an era when many people did not have one. When almost everyone has a diploma, it is no longer enough to stand out from the crowd. Effective signalling now requires a college degree. As more and more people become college graduates, you now often have to have a graduate degree to stand out. And so on. Caplan argues that we might all be better off if we cut education expenditures, thereby curtailing the resources devoted to the arms race.
Caplan's analysis of the data is far from indisputable. For a good discussion of the issues, see this debate between Caplan and leading education economist Eric Hanushek. But even if you don't put as much stock in signaling theory as Caplan, he does compile a lot of evidence indicating that much education spending is wasteful, and particularly that it has very poor returns for weak students and many of the poor (two groups with a large overlap). Hanushek's own previous work suggests that increased education expenditures usually do little or nothing to improve educational outcomes. This bolsters the case for Caplan's argument that we can radically cut education spending, without losing much of value.
II. Does Education Make Us Better Citizens?
While much of the commentary on Caplan's book focuses on his analysis of the narrowly "economic" utility of education, he also has a section analyzing claims that education increases political knowledge. Historically, one of the most important defenses of public education spending is that it is necessary to make us better-informed voters. Even such libertarians as Milton Friedman argued that this justifies government subsidization of education (though Friedman argued it should take the form of vouchers rather than traditional public schools). Unfortunately, as Caplan documents, there is little evidence that the education system actually performs this role at all well. Over the last 75 years, educational attainment has greatly increased, with the average American adult today now having several years more formal education than that of 1960. Yet political knowledge levels have stagnated at roughly the same low level. As the 2016 election dramatically confirmed, widespread political ignorance remains a very serious problem, and massive increases in education spending over the last several decades have done little or nothing to alleviate it.
It's easy to say that the problem is simply that schools are teaching the wrong things. If we could only revamp the curriculum properly, maybe they could do a far better job of preparing students to be good voters. But incentivizing real-world politicians, school administrators, and teachers to improve civic education is a much tougher proposition than it may seem. The failure of several decades of efforts along those lines is one of several indications of that. Among other things, it is unlikely that incumbent politicians - who owe their election to largely ignorant voters - will want to alleviate the very ignorance that helped them get into power in the first place. There are also several other major obstacles to increasing political knowledge through education, which I discuss in more detail in my book Democracy and Political Ignorance.
What is true for increasing political knowledge is also largely true for the pursuit of educational enlightenment for its own sake. Many people argue that learning about and appreciating high culture is intrinsically valuable, regardless of its effect on your future job prospects. Caplan agrees. But, as he documents, the current education system provides such enlightenment to only a very small fraction of students.
As in the case of political knowledge, many people's immediate reaction to evidence that the education system provides little intrinsically valuable enlightenment is to say we should reform the system rather than cut spending. But improvement is extremely difficult, given the many perverse incentives in the system. And, as Caplan emphasizes, cutting wasteful spending is desirable even if you do not have an immediately viable way to achieve the same goal. Imagine you have been taking medicine to cure a skin rash. New scientific evidence proves that the remedy you have been using is actually snake oil that does nothing to cure the rash. In that event, the rational thing to do is stop buying snake oil - even if you don't have an alternative treatment that actually works. The resources wasted on snake oil can still be better used on other things, even if they are purposes unrelated to curing the rash. The same point applies to wasteful education expenditures that result in little or no useful learning. There is much to be gained from cutting them, and using the resources for more valuable purposes. Some of those other purposes may themselves be education-related, but others will not.
Caplan does, however, find two interesting political and social effects of education that may cut against his thesis. Controlling for many other variables, increased education makes people more economically conservative and more socially liberal - in other words, more libertarian. This should lead populists, especially those who are also social conservatives, to take a more negative view of education. But it should have the opposite effect on libertarians and other free market advocates, including Caplan himself. In addition (again, controlling for other variables) education increases marriage rates. Marriage is not the best option for everyone. But many studies suggest it increases happiness and correlates with a variety of beneficial social outcomes. Increased libertarianism and higher marriage rates do not prove that all of our gargantuan education spending is justified. But Caplan should give these effects greater weight than he does.
III. Why Caplan Overstates the Role of Educational Signaling.
My biggest reservation about Caplan's theory, however, has to do with the core of his analysis of signaling: ironically, for a libertarian economist, the theory relies on the notion that the employment market suffers from a gigantic market failure. Sending the "right" signals to employers turns out to be very costly if the only way to do it is to spend many additional years in the educational system during which you learn very little of actual value. Both employers and workers could get a huge payoff if they could figure out a cheaper signaling mechanism. Employers would have access to a valuable new pool of labor. Workers would be able to start their careers earlier, make more money, and carry less college debt.
As Caplan recognizes, most of the putative problem here cannot be explained by inefficient government subsidies to education. Although those subsidies reduce the cost to students and employers, the remaining costs borne by the students are still huge. Even those students who get completely subsidized tuition are rarely compensated for the foregone wages, time, and effort they lose during the additional years of education they take on largely for signaling reasons.
There is no inherent contradiction between being a libertarian (or even a libertarian economist) and recognizing the existence of market failure. Few serious scholars of any ideology believe that markets always function perfectly. But the theory at the heart of Caplan's argument posits a truly enormous market failure. And the logic behind it is not as compelling as it may seem.
Consider the three main traits Caplan believes education enables workers to signal to employers: intelligence, conscientiousness, and conformity. As Caplan recognizes, it isn't actually that hard to signal intelligence. For example, employees could submit standardized test scores to employers (which have a high correlation with intelligence). They could even take just one or two intellectually demanding courses and present evidence that they scored high on the final exam. It certainly does not require four years of college to prove your intelligence. Even if laws or regulations prevent employers from requiring applicants to submit test scores, nothing prevents the latter from submitting them of their own initiative.
Conscientiousness also is not that costly to signal. If you work at a demanding blue collar or service job for a few months, and get good reviews from your supervisor for customer service, showing up on time, and the like, that is pretty strong evidence of conscientiousness. Indeed, it might well be better evidence than getting a college diploma. Conscientiousness is better signaled by good work at boring and unpleasant tasks than at relatively interesting ones. If you do the latter well, it could just be because you enjoyed them, not because you are conscientious. At most four year colleges, students have considerable choice as to which courses to take, and can usually avoid those they especially dislike. By contrast, unpleasant drudgery is a major part of many service jobs. A year of good performance at McDonald's is probably a better signal of conscientiousness than a year of passing courses at most colleges. And, unlike college, McDonald's does not charge tuition and pays you a salary (even if a small one).
Even if, for some reason, people are averse to proving their conscientiousness by working at low-wage jobs, they could do so by taking one or two demanding courses and doing well on the final exam. Indeed, one can even imagine entrepreneurs developing and selling "conscientiousness certification" courses (a real entrepreneur would come up with a catchier name!) that fill this market niche. Such classes could be much cheaper and less time-consuming than college.
Caplan's response to arguments like this is to double down on the importance of conformity. Indeed, it turns out that conformity ends up doing the lion's share of the work in his signaling theory. People who find unconventional ways to signal intelligence or conscientiousness prove themselves to be nonconformists, almost by definition. And most employers don't want workers like that.
A certain degree of conformity is indeed important in most jobs. But it does not follow that employers want workers who are conformist across the board. Rather, as a general rule, they want people who are conformist in the sense of following instructions and "fitting in" with the other people at the firm. Across-the-board conformism that goes beyond this is as likely to be a detriment as an asset, particularly in jobs that require a degree of initiative or original thinking. If people say you are a conformist, it usually isn't a compliment!
As with intelligence and conscientiousness, there are likely to be cheaper ways to signal the types of conformity that are actually desired by employers than spending years on unnecessary formal education. Working at a relatively unpleasant job for a few months and getting good reviews from superiors can be a powerful signal of that kind of conformity.
When I was in middle school and high school, I did a lot of babysitting and lawn work. My generally effective performance of these jobs was a much better signal of conscientiousness and conformity than anything I did in college. Indeed, what could be a better signal of conscientiousness than the fact that people were willing to entrust their children to me?
The existence of relatively cheap ways to signal the traits that Caplan emphasizes weakens the claim that income gains from education are overwhelmingly due to signaling. It is difficult to believe that people devote such vast resources to a signaling method when there are far less costly and time-consuming ways to achieve the same result.
None of this proves that educational signaling is not an important phenomenon. Far from it. But it does suggest it may not be as omnipresent as Caplan argues. It may well be that human capital gains account for a significantly higher percentage of the income benefits of education than his analysis concludes. One possible mechanism by which it does so could be improvements in general cognitive skills, of the sort emphasized by Eric Hanushek in his debate with Caplan. At the same time, however, Caplan's book still provides powerful evidence that the education system wastes enormous amounts of money and that the payoff from it - especially for the poor and relatively weak students - is often unconscionably low. That should trouble even those readers who are not willing to endorse the more radical aspects of Caplan's thesis.
