Labor economists generally assume that education makes people better off financially because it teaches people to do things useful on the job. Bryan Caplan argues that this accounts for only a portion, maybe only 20 percent, of education’s payoff. Education pays primarily because it signals that you have the traits that make you a productive employee, not because it teaches you anything useful. In short, education is an indicator, not a cause.
This is no mere academic dispute. The skill-creation model implies that more education makes society better off. The more educated we are, the more skills we have, and the more we can produce. The model encourages the goal of getting as many people as possible as much education as possible. Government subsidies, grants, loans, even promises of free college for all, make sense on the skill-creation model.
The signaling model implies the opposite. The more educational credentials people have, the less each credential indicates about a person’s productivity. We suffer from an inflation of credentials, driving more people to seek more degrees—degrees they need to get their jobs but don’t need to do their jobs, at great expense of time and money to them and to the rest of society. This model implies that we already have too much education, and that government subsidies for education should be cut or eliminated.
Caplan marshals statistical and common-sense data to support the idea that education today follows the signaling model. First, common sense points about student behavior:
Students celebrate when teachers cancel classes or assignments.
People forget much of what they learn after a relatively short time.
Students are tempted to cheat; many do.
Anyone can attend classes or watch online videos of classes for free; few do.
If the point of education is to signal preexisting traits, this is predictable. Most students seek degrees, not skills, because that’s what will bring them rewards. Skills actually turn out to be mostly irrelevant. Some of what students learn will be useful to them later. But few students are in a position to know what small fraction that will be.
If the point of education is to acquire skills, however, none of this makes much sense, Caplan contends. Students should feel betrayed if classes or assignments are cancelled. They should work to keep their skills from fading. They should see cheating as pointless. They should flock to audit courses and watch lectures online. But, of course, weakness of will is pervasive; people drink, smoke, overeat, avoid exercise, and so on, but no one would conclude on that basis that those things aren’t harmful. And online education is still at an abecedarian stage; EdX, Coursera, Udacity, Khan Academy, Prager University, Jordan Peterson, and a host of others attract a broad audience. My own YouTube channel, with 45-minute-long philosophy lectures, has well over a million views. So, these arguments strike me as inconclusive.
The behavior of schools, Caplan argues, also supports the signaling model. It’s important to learn to read, write, and do basic mathematics, so it makes sense on both the skill creation and signaling models to teach those subjects. But why require subjects of little or no utility? One can quibble with Caplan’s list—he considers music, art, foreign languages, history, social science, and physical education mostly useless, and natural science of limited utility!—but most people will agree that they use little of what they learned in school. When was the last time you needed to know the capital of Malawi or the countries in the Triple Alliance? How long since you had to prove the Pythagorean Theorem or balance a chemical equation? When did you last play kickball or do gymnastics?
Nevertheless, it’s obvious that something is wrong with his analysis. The idea that Americans spend too much time learning history, geography, music, art, and science is absurd on its face. Caplan recognizes that education is more than job training, but he focuses on little else.
Caplan concedes that there is a difference between relevance to a career and relevance to a life. Music, art, Latin, history, and literature have done relatively little to increase my productivity on the job—though I have ended up teaching courses that cover literature, history, and the arts as well as philosophy, and have served on dissertation committees in the philosophy of music—but they’ve enriched my life immensely. Philosophy is probably the same for many of my students. Studying it may pay off financially; philosophy majors learn how to read carefully, analyze arguments, find weaknesses, and write effectively, all of which are important professional skills. But understanding yourself and the world around you enriches your life even if it doesn’t lead to a higher salary.
Caplan ignores completely an even more important function of education: it transmits a civilization from generation to generation, unifying a diverse society as it does so. The areas most critical to transmitting culture—music, art, literature, history, philosophy—are those on Caplan’s list of supposedly useless subjects. They may not be those most useful on the job, but they’re nevertheless among the most useful to society. They communicate an identity and an accompanying set of values. They train sensibilities, showing students what virtue, leadership, and, generally, excellence looks like. That teachers hostile to our civilization have seized control of those areas poses a deep problem. But the answer surely isn’t eliminating the areas altogether. That would be to surrender the task of building and transmitting our culture. It would, in short, be to surrender civilization and the pursuit of excellence.
Caplan’s response to these concerns is simple: it doesn’t work. People learn little about music, art, history, etc., and forget most of what they do learn. They continue to prefer Minaj to Mozart. My answer: it used to work. We succeeded in educating people in a shared culture when we tried. To paraphrase Chesterton, education in a common civilization hasn’t been tried and found wanting; it has been found difficult and left untried.
Caplan’s statistical arguments are powerful and more carefully developed. First, the “sheepskin effect”: earnings rise somewhat for each additional year of schooling. But getting a diploma or degree spikes the reward: the senior year of high school and college pays off far more than any other year. Since job-relevant learning isn’t concentrated accordingly, that suggests that the credential, not the learning that earned it, is what matters.
Second, malemployment: employers demand degrees not needed to do the job, and those hired may have even more irrelevant qualifications. Many jobs now require degrees they did not require a generation ago. High school dropouts, meanwhile, find opportunities scarce.
That isn’t because the jobs changed. In many cases, in fact, technology made them simpler. Positions that used to require an ability to do and understand complex calculations now require little more than knowledge of Excel. The explanation is that credentials became more commonplace. More credentials are required to set yourself apart.
There’s a real social cost associated with credentialism. Added costs in time and money are only a start. When society bases its decisions more on credentials rather than skills, the quality of the decisions it makes declines. Degrees substitute for performance. As Glenn Reynolds is fond of saying, the result is a ruling class that is credentialed, but not educated. Class divisions are exacerbated; inequality increases, mobility decreases. Meanwhile, competence declines, and respect for authority and supposed expertise erode.
Credential inflation is undoubtedly a problem, but it’s hard to see what to do about it. The signaling model implies that pursuing additional education is often wise for an individual. College graduates make 73 percent more than high school graduates, on average. Some of that results from having more ability rather than more education. Taking ability bias into account, Caplan argues, probably reduces the 73 percent figure to 40 percent. That’s still worthwhile, at least until student debt is added to the equation.
Whether college is a good investment, however, varies with ability level. Good students tend to graduate and see improvement in skills. Mediocre students are less likely to graduate and may see little improvement. Poor students tend to drop out. Caplan argues compellingly that high school is a good investment for almost everyone, but that college pays off only for good students. For mediocre students, it’s a bad investment. For poor students, it’s a terrible investment. Policies that encourage marginal students to go to college do neither them nor society any favors.
For society at large, he argues, college is almost always a bad investment. Signaling is a matter of standing out from the crowd. So, if everyone gains an additional credential, no one is better off, and society is a net loser. Status seeking is a zero-sum game, and the pursuit of those additional credentials is expensive. Caplan’s conclusion: “Schooling’s numerous social benefits pale before its staggering social cost” (187).
Caplan makes his assumptions explicit. The book has a remarkable feature: the spreadsheets underlying his analysis are all online. So interested readers can change the assumptions and see how that affects his conclusions. Would that all authors and economists were so transparent and invited such interaction!
What to do? Caplan suggests cutting fat from the curriculum. For him, the K-12 fat includes history, social science, art, music, foreign languages, and physical education. For reasons I’ve already mentioned, that proposal strikes me as disastrous. At the college level, he estimates that about 40 percent of students major in useless degrees. He outlines several options, ranging from the moderate to the extreme—from no longer requiring such subjects to raising standards to weed out most students to eliminating them altogether. Here, I would ask different questions: are these areas constructive or destructive? Do they build and convey a civilization? Or do they seek to undermine it? Do they rest on solid methodological foundations? Or are they driven by emotion, political ideologies, or intellectual fads?
Caplan proposes to eliminate subsidies, transferring the cost of education from society at large to parents and students at the college level and above. That would discourage people from pursuing education—in Caplan’s view, a good thing—and push people toward learning things that increase their productivity. Engineering will thrive even if people have to pay the full cost of schooling. Gender studies, sociology, French, anthropology, philosophy, global studies—maybe not so much.
Caplan’s critics charge that he ignores questions of equality. Education, they contend, is the path of promise for historically marginalized groups. But Caplan has a response: “subsidies raise the correlation between educational attainment and employability” (214). That means, in effect, that subsidies increase what members of marginalized groups have to do in order to be competitive. By producing credential inflation, they raise the bar, increasing the burden on the students they mean to help. They benefit the groups whose students are already the most able to compete, and harm the groups whose students are currently least able to compete. In short, whatever their intended purposes, they advantage the advantaged, and disadvantage the disadvantaged.
Caplan argues in favor of vocational education and apprenticeships instead. Unlike high school or college educations, they teach job skills. Apprentice plumbers learn plumbing, not hydrodynamics, plumbing theory, or “plumbing studies.” They learn what they need to do the job.
Caplan is right: “the average person shouldn’t go to college. Indeed, the average college student shouldn’t go to college” (285). The best of our students are outstanding, likely to succeed in whatever career they pursue. Their educations give them real skills. But they also came to the university having the intelligence and work ethic needed to thrive. The bottom half are academically weak, and learn little of any relevance to a career.
That said, I have a few doubts. Comparing salaries across degrees and majors is tricky. The timing of payoffs is different in different professions. It’s hard to get data showing the full picture. Philosophy majors lag many other majors shortly after graduation. But many philosophy majors go on to graduate study (most commonly, law school). It takes time for their professional careers to develop. Some fields are highly cyclical; I'm old enough to remember the joke, “What do you call a petroleum engineer in Houston? Waiter!” (Today the answer would be, “Boss!”) Other fields are more stable and offer more security.
A more basic doubt concerns the direction of causation. Maybe education acts as an indicator of desirable traits rather than a cause of those traits because it’s being done poorly. Contemporary university education is a shambles; it lacks the core needed to build character, develop sensibilities, and convey a common culture. The solution, I would argue, is to rebuild the core.
Caplan is right that our education system wastes time and money. Whether that’s avoidable is unclear. The signaling model implies that any signal, to be effective, must require sustained effort over a period of years. It would be nice if more of that effort were focused on what matters. But, absent a core, different things end up mattering for different students. To paraphrase John Wanamaker, we know much of the money we spend on education is wasted. The problem is, we don’t know which part.