The Case Against Education: Why the Education System Is a Waste of Time and Money

by Bryan Caplan

Popular support for education subsidies rests on the same fallacy. The person who gets more education, gets a better job. It works; you see it plainly. Yet it does not follow that if everyone gets more education, everyone gets a better job. In the signaling model, subsidizing everyone’s schooling to improve our jobs is like urging everyone to stand up at a concert to improve our views. Both are “smart for one, dumb for all.”6

Now we’re up to three broad traits that education signals: intelligence, conscientiousness, and conformity. We could easily extend this list: education also signals a prosperous family, cosmopolitan attitudes, and fondness for foreign films.21 For a profit-maximizing employer, however, the extensions are a distraction. The road to academic success is paved with the trinity of intelligence, conscientiousness, and conformity.22 The stronger your academic record, the greater employers’ confidence you have the whole package.

Education signals a package of socially desirable strengths. People at the top of their class usually have the trifecta: intelligent, conscientious, and conformist.

Imagine this stark dilemma: you can have either a Princeton education without a diploma, or a Princeton diploma without an education. Which gets you further on the job market? For a human capital purist, the answer is obvious: four years of training are vastly preferable to a page of paper. But try saying that with a straight face. Sensible versions of the signaling model don’t imply the diploma is clearly preferable; after all, Princeton teaches some useful skills. But you need signaling to explain why choosing between an education and a diploma is a head-scratcher rather than a no-brainer.

My father, a Ph.D. in electrical engineering, routinely denied that “soft” majors pay. When I was growing up, he gave me the impression there were only two education/career tracks. Some students study engineering to become engineers; the rest study liberal arts to become taxi drivers.

Affirming the financial rewards of education often confuses critics of the signaling model. Isn’t it contradictory to claim the market rewards irrelevant education? No; a thousand times no. The signaling model’s whole purpose is to explain why education raises income more than job skills. Unless education has a larger effect on income than on job skills, there’s nothing for the model to explain. The case for the signaling model is strongest if students learn zero in school—and employers treat graduates like kings.

When pay spikes, so does education itself. “Finish your degree, then rest on your laurels” is the classic student strategy. One-third of the U.S. population spends 12 years in school, gets a high school diploma, then stops. Only 2% quit high school right after eleventh grade. One-seventh spends 16 years in school, gets a bachelor’s degree, then stops. Only 2% quit college right after their junior year.6 Signaling has an instant explanation for all the spikes. Why does pay spike for degree years? Because finishing sends employers a much better signal than quitting. Why does education spike for degree years? Because students run, walk, or crawl to grab the handsome cash prize they see just over the finish line.

Signaling weaves a contrary tale—one where malemployment reflects workers’ never-ending struggle to outshine each other. Picture the labor market as an arms race. Rising education automatically sparks credential inflation; as credentials proliferate, you must study harder and longer to convince employers to hire you. In an everyone-has-a-B.A. dystopia, an aspiring janitor might need a master’s in Janitorial Studies to land a job scrubbing toilets.30 When a B.A. bartender asks, “Why oh why can’t I get a better job?” signaling ruefully answers, “Because too many competing workers have even more impressive credentials than you do.” As two noted sociologists explain: In an over-qualified labor market, employers will fill the “highest” jobs with those who have the “highest” credentials. Since over-schooling means there are too many workers who are highly educated, some of these workers are necessarily allocated to “mid-level” jobs. This process is repeated for those with mid-level qualifications, where, since there are not enough mid-level jobs, many are forced to compete for low-level jobs.31

Malemployment is not mere “man bites dog” hype designed to terrify English majors’ parents. The amount of education you need to get a job really has risen more than the amount of education you need to do a job. Bartender, cashier, cook, janitor, security guard, and waiter are now common jobs for college grads. Education helps workers advance in almost any line of work—whether or not they tap their education on the job. Technology has made many mentally undemanding jobs like cashiering simpler than ever, but the market still pays educated cashiers a hefty premium.

Signaling is the only theory that explains the totality of these otherwise baffling facts. In our society, education is a seal of approval. Employers know it. Workers know it. As seals proliferate, workers need extra seals to upstage the competition. You’ll never apply most of what you study, but so what? Academic success opens doors. A dysfunctional game, but if you refuse to play, the labor market brands you a loser.

According to the pure human capital model, education lifts income by making you more productive. A worker gets more education; their productivity and income go up. A nation gets more education; its productivity and income go up. If human capital is the truth, the whole truth, and nothing but the truth, education is a path to individual and national prosperity: education makes the pie bigger, so every worker can enjoy a bigger slice.

According to the pure signaling model, education raises income by making you look more productive. A worker gets more education; their productivity stays the same, but their income goes up. A nation gets more education; its productivity and income stay the same. The personal and national effects diverge because signaling is a rat race. Only one worker can look like the Best Worker in the Country, and only 25% can look like the Best 25%. If signaling is the truth, the whole truth, and nothing but the truth, education is a path to individual prosperity and national stagnation: education fails to make the pie bigger, so bigger slices for some means smaller slices for others.58 Actual education lies between these extremes. Yet the polar cases of pure human capital and pure signaling highlight another strategy to nail down the human capital/signaling split. Step 1: Estimate the effect of a year of personal education on personal income. Step 2: Estimate the effect of a year of national education on national income. Step 3: Compare. Human capital’s share of the truth equals the national effect divided by the personal effect.59 The rest is signaling.

Economists are so eager to argue education is underrated they neglect a strong reason to think education is overrated: reverse causation. Instead of “When countries invest more in schooling, they get richer,” the real story could be, “When countries get richer, they consume more schooling.” Almost everyone buys reverse causation at the personal level. Why do the rich spend more money on fancy prep schools and bloated college tuition? Because the rich have more money to spend.

Socially speaking, however, the human capital/signaling split is all important. The closer we get to human capital purism, the more education benefits mankind. As signaling’s share rises, education’s social benefits fade. When we near the pure signaling pole, education becomes an incinerator that burns society’s money, time, and brains in a futile attempt to make everyone look special. When individuals weigh how to spend their

The point: When you weigh the value of education, knowing the benefits of education is not enough. You also need to know costs and timing. If you invest in education, how much comes back (or “returns”) to you—and how long must you wait to collect? Economists answer both questions with one number. They call that number the “rate of return to education” or just the “return to education.”

Compensation. We’ve already learned three vital facts about the effect of education on compensation. First, thanks to ability bias, education’s financial benefits—cash and noncash—are smaller than they look. A reasonable estimate is that only 55% of the apparent premium is genuine.11 Second, most of education’s financial benefits are sheepskin effects. In percentage terms, the last year of high school is worth about 3.4 times as much as a regular high school year, and the last year of college is worth about 6.7 times as much as a regular college year. Since sheepskin results for the master’s are sparse, I assume they match the bachelor’s ratio.12 Third, noncash benefits are a big deal. Compensation exceeds income by an average of 44%.13 While more educated workers earn higher benefits, the ratio of benefits to income mildly falls as education rises.

Unreflective researchers naturally overlook noncompletion because it falls far outside their personal experience. The researchers finished their degrees. So did almost everyone they personally know. How bad can attrition be? Dismal. Overall dropout or “noncompletion” rates are high at all levels of American education. About 25% of high school students fail to finish in four years. About 60% of full-time college students fail to finish in four years. Half of advanced degree students never finish at all.66

Incidentally, the marriage market is probably the strongest reason to pay for expensive private schools. Going to Harvard may not get you a better job but almost certainly puts you in an exclusive dating pool for life.

From compensation to productivity. The selfish return to education hinges on compensation: How much pay do you forfeit while you’re in school, and how much extra pay do you capture after you finish? The social return to education, in contrast, hinges on productivity: How much stuff does society forfeit while you’re in school, and how much extra stuff does society capture after you finish?

To calculate education’s social return, then, you must know why education raises pay and benefits. If education boosts compensation solely by raising worker productivity, society’s gain equals the worker’s gain. If education boosts compensation solely by revealing worker productivity, society gains far less. For most purposes, in fact, society gains zero. True, the economy is more productive—and society richer—when employers know which workers are good and which workers aren’t.3 Knowledge is wealth, so ranking students has social value. But once students have been properly ranked, the social value ends. Employer’s knowledge of worker quality would be essentially identical if everyone had one less degree, right? In economic jargon, the marginal social benefit of signaling is roughly zero even though its total social benefit is substantial.

From a selfish point of view, the pay you miss while you’re in school and the raise you get once you’re out of school are symmetric. The pay you miss is all selfish cost. The raise you get is all selfish benefit. From a social point of view, however, what counts is productivity, not pay. The social cost of school is the stuff you fail to produce. The social benefit of school is the extra stuff you learn to produce. If education’s payoff is 80% signaling, and a year of education raises annual earnings by $5,000, only $1,000 is a true gain to society. The other $4,000 is your reward for convincing employers they’ve been underestimating your value.

Employment. Suppose four years of college cut your unemployment risk from 9% to 4%. If education’s rewards are 80% signaling, this means 20% of college grads’ reduced risk of unemployment—one percentage point—stems from their enhanced productivity. The rest is zero-sum. Your college degree lowers your unemployment risk by raising the risk for your unadorned rivals.

From a social point of view, education’s effect on job satisfaction and happiness is even more questionable. Humans savor status—a high rank in the pecking order. In our society, status heavily depends on education.8 Unfortunately, status is, almost by definition, zero-sum.9 As society’s education rises, so does the education one needs to feel socially superior. The disturbing implication: even if education were a path to personal happiness, it could remain a dead end for social happiness.

Start with the full cost of public K–12. The per-student bill varies massively from state to state. In 2009–10, the latest available year, Utah spent $7,916. Washington, DC, tripled that, for a grand total of $23,816. The U.S. average was $12,136.17 This figure is all-inclusive, counting cost of instruction, support services, food, enterprise operations, capital outlays, and interest payments.18 When calculating the social cost of the typical high school student’s education, however, this number isn’t quite right.

What if a government program shears the risk of educational failure? All else equal, selfish and social returns rise. But “all else equal” is key. The cheapest and surest way to raise high school graduation rates is indiscriminately awarding diplomas to all. Before long, however, this infinitely merciful system erases the selfish and social gains of graduation: if everyone gets a diploma, no one does.

Unfortunately, this stirring sermon is wishful thinking. Chapter 4 already reviewed research on the national education premium.31 While the evidence is messy, education seemingly does less for countries than individuals. At the national level, it’s not clear that education increases living standards at all, much less that education makes countries’ living standards increase at a faster rate. If

Signaling, as usual, is the final wrinkle. Give an individual more education, and they get better offers so they’re more likely to want a job. Give everyone more education, and you ignite credential inflation. Implausible? Ponder this: in 1950, only 33% of adult males had finished high school, but male workforce participation was higher than today.40 Crime.

Because education’s powers of social transformation are galactically overrated. The observed gap between, say, dropouts and engineers, is only one term in what could be called the Educational Drake Equation. For workers, education’s social benefit equals the observed dropout-engineer gap, times the probability of successfully completing the education, times the fraction of the gap not due to preexisting ability differences, times the fraction of the gap not due to signaling.

A classic bumper sticker muses, “It will be a great day when our schools get all the money they need and the air force has to hold a bake sale to buy a bomber.” By most measures, this great day arrived in the United States long ago. The air force may not hold bake sales, but total education spending far surpasses total military spending. For the 2010–11 school year, education was 7.5% of the American economy, versus 4.7% for defense. Spending came to over $1.1 trillion on education, and a bit over $700 billion on defense. Schools overtook the military back in 1972 and sharply widened their lead after the Cold War.12

The signaling model highlights two desirable forms of educational austerity. The first: cutting fat from the curriculum. The second: cutting subsidies for tuition. Let’s explore both.

Since education is mostly signaling, however, the social justice catechism is wrong. Yes, awarding a full scholarship to one poor youth makes that individual better off by helping send a fine signal to the labor market. Awarding full scholarships to all poor youths, however, changes what educational signals mean—and leads more affluent competitors to pursue further education to keep their edge. The result, as we’ve seen, is credential inflation. As education rises, workers—including the poor—need more education to get the same job. Where’s the social justice in that?

This is a book about education policy, not political philosophy. My top policy recommendation—austerity—is less jarring for some political philosophies than others. Yet once you buy my central thesis—education has a low social return because it’s mostly signaling—almost every political philosophy urges cuts. Liberals and conservatives may decry each other as evil incarnate, but both affirm that wasting taxpayer dollars is bad.

When I explain the centrality of signaling, audiences often think I’m endorsing the technophiles’ story. This utterly misunderstands me. Education is not a bubble, but stable waste. As long as traditional education receives hundreds of billions of taxpayer dollars every year, the status quo will stand. Online education will slowly carve out a niche, but that is all.

Technophiles would have a compelling case if education’s sole function were teaching job skills. Online education has clear pedagogical advantages over traditional education. Coursera, Khan Academy, Marginal Revolution University, and their rivals hire the best teachers in the world.47 Students can learn at their own pace—pausing whenever they need to reflect, rewinding whenever they need review, fast forwarding as soon as they master the material. Anyone who’s lost can drop a level without looking like a loser. Anyone’s who’s bored can jump a level without looking like a nerd. Online education is an awesome way to build human capital. Unfortunately, students aren’t hungry for human capital. They’re hungry for signals. Why? Because the labor market mainly pays for credentials acquired, not skills learned. After years of hype, sophisticated technophiles are finally coming to terms with this fact. Kevin Carey of the New America Foundation explains the intellectual evolution: “Three years ago, technology was going to transform higher education. What happened?”

When you chew this bullet, however, it melts in your mouth. Collective folly is inherently plausible. Weigh the intellectual payoffs. When spending our personal resources, we retain a clear incentive to second-guess popular ideas. Consumers who discover a best-selling product is junk can unilaterally save their own money. Don’t like it? Don’t buy it. When spending collective resources, in contrast, second-guessing popular ideas is selfishly futile. Taxpayers who discover a beloved program is junk must pay their taxes like everyone else. Don’t like it? You’re only one person. Without the prominence and charisma to reverse public opinion, nay-saying won’t save you money but will make you enemies.

The moral: in politics, critical thinking is an act of charity. Objective truth has to beg for spare change to survive. Owing to these perverse incentives, almost any political idea that becomes popular tends to remain popular.49 Even if it’s false. Even if it’s always been false. Why would the false become popular in the first place? Because human beings don’t like expressing—or believing—ugly truths. Instead, we gravitate—in word and thought—to views that “sound good.” Psychologists call this Social Desirability Bias.50 “There’s no such thing as a stupid child” sounds better than “10% of children are stupid.” “We will win the War on Terror” sounds better than “There’s a 50% chance the War on Terror reduces terrorism, a 30% chance it makes no difference, and a 20% chance it makes terrorism worse.” Isn’t “what sounds good” occasionally true? Certainly. When skinny people ask “Am I fat?” the unbiased response is “No.” Social Desirability Bias distorts responses to “Am I fat?” because we want to say “No” to everyone. Social Desirability Bias takes many forms. The bald lie—declaring “I love you” when you feel no love—is only one. More typically, the bias works through intellectual laziness. It’s not a lie if you believe it—and if you avoid calm deliberation, you can believe almost anything. What does this have to do with education?

Education and Social Desirability Bias are a perfect fit. “Give a man a fish, and he’ll eat for a day; teach him to fish, and he’ll eat for a lifetime” sounds lovely. So does “In a modern society, every child needs the best possible education.” “Education is the most important investment we make in our children’s future” leaves a warm glow. “We have to ensure everyone who might benefit from college attends” is music to our ears. Such statements aren’t blatant falsehoods. But we’re inclined to hastily accept them regardless of their truth because they’re emotionally appealing. With intellectual security this lax, we should expect proeducation slogans to permeate our culture, whatever their merits. The only way to sort out the mess is to put crowd pleasing aside and crunch the numbers. How can Social Desirability Bias explain the ubiquity of proeducation sentiment? There are three top stories. One appeals to human universals. Underneath our diverse cultural baggage, we’re much alike.51 Homo sapiens around the world cotton...

A complementary account: calling any popular view a “fallacy” is socially undesirable—and the human mind is naturally prone to the fallacy of composition. Since education has pretty good selfish returns, humans hastily infer matching social returns; what’s true for the part must be true for the whole, right? Social Desirability Bias deters us from using our firsthand knowledge of the irrelevance of the curriculum to challenge this fallacy—even in the privacy of our own minds. The final story appeals to global elite culture. Non-Western elites straddle two worlds: Western elite culture, and their own traditional cultures. After Western elites fell in love with education in the nineteenth century, they won over Western masses and non-Western elites. Non-Western elites, in turn, gradually spread the gospel of education to their own cultures. The global ubiquity of proeducation sentiment is no more puzzling than the global ubiquity of Abrahamic religions. What’s so bad about Social Desirability Bias? It is the fountainhead of wasteful and counterproductive policies. You know the mechanism. Almost all governments ...

In any case, matching course content to job openings remains the most direct way to ballpark vocational ed’s signaling share. All classes prepare students for some job. Auto shop teaches students how to repair cars; history teaches students how to do history. From a signaling standpoint, the issue is always, “How often do students use the skills they learn?” Vocational ed stands out because it prepares students for common jobs. According to the Bureau of Labor Statistics, the United States has roughly 900,000 carpenters, 700,000 auto mechanics, and 400,000 plumbers. Classic college-prep classes like literature, foreign language, and history fall short because they prepare students for rare jobs. The whole U.S. employs only 129,000 writers, 64,000 translators, and 3,800 historians.16

What makes vocational ed’s social return so ample? Status is zero-sum; skill is not. Conventional education mostly helps students by raising their status, but average status cannot rise. Vocational education mostly helps students by building their skills—and average skill can rise. Why are social returns especially ample for Poor Students? Because vocational ed trains these crime-prone students for productive work without igniting severe credential inflation.17

The silliest objection is that businesses “exploit” our children, handing them a pittance for their toil. No one expects schools to pay their students; the training kids receive is payment enough. Why hold firms to a higher standard? College students ferociously compete for unpaid internships because training is valuable compensation—and total compensation, not cash alone, is what counts.22 In any case, if the young were really grossly underpaid, employing them would be extraordinarily profitable—and thanks to competition, few business models stay extraordinarily profitable for long.

Egalitarians picture college prep as a free lunch: anyone who fails academically can switch to the vocational track, so everyone might as well start with academics. This ignores the disturbing possibility that after academic students crash, they’ll be too embittered to learn a trade.36 When such students start on the academic route, they learn how to do zero jobs. When they start on the vocational route, in contrast, they probably learn how to do one job.

Let’s start with books. Consumer demand is shockingly low overall: Americans spend 0.2% of their income on all reading materials, barely more than $100 per family per year.12 Americans used to spend more on reading but never spent much: back in 1990, well before the rise of the web, reading absorbed 0.5% of the family budget.13 Today’s Americans spend about four times as much on tobacco and five times as much on alcohol as they do on reading.14 Within

In music, pop culture’s victory over high culture is even more decisive. The Three Tenors in Concert is the best-selling classical album ever.16 With twelve million copies sold, it does not even break into the top fifty albums of all time.17 Looking at overall sales, classical music is only 1.4% of the U.S. music market. Country is eight times as popular, and rock/pop over thirty times as popular.18 Classical does better globally but still only commands a 5% share of the world’s music marketplace. Well, at least it beats jazz.19

First, despite years of study, most adults are historically illiterate. Either they never learned basic history, or they swiftly forgot what they learned. If a world of historical ignorance is scary, you should be scared already, because that’s where we live. Second, humans’ ability to transfer knowledge from one domain to another—such as from history to policy—is poor. When Bush invaded Iraq, was he ignoring the lesson of the Vietnam War or heeding the lesson of the Korean War? So even if citizens knew the details of history, it is far from clear they’d fruitfully apply their knowledge.

Humans care far more about how their “social equals” are living today than what their teachers said or insinuated years ago. Consistent with this story, social class fully accounts for education’s effects on marriage and divorce in modern America.55 Staying in school gets individuals into the elite club, which in turn helps them get married and avoid divorce. When everyone stays in school longer, however, the elite club jacks up its membership requirements, leaving the prevalence of marriage unchanged.

Many idealists object that the Internet provides enlightenment only for those who seek it. They’re right, but petulant to ask for more. Enlightenment is a state of mind, not a skill—and state of mind, unlike skill, is easily faked. When schools require enlightenment, students predictably respond by feigning interest in ideas and culture, giving educators a false sense of accomplishment.

DEREK: Yea, because employers care more about a scrap of paper than what you can do. BRYAN: You’re getting ahead of me. When education raises your income, economists call that “the education premium.” Part of the premium exists because school makes you more productive. That’s the human capital share. The rest of the premium exists because school makes you look more productive. That’s the signaling share. The “human capital/signaling breakdown” is both shares, side by side; 70/30 means “70% human capital, 30% signaling.”

BRYAN: Federal, state, and local governments massively tilt the scales toward the status quo. Without hundreds of billions of annual subsidies, who knows what alternative worker certification systems would have arisen? ALAN: [dubious] So it’s all government’s fault?

BRYAN: I keep telling you it’s not all signaling. Eighty percent signaling is my best guess. That said, you’d have a strong point if education signaled intelligence alone. In the real world, however, education signals a package of desirable employee traits: intelligence, work ethic, and sheer conformity for starters.