Day 5: Formal education is a stupid waste of time and money
A series following my new class: The Economics of Higher Education
For the Friday of Week 1 I assigned some fun readings. Bryan Caplan is an economics professor at George Mason University, a libertarian, and a contrarian. I asked my students to read a few chapters of his “Formal Education is Stupid” book because I thought it will be sure to spark interesting conversation. And I was not wrong!
I don’t agree with many of the ideas in these readings, and that is exactly the point of assigning them.
Today’s readings
Bryan Caplan’s The Case Against Education: Why the Education System is a Waste of Time and Money
Introduction
Chapter 1: The Magic of Education
Chapter 2: The Puzzle is Real
Introduction
Economists often think of labor markets as ones that include a lot of asymmetric information. Workers know more about their own attributes than employers do, but employers know more about what they need from a worker than workers do; workers know how much effort they’re really willing to put into a job, but employers know more about job quality and how effort will or will not be rewarded with money or perks. Etc. Markets that include a lot of asymmetric information don’t usually work very well because nobody knows exactly what they are buying or selling. It’s then difficult to settle on the right price for each match.
Of course markets often find a way to mitigate the problems of asymmetric information and economists have been following along. In 2001, George Akerlof, Michael Spence, and Joseph Stiglitz won the Nobel prize for their work on such markets, Akerlof for his ideas about adverse selection, Spence for his ideas about signaling, and Stiglitz for his ideas about screening. Caplan thinks formal education is mostly about signaling.
Caplan’s main argument is that formal education is bad at inculcating actual knowledge and skills, but pretty good at separating out which people are intelligent, conscientious, and socially conforming. Those are the attributes that employers really want, he says. He doesn’t think that all education is signaling; he just thinks its mostly signaling. “Personally, I think the true fraction exceeds 50%. Probably more like 80%.”
Because education is mostly signaling, Caplan makes the argument that its ultimately pretty wasteful.
“If education merely certifies labor quality, society would be better off if we all got less. Think about it like this: A college degree now puts you in the top third of the education distribution, so employers who seek a top-third worker require this credential. Now imagine everyone with one fewer degree. In this world, employers in need of a top-third worker would require only a high-school diploma. The quality of labor would be certified about as accurately as now - at a cost savings of four years of school per person.”
“To be maximally blunt, we would be better off if education were less affordable. If subsidies for education were drastically reduced, many could no longer afford the education they now plan to get. If I am correct, however, this is no cause for alarm. It is precisely because education is so affordable that the labor market expects us to possess so much. Without subsidies, you would no longer need the education you can no longer afford.”
Conversation inspiring for sure! Onward.
Chapter 1: The Magic of Education
Caplan starts by pointing out that much of what formal education proposes to teach is not all that useful. “Why do English classes focus on literature and poetry instead of business and technical writing? Why do advanced math classes bother with proofs almost no student can follow?” It’s odd then that mastery of esoteric academic disciplines is so correlated with climbing the career ladder.
And here is Caplan’s primary argument: “despite the chasm between what students learn and what workers do, academic success is a strong signal of worker productivity. The labor market doesn’t pay you for the useless subjects you master; it pays you for the preexisting traits you reveal by mastering them.”
“Signaling models have three basic elements. First, there must be different types of people. Types could differ in intelligence, conscientiousness, conformity, whatever. Second, an individual’s type must be nonobvious. You can’t discover a person’s true work ethic at a glance. You certainly can’t ask, `How good is your work ethic?’ and expect candor. Third, types must visibly differ on average; in technical terms, `send a different signal.’ Deviations from average are okay. A signal doesn’t have to be definitive, just better than nothing. … Visible signals must be less costly for types in higher education. This cost could be measured in money or time. Or it could be purely emotional.” Caplan quotes Tyler Cowen as saying “`It does not suffice to give everyone a test and hire people with the highest scores… Doing well on a test is no guarantee of perseverance. The signal must be costly and grueling, otherwise it fails to sort out the best job candidates.’”
Caplan walks through a few vignettes showing why he thinks that education signals intelligence, contentiousness, and conformity and states why he thinks these are the three traits most employers look for: “An intelligent worker learns quickly and deeply. A conscientious worker labors until the job’s done right. A conformist worker obeys superiors and cooperates with teammates.” He then continues to argue why non-traditional signals of those same three traits don’t work as well as formal education. “Heterodox signals of your strengths, in contrast, automatically suggest offsetting weaknesses. Suppose you scored well on the SAT but never went to college. Employers will readily believe you are smart. But if you’re so smart, why didn’t you go to college? As long as your conscientiousness and conformity were in the normal range, finishing college would have been a snap. Once employers see your SATs, they naturally infer you’re below average in conscientiousness and conformity. The higher your scores, the more suspicious your missing diploma becomes.”
Caplan continues his argument by pointing out that when firms don’t like firing workers, signaling and credentials become even more important. Suppose a firm hires a new worker and then discovers they are a bad apple. If the firm doesn’t fire them, average productivity will fall, both because the bad apple is bad and because the bad apple may spoil the bunch, i.e. drag other workers down too, either with incompetence, too much chit chat, drama, or whatever. But if the firm does fire the bad apple, the manager doing the firing may be sad about having to do it. The other workers may be sad too. People make friends with their coworkers, even when those coworkers aren’t great workers. That could reduce productivity too. Firms will then be hyper focused on never hiring a worker that could turn out to be bad. That makes firms more risk averse and more interested in looking for signals that a job candidate is not bad quality. Note that I didn’t say signals of good quality. Firms do want to hire the best workers. But when they don’t like firing workers, firms are more interesting in NOT hiring bad workers than they are in hiring good workers. A college degree may not be the perfect signal of a very high quality worker, but it is a good signal that the graduate is not a bad worker.
I think these arguments are really interesting because they don’t seem to apply to a specific high profile labor market right now, and that makes me wonder what will happen to the labor market more generally.
Which market am I thinking of? Tech.
My husband is currently a Senior Software Engineering Manager at a tech startup. He’s been in tech for almost nine years. Is my husband a fresh faced 31 years old, having entered his first tech job straight out of an undergrad computer science degree? Why no, that is not his story. Does my husband even have a degree in computer science? No again. Does my husband have any formal training in software development? Nope, not even a little bit. What’s more, my husband tells me that when his current and previous startups were searching for new talent, they did not require a BA/BS in computer science even for junior developers. They did not look at coursework or grades for interns, and in fact, formal schooling in comp sci was sometimes viewed as a negative.
What makes the current market in tech so disinterested in signaling through college degrees, while the labor market Caplan describes depends on them? And will the relevant features of the market in tech soon apply to other parts of the labor market?
Could this be part of why the college wage premium is falling?
Chapter 2: The Puzzle is Real
This chapter reviews the educational psychology literature and presents several inconvenient facts:
Most subjects taught in high school and chosen as college majors are not directly relevant to jobs.
Most people don’t learn as much in class as you (or they) might think.
Students don’t retain much knowledge even a few weeks after class is over (think about the “summer learning loss” we hear about as parents) and they really don’t remember much 5 years later.
Many (most?) American adults are pretty ignorant about basic science, history, and civics facts. We should see this as the ceiling of what is possible for formal education to teach because most Americans both received formal schooling AND could have learned things on their own.
People are quite bad at transferring knowledge and concepts across domains and settings, even in highly contrived best case scenario experiments.
Before reading this chapter, I would have told you that my job as an educator is to help people learn how to learn, and I feel pretty good about that. Not specific facts, not what to think, not even how to think, but rather how to learn. None of the studies Caplan mentions measures learning speeds before and after receiving different types of education, so his chapter doesn’t directly speak to my professional raison d'être. However, he does make a good point when he says that if formal education is so measurably bad at teaching lower-level learning outcomes like facts and specific skills, why do I think formal education is good at teaching higher-level meta-skills like how to learn? How is that not just unproveable wishful thinking?
“Education seems to pay. Human capital purism advances a single explanation: education pays because education teaches lots of useful job skills. A tempting story… until you stare at what schools teach, what students learn, and what adults know. Then human capital purism looks not just overstated, but Orwellian. Most of what schools teach has no value in the labor market. Students fail to learn most of what they’re taught. Adults forget most of what they learn. When you mention these awkward facts, educators speak to you of miracles: studying anything makes you better at everything. Never mind educational psychologists’ century of research exposing these so-called miracles as soothing myths. An optimist could admittedly reframe my summary of the facts. If most of what schools teach has no value in the labor market, then some of what schools teach has value. … Fair enough. Yet the question remains: Can the modest job skills we learn in school explain the extra pay we earn after graduation?”
And I’ll add: Can it justify the expense?
Class discussion was fabulous today. The students were super engaged and did a great job connecting their ideas to the reading for today, to previous days’ readings, and to each other’s ideas.
As a hint of how fun our discussion would be, one of the student discussion leaders started us off with this perfectly selected clip.
There were some students who were strongly in favor of the idea that the value of education is mostly signaling. Others were strongly against. Interestingly, not very many people defended the purpose of education as building knowledge. (They did think that knowing some history was important for being a good citizen, but also believed Caplan’s stats that many Americans don’t know very much history. They said there’s a very funny TikTok trend right now that asks random people on the street about history facts.) They were more into the idea that the purpose of education was to build meta-skills like self-discipline, managing multiple commitments, attention span, sticking with things that are boring, teamwork, and dealing with difficult people. While they thought that some of that conscientiousness could have been innate, they firmly believe that college builds and hones those skills.
They were very strongly against reducing government subsidies for education. Education is extremely valuable for people who need a credential to combat discrimination, even if it is mostly signaling. Heck, especially if it’s signaling.
We also had a very interesting conversation about the value of talent versus discipline, and how different jobs/industries need different amounts of talent and discipline. Tech wants creativity and formal schooling can beat creativity out of you (their words, not mine), so it makes sense that tech wouldn’t always require a formal degree. But there are other jobs that require more discipline. This is especially true for industries that are exposed to legal liability: law, medicine, engineering. Discipline is also important in industries in which it’s difficult to fire people: teaching, college professors, and government workers. They liked Caplan’s argument that part of why firms like workers with college degrees is that firms like to mitigate the risk of hiring a bad worker, and I liked how they extended the idea of risk mitigation into the specific needs of different industries.
And that wraps up Week 1! Overall I thought today’s readings were a fantastic addition to my syllabus, especially for a discussion on a Friday. I will definitely assign them again next year.
On Day 6 (Monday of Week 2) we’ll pick up with peer effects in higher education.
Tomorrow’s readings
Winston, G. C. (1999). Subsidies, hierarchy and peers: The awkward economics of higher education. Journal of Economic Perspectives, 13(1), 13-36.
Carrell, S. E., Fullerton, R. L., & West, J. E. (2009). Does your cohort matter? Measuring peer effects in college achievement. Journal of Labor Economics, 27(3), 439-464.
I think Caplan is right that signaling really matters and we provide a very good signal, but he's wrong about the contribution of education to the process of actually making you more valuable to the workforce. One of the cool things about being in a leadership role is you contact that same students from recruiting and orientation through graduation, and the intellectual maturation is palpable. They are *different* people at the end, and in a way that is not hard to connect to the experiences they've had while in college.
I also think its intriguing that companies that produce software are fairly unique in their move toward identifying skills versus credentials. Are they more sophisticated? Or is coding a more easily assessed skill and/or a skill that is more easily self-taught ?
Commenting to note that these posts are the thing I most look forward to reading these days. Awesome stuff. I was wondering if you might, maybe toward the end, solicit and share with us some student feedback and reflection on the course? I’d love to know what they think when it’s all over. I’d also love to know what non-economics students would think and how it’d be different without that disciplinary lens but that’s obviously something else entirely. Thanks for teaching and sharing this course with us!