a16z Podcast | The Case Against Education, From Signaling to Rainbow's End

Signaling and credential inflation -- not learning -- can explain why education pays in the labor market, and why we shouldn't invest (any more) in it, argues Bryan Caplan, economics professor at George Mason University and author of the book The Case Against Education: Why the Education System is a Waste of Time and Money. But is it really... a waste of time and money? Doesn't education have other benefits at least, like "learning to learn"; or sorting personality traits for employers at least; or helping developing economies even? And isn't it interesting that all the people (not just Caplan, but many in Silicon Valley and elsewhere) who argue against education are in fact, ahem, educationally credentialed themselves? This episode of the a16z Podcast, hosted by Marc Andreessen with Sonal Chokshi, takes on Caplan's "cynical idealist" take to probe both the cynical (problems, realities) and idealist (implications, solutions) aspects of education, no matter one's politics. And finally, where does tech (and a bit of sci-fi) come in??

Speaker names marked ? are inferred from speaking time and the show’s known hosts — they are not voice-verified.

• 0:00Sonal?

  Hi, everyone. Welcome to the A6 and Z Podcast. I'm Sonal. Today, Marc Andreessen is co-hosting one of our A6 and Z Podcast book conversations with Brian Kaplan, the author of the book titled "The Case Against Education" and subtitled, "Why the Education System is a Waste of Time and Money." Brian is a professor of economics at George Mason University and is also the author of other popular books, including "The Myth of the Rational Voter," which we also touch on towards the end of the episode. But we spend most of the episode talking about the themes of his latest book, which came out earlier this year, on the role of signaling in education and workplace hiring. And since we love how Brian is a self-professed cynical idealist, we decided to focus first on probing the cynical side of his arguments, especially

• 0:42Sonal?

  in terms of problems, and the rest of the episode on the idealist side in terms of implications and solutions, including tech that's already here and that may be sci-fi. But first, we asked Brian to summarize the key ideas of the book. That's what book podcasts are for after all.

• 0:56Speaker 01

  So, the key idea of the book is that there's really two different ways why education might pay in the labor market. One of them is the usual one, where you go to school, you learn some useful skills, and then employers like you more because you can do more stuff for them. But there's a second, totally different story, and this is that you go to school to impress employers. And even if what you learned in school will never come up on the job, employers still might prefer you and pay you more and give you a better job because you have convinced them through your education. Some of what you learn in school goes in the first category, literacy and numeracy. Those are useful job skills. But a lot of what you do seems more like the second thing, which economists call signaling. Looking better than other people is a great way to advance yourself, but it is not a way for

• 1:38Speaker 01

  society to advance. We can't all be richer if we all look better to other people. The result of this is what's called credential inflation, where you need more education to be considered worthy of employment. And then, the heart of the book is that while, selfishly speaking, it doesn't really matter exactly why education pays, from the point of view of what it makes sense for taxpayers to fund, it makes all the difference.

• 1:58Sonal?

  You left out the most controversial part of the book, which is that you're arguing for austerity as a policy solution in that we should not invest in education policy-wise anymore, because it may pay off for individuals, but the inputs are not clearly linked to outputs.

• 2:12Speaker 01

  Yeah, a release investment. If you really think that... The expansion to education is just leading to a lot of fruitless credential inflation. Is there any way that we could slow the credential inflation or stop it or reverse it? The most obvious solution there is educational austerity. If the reason why you need so much education to be worthy of employment is that education is so accessible, then if education were less accessible, you wouldn't need as much of it. The way that I often put it to my students is if I had my way, a lot of you couldn't afford to be here. That's the bad news. The good news is you wouldn't need to be here to get a job.

• 2:46Marc Andreessen?

  So one of the things I found really kind of entertaining about the book is when I was in sixth grade, I and everybody I knew believed basically everything you're saying, maybe from a completely naive standpoint. And of course, all the authority figures and experts, you know, including our own teachers assured us it wasn't true. You ran the whole gauntlet all the way through the system, come out the other side, and you're sort of coming back with the most sophisticated possible argument of the exact same points. So basically, the 12-year-olds are all basically correct.

• 3:09Speaker 01

  Well, here's the thing is that kids tend to think that either the schooling is not going to be needed in the future, you know, you're not going to use it and therefore it won't pay. Or, of course, there's the propaganda story of it's going to pay because it is useful. And the whole story that I put forward is each side has half the truth. It's not going to be useful, but that doesn't mean that it isn't going to be profitable. And that's really what the signaling model does is explain this paradox. How can it be that employers care so much about your acquisition of skills that they don't really care if you know how to do? And signaling says, well, if you go and do well in your Aristotle class, your foreign language class, this, though useless in terms of what you will do, it is impressive. It is convincing.

• 3:53Speaker 01

  It's a way of selling people on your value as a worker. And just to say, look, I'm really great at learning stuff. So, you know, I mean, the kids often are like, they say, like, when are we going to use this? But then the inference of therefore we can fail this and it won't affect our futures, that's what I say is incorrect. And that's where the adults are making a point, although they too are messing up half the story, unfortunately.

• 4:13Marc Andreessen?

  Or they have the conclusion correct, but they have the cause wrong.

• 4:16Speaker 01

  Yeah, yeah. There's a whole story that people tell of you come here, you learn all this great stuff, and then you're employable. Right, and you will then have a, let's see, I think it was the dad on Arrested Development says, "A happy life full of hard work." Alright, so that is the story that you're told, but it's partly true, which I think is where the plausibility comes from. And, I mean, I remember, you know, not just being in sixth grade, I think even in kindergarten, to say, "Why are we doing all this stuff? We're never gonna use this stuff, are we?" And I would go and ask adults, like, "When are we gonna use this?" And, you know, often they'd level with me and say, "Yeah, you're not gonna use it, but you need to do this to get a good job."

• 4:48Marc Andreessen?

  So then you basically do, like, agree with every 12-year-old, or to your point, maybe every precocious kindergartner of like, "Guess what? You really aren't going to use trigonometry."

• 4:54Speaker 01

  Yeah, exactly. I mean, I think overstatements say they all think this. There's always the goody-two-shoes who say, "Oh, no, no. Our teachers wouldn't teach us this unless it was really important."

• 5:04Marc Andreessen?

  But you also argue even a step even further, maybe darker than that, which is, you argue both that students aren't going to use the things they're taught. You also go a step further to argue, for the most part, they're not even learning the things that they're taught.

• 5:12Speaker 01

  Well, there's no reason for employers to pay you for things that you used to know but have forgotten since then. And there is a huge literature of psychology confirming the super-obvious point that people forget stuff that they aren't using. So what I do in the book is I try to get as many measures as I can of adult knowledge of subjects that are heavily taught in school for years. I've got data on knowledge of history, government, science, and foreign languages, in addition to literacy and numeracy. Literacy and numeracy, at least, there's a credible story, but the other things that I just mentioned, there, if you go and look at what adults know about those subjects on average, it is so close to zero. It is pretty shocking, which again, maybe means that they never learned in the first place. But again, it's also consistent with they knew it on the day of the final exam and have subsequently forgotten almost all of it.

• 5:56Speaker 01

  So basically, make up a list of things that seem so obvious you can't imagine anyone not knowing them. It's like, "How many senators each state have?" You might say, "Well, people have to know that." No, half don't know that. Right? Or, "In what century was the Civil War fought?" Questions of similar difficulty, Americans get about half of them right. So, you know, what this means is, even if you thought that the subjects were gonna be used on the job, it can't explain the earnings because people know next to nothing about the subject, so they just have not retained much, anyway.

• 6:25Sonal?

  But Brian, what do you make of the argument that it's not about the domain knowledge, but the tools for learning? I would argue that the one purpose of school is to provide people with tools for learning that they can then bring into other settings. It doesn't matter what the domain expertise is.

• 6:37Speaker 01

  So learning how to think, learning how to learn, or teaching critical thinking.

• 6:40Sonal?

  How to think, analyze.

• 6:42Speaker 01

  So what's interesting to me is these are all claims that economists will retreat to when you start pointing out how useless the subjects seem to be.

• 6:49Sonal?

  Mm-hmm.

• 6:50Speaker 01

  You know, economists who normally have as much contempt for psychology as a person can muster in scare quotes. They've been studying this issue for 100 years. And in the book, I review what educational psychologists have determined about this. And the main result is that most of this learning how to learn or learning how to think stuff is wishful thinking. Most of the time, when someone learns a subject, the most that they take away is exactly what they're taught. People are very bad at applying what they've learned, especially without external prompting. So there's a lot of experimental evidence where you'll teach, for example, either math and science, or science and math, and you'll sort of randomly assign which order you get, and then to see whether people apply something that is relevant to the other subject.

• 7:34Speaker 01

  And the usual result is people are terrible at applying this unless you give very heavy-handed instructions, like use the material that you learned in the previous section to solve the problems in this section. There's very little sign that people actually do apply it. There was one study that I cited where they had a big science class. I think these are actually science majors. They've been taught the scientific method. First, form a hypothesis, then come up with an experiment to test it, then change your mind depending upon it. And then they ask them questions like, so kids who get more sleep are getting better grades. Does this show that sleep causes good grades? And then instead of the students going and saying, well, actually, it could be a lot of things, and there are many different possibilities, and this seems totally unconvincing, and we need to design an experiment. Instead, even science students just give you a bunch of platitudes, like, well, it

• 8:19Speaker 01

  can't hurt to get good sleep. Why is it that you have not even considered applying a scientific method to this? And I've talked to biologists and said, so what do we really know about nutrition as an experimentalist? How do the practical problems with running an experiment on human nutrition really affect your judgment of what we know?

• 8:36Sonal?

  What do they say?

• 8:37Speaker 01

  The normal thing is just to go and repeat a bunch of stuff they heard on TV. And they still do it because for most people, there is this chasm between the classroom and the real world.

• 8:46Marc Andreessen?

  So Sotl and I both, by default, agree with a lot of the book. In that spirit, let me ask you kind of the harshest possible question, which is like, okay, you're a tenured professor of economics at a public university. You're not the only person who makes these arguments. My friend Peter Thiel makes very similar arguments. He himself is a highly accomplished philosophy major from Stanford, law degree from top law school from Stanford. There are other tech founders out here who make the same argument that have started tremendously successful companies. They all tend to be very highly credentialed, have gone to top schools, finished their degree programs and so forth. And then there's the fact and this correlation and not necessarily causation, but just a cursory look at the charts of income levels, unemployment levels and upward mobility is the college is the dominant correlation, at least just looking at the surface level data. And so, given that there aren't kind of obvious routes, in particular, other obvious routes

• 9:31Marc Andreessen?

  to upward mobility, couldn't the argument be made that collectively, you're telling people to do as you say, not as you do? And in that sense, you're kind of pulling up the sort of upward mobility ladder or the achievement ladder behind you?

• 9:41Speaker 01

  So, I mean, the first thing I would say about that is, you know, I could not be a professor unless I jumped through these hoops. So my behavior is totally consistent with the model that I'm telling you where if I had not gone through each separate layer of the educational system, I couldn't be a professor here. Professors have an especially credentialist and rigid hierarchy. Basically, the way that I would describe myself, I think of myself as a whistleblower. So if I were not a tenured professor, wouldn't people just think it was sour grapes and me complaining and feeling jealous of people that had actually been successful? And again, the main thing to realize is the whole idea of the signaling model is that it predicts everything you're predicting. It predicts that people with more credentials will make more money, get better jobs, have lower unemployment and have all these otherworldly gains. It's just an alternative theory of the mechanism, right?

• 10:24Marc Andreessen?

  So the other version of the harsh question, then, would be kind of the survivorship bias aspect that you alluded to, but let's just say there is an alternate quantum universe in which Brian Kaplan today is a motorcycle mechanic and we'd be having a very different interview right now where you'd be presumably saying, boy, I sure wish I had gone down this. Track of my parallel universe, Brian Kaplan, who did go down the track.

• 10:41Speaker 01

  Well, in terms of understanding, measuring the personal payoffs, for any one individual, you can always sit around wondering and speculating what would the alternative path have been. But this is again, why I mostly rely upon statistical evidence and say, look, you know, this is true on average. There's a whole bunch of adjustments that people have made to try to get a more realistic estimate of how much the true causal effective education on success is. But even after you throw the whole kitchen sink in, every possible adjustment anyone's ever thought of, there still seems to be this lingering and substantial labor market reward of education that is unexplained by anything other than you've got the credential. And that's where you need to say, all right, well, does this show that you learned something useful? And as Sonal was pointing out, maybe it doesn't really matter what you studied.

• 11:25Speaker 01

  Maybe what just matters is that you exercise your mind and built up your intellect. It put a lot of weight on this transfer of learning research because the people doing it didn't want to get the results. People who do educational psychology, they want to find that education is wonderful. This is one of my heuristics, by the way, is you're putting more weight on results people don't want to find.

• 11:45Marc Andreessen?

  So you highlight in the book, the underlying personality traits that are predictive of success in education, which is relevant because then those are the things presumably education is signaling employers about, right? And so it seems like you highlight three, which I think you describe as intelligence, conscientiousness and conformity?

• 11:58Speaker 01

  I usually say work ethic because conformity is often seen as being part of conscientiousness. So it's just a little bit more specific, but yeah, close enough.

• 12:05Marc Andreessen?

  Then in terms of psychology, so in the field of psychometrics, I think you translate this into IQ trait conscientiousness and trait agreeableness. Is that right?

• 12:13Speaker 01

  Not really. I mean, here's the thing is that there's a pretty limited body of work that uses standard personality tests to project income, and they usually have a lot of trouble actually getting very much, which I will say is something that kind of bothers me. I mean, intelligence tests are quite predictive. But personality tests, at least for labor market outcomes, don't seem all that predictive. And then you are sort of left with one possibility is personality doesn't matter and everybody's wrong. But employers so strongly say that they care about these attitudinal traits. But again, like the simplest story is just that the tests are not very good measures of the traits, at least when there's some incentives on the line.

• 12:51Marc Andreessen?

  Everything you've been talking about is sort of measuring these traits and then applying them to success in the workplace. But let's back up a second. My understanding is measuring these traits are very good predictors of performance in the educational system. Is that right?

• 13:02Speaker 01

  So again, so intelligence, definitely for the conscientiousness, at least for the measured conscientiousness, I think it does work a bit better within the educational system, especially perhaps if you're doing it on younger kids who just aren't trying to impress anyone. Maybe they actually like flaunting their antisocial traits to some degree. But I don't remember seeing like a nice, thick stack of evidence that lays it out solidly. I know that there's something called the Handbook of Employee Selection where they do put conscientiousness as the second best thing they've got, although I think acknowledging that it's just a lot worse than the first one, which is intelligence.

• 13:33Marc Andreessen?

  So the question that I'm heading towards is if employers are in fact looking for these traits, even if these traits are predictive of success in the workplace, if it's what employers are hiring for, and then because they for some reason feel like they can't or won't or don't want to run those tests directly, by the way, up to and including legal reasons, then to that extent, the argument gets made that they're then basically outsourcing that testing to the educational system. And the educational system is sort of a de facto intelligence conscientiousness test. And so if you follow this logic all the way through, could the signaling effect of the education system, could you simply at some point substitute an IQ and conscientiousness test and get the same signaling benefit without the four years?

• 14:08Speaker 01

  So the intelligence part seems very clear that you could, which to me shows that there's got to be something much bigger going on in education than just signaling intelligence, because there is such a quick and cheap way of getting at least a quite accurate measure. The problem with hiring based upon personality tests is they're very easy to fake. Fake being dumb, that's not hard, but fake being smart, that's hard. On the other hand, fake being conscientious on a test, that's something where anyone can do it. So there's that problem, you know, just at the level of the administration. But then there's the issue, well, maybe we could at least get over a lot of this by saying if you've got a good enough IQ, then we'll hire you without the education. Right, so there I say there is what economists call an adverse selection problem, which is there's someone who's really smart but did poorly in school, what do you think

• 14:51Speaker 01

  about their work ethic? What do you think about their conformity? Usual view is it's not just that they're average, they're probably really below average because it's so easy for a person that's smart to do well in school that you really have to wonder, why? Why did you not just have to go and do this alternate path? Now, you know, there's a lot of economists who are under the impression that this Supreme Court case from around 1970, the Griggs case, makes it illegal to use IQ testing for hiring purposes. I did a lot of research around this and really changed my mind on it. So the main thing is that if you look at the original case, it actually said that not only is IQ testing legally suspect, but educational credential hiring legally suspect. Oh, wow. So it's like, hmm, it's somewhat legally frowned upon, but still a lot of people do it.

• 15:36Marc Andreessen?

  But we live in a world in which IQ has become a very politically, say, controversial topic. I'm not aware of anybody out here who does an IQ test in the hiring pipeline. I think anybody who did would probably find themselves under tremendous firestorm of controversy. And the sort of cynical view would be therefore what's happening is the employers, if you believe everything else you're saying, then what you say is basically the employers are outsourcing the IQ test to the educational institution, which by the way is front-ended by an SAT test, which is basically a proxy for an IQ test. And so the cynical view would be this entire exercise is basically one gigantic sort of bloated IQ test. They could be shortcutted, and if the politics were to ever change, you could sort of collapse a lot of what is the current education system down to that test. Would you take it that far? Or not?

• 16:13Speaker 01

  Yeah, so I love cynical theories in general, but I think this particular cynical story is wrong. I mean, so for one thing, if that story was right, then people could hire based upon college admission letters rather than graduation.

• 16:22Marc Andreessen?

  So there is a phenomenon in Silicon Valley where if you get accepted at Harvard and you drop out on day one, you are much more attractive.

• 16:28Speaker 01

  Ah, you're much more attractive, but although still probably not as attractive as a graduate, I'm thinking.

• 16:32Marc Andreessen?

  Well, it depends. For entrepreneurs, possibly, possibly actually more attractive. Could be a special case. The next question I was going to ask, which is directly on the agreeableness or conformity point. Which is, as you said, it seems like if you ask employers and if you'd kind of observed them, it seems like they want conformant behavior, especially in large organizations. However, I'm sure you've seen the research, over time, disagreeable people, less agreeable people seem to earn more money.

• 16:53Speaker 01

  Mm-hmm.

• 16:54Marc Andreessen?

  Do you have a, do you have a sense of like, how to reconcile that?

• 16:56Speaker 01

  So, my reading in the Pyrrhax is that that effect is very small, and if you slice the data a lot of different ways, you could easily believe you find a subset, but it's not a general result that I've ever seen, except among people who look at the numbers. Yeah, I mean, in terms of maybe at the very, very top, then people like the dropouts better, you know, that is possible. For you, of course, as a venture capitalist, you might really like these Harvard dropouts, but as an employer of a functionary, on the other hand, not so interested in getting someone like that.

• 17:23Sonal?

  Do we know that people who are outliers, like dropouts, et cetera, those who've taken a non-conformist path, are they the ones who have the biggest outcomes? What's the data on that? I don't know if we know.

• 17:33Marc Andreessen?

  It's really, really, really hard to run these analyses, because the, the nature of what we do is, is sort of, all the outcomes that matter are flukes.

• 17:39Speaker 01

  And if you were just to average the 100 most successful tech people, Bill Gates and Zuckerberg could change the entire average of the whole distribution, personally.

• 17:47Sonal?

  Right.

• 17:48Marc Andreessen?

  That's right. And so, this is kind of the situation you get into, which is, if you actually try to run the analysis, you kind of end up maybe optimizing away from the flukes, and it may be that the flukes are exactly what you want.

• 17:56Sonal?

  Yep.

• 17:56Marc Andreessen?

  So, Brian, a lot of people disagree with your book, for whatever sets of reasons, but I think there's bipartisan consensus that a K through 12 results are not that great, and they certainly aren't getting better over time, for the most part. And then, you know, everybody agrees that the kind of hyperinflation of tuition, and some of the student loan crisis and so forth, something's wrong in, in higher ed. The people on the left of the political spectrum, their answer seems to be, pour more money in, invest more, anti-austerity, and thereby improve the system through greater levels of investment, pay teachers more, get better teachers, and so forth. People on the right side of the spectrum want to introduce competition, so they want voucher programs, and they want more universities to be accredited, and fundamentally, they want choice, which would lead presumably to innovation and to improved outcomes. Are you equally skeptical of both of those routes to improvement?

• 18:39Speaker 01

  Yeah, so, I'm super skeptical of the first one. It's given the enormous amount of waste in the current system, to go and put more money in the hands of people that would blow so much money on stuff where very basic data shows that almost nothing's been achieved. I've looked at at least some of the research, and it seems like the idea that if there's a lot more school choice, this will raise test scores a lot, doesn't seem to be true. I think the obvious explanation is that parents aren't that concerned about getting their kids' test scores up. So they may want them to have high SAT scores, but in terms of state standardized tests, most parents don't care about that stuff very much. And instead, they're mainly concerned about getting their kid into a good college. Probably a lot of what people are shopping around for is just a more pleasant experience for the kid, just a school that's a better emotional fit for the kid, which I say, you know, is worthwhile.

• 19:24Speaker 01

  So I mean, I say at least, you know, like, you know, kids could either be happy while not learning much or sad while not learning much. I'd rather they be happy while not learning much. But this is, you know, like a very deflated version of the usual view in favor of school choice, which I would say if parents really wanted their kids to get great reading and writing, I think choice would deliver that. It's just that it doesn't seem that this is actually.

• 19:46Sonal?

  Well, along those lines, I wanted to ask about sort of picking up on Mark's earlier point about pulling up the ladder behind you, which is that this disproportionately will affect poor and underprivileged kids.

• 19:56Speaker 01

  So when you think about this thought experiment, you'd rather be a high school dropout today or 1950, right? Ignoring just the greater wealth of the world, but just in terms of competitiveness, it's much better in 1950 to be a high school dropout because in those days, there were so many high school dropouts that there just wasn't that much of a stigma against it. And the reason is precisely that the higher the education level gets, the greater the stigma against people that have not kept up with the rising expectations. So when we're thinking about the effects on poor disadvantaged families, we shouldn't just be focused on the one kid in a thousand or more families, a fantastic student. We should also be thinking about the typical kid from a family who now in many ways has fewer opportunities than they would have had in the past because employers expect so much education.

• 20:39Speaker 01

  There is this treadmill or rat race effect where for any one individual that would give them more opportunities, and that's true, but if you go and give a whole generation more opportunities, the result is just that employers jack up the educational expectations to be able to even to give you an interview, much less to hire you or train you. And because most of the job skills people actually have are gained on the job, that's really the crucial step is finally getting, you know, convincing an employer that you deserve to be trained to do something real.

• 21:06Sonal?

  Actually, I want to quote back something you said because I thought it was super interesting. And you made this argument that bottom line, people as it is are just not paid what they're worth, that there are a lot of workers, especially less educated workers, who are paid less. And so if signaling is important. There's diamonds in the rough, people who are underpaid because they lack the right credentials to convince employers of their quality. And that immediately made me think of a lot of entrepreneurs and people who feel that workplaces don't value them, and so they find ways of creating value by just going off and doing their own thing. It's a way to sort that group out. So to me, that's an opportunity for people who are underprivileged, don't have the credentialist advantage. It did make me wonder if this kind of approach might also then not find the entrepreneurs in the system, because there's sort of this interesting argument that that's how entrepreneurs

• 21:51Sonal?

  are found and made. And by the way, this is disproportionately true of immigrant children, where education is their number one tool for mobility.

• 21:57Speaker 01

  So one of the earliest tests of signaling was to see whether self-employed has the same kind of payoff for education as the non-self-employed. Critics say, well, it may be that you use your credentials to impress clients, in which case it's not all that clear what the prediction is. So you can go to a doctor's office, he's got a wall plastered with diplomas on it to go and impress you. In terms of how much it matters for diplomas in Silicon Valley, I think crucial is just like getting the first meeting.

• 22:22Sonal?

  So you make the argument that part of the problem is that because we have so much focus on this model of education, that vocational school is an afterthought. And so my question for you is, first of all, what do you mean when you say vocational education? Like, what are some of the forms it can take? And then what are some of your theses around vocational education and making it work?

• 22:39Speaker 01

  Yeah, so there's many different kinds of vocational education. There's actually being in a classroom and just, you know, learning by doing while the teacher teaches you. There's apprenticeships. There's just on-the-job training. So there's many different kinds. Again, the most famous systems in the world are probably like the Swiss and the German systems of vocational education, where when you are in your early teens, if you're not going to the academic high school, instead you go to a vocational school where they teach you a trade. Now, in terms of what's so good about it, first of all, there's quite a bit of research on just the selfish gains to the individual student. So even for students that do go into college, there's still pretty good evidence that vocational education, in the long run, raise your earnings and employment rates, especially for kids who just don't like regular school.

• 23:23Speaker 01

  For kids who resent academics, for them vocational education seems really good because you know it leads them to finish and they then acquire some actual useful skills. Now in terms of the social value, that's where I really push it and say, if you take signaling seriously, the really good thing about vocational education is it seems like it's much less about impressing others. In fact, there may even be a little bit of a stigma against it and more about acquiring actual, concrete, practical skills.

• 23:49Marc Andreessen?

  So let me make the classic American counterargument against vocational education, which I think I might actually, in this case, agree with. I'm half German, so let me characterize utilitarian society, let's say. But that kind of school of thought cuts against sort of the fundamental American value of egalitarianism. And so, just picturing myself, I very easily could have ended up being tracked onto the vocational track. Mostly kids I grew up with, you know, certainly could have as well. And so, again, there's like an alternate universe, Marc Andreessen, that got tracked into a machine shop curriculum. Not that there's anything wrong with people who want to do that, but that would not have been the right path for me. And I would have gotten tracked at like, what, age 14? And I would have had all these other life paths kind of cut off by an authoritarian system that would have kind of decided for me, because that's kind of how that works.

• 24:29Speaker 01

  Yeah, so at least in Germany, it's not true that they decide for you. What they do is you take a test and they give you a recommendation and you people usually go along with the recommendation. So I guess I would just say the main thing is that if you're designing like any kind of system and spending taxpayer money, you've got to think about what is likely to work and the fact that, you know, like any system where you go and classify people or rank them, there's going to be some people who are misranked. But still, the question is, on average, which is the better system? Right? And the current one, where we try to keep virtually every American kid on the college track, is one that doesn't work out for about two-thirds of kids. I do realize there is this egalitarian idea, this egalitarian norm, although I mean, this is a norm that no society can actually really live up to, because right now, our norm is everyone's entitled to fail out of college once.

• 25:13Speaker 01

  But then it's like, so like, why do we draw the line there? And what is the actual evidence? That is the point where now we know the truth about whether or not you have a potential or not. I think like a more practical objection of vocational education is just it's a mistake to lock a 14-year-old into a job. And I say, you know, that makes perfect sense. Like, how about when you're 12 or 13? They say, you know, we're going to spend a year exposing you to 20 different jobs for two weeks each, and then see if anything sticks, sees if there's anything that you like it or good at. So again, there's no reason for vocational education to be narrow or limiting in this way. A common defense of the regular academic education is we expose people to all these different options. Well, you expose them to a bunch of different pipe dreams, but you expose them to a bunch of options, most of which don't work out for

• 25:56Speaker 01

  the people, even those who have enormous enthusiasm for them. So again, yes, it's a great idea to expose kids to a bunch of different options, but it's a good idea to have them be realistic options.

• 26:05Marc Andreessen?

  So Brian, there's a fairly common story people will tell, which is, yeah, I didn't necessarily get that much out of school, but boy, I had this one teacher, and boy, that teacher, and they always remember the name, Mr. And Mrs., you know, can we extrapolate anything from that? Could you imagine a system in which there are a lot more teachers that generate that kind of response from the kids?

• 26:19Speaker 01

  Yeah, so, you know, there's a lot of work done on teacher quality. The main thing known is none of like the obvious social science measures of teacher quality predict teacher quality. So having credentials from better schools, or having more credentials, or even more experience, none of the standard things that are in the salary formula actually seem to be very useful predicting whether or not you're a good teacher. But there is such a thing as a good teacher, which we know from the fact that if you do a random assignment of kids to some teachers rather than others, the same teachers who have good value added one year seem to have it the next year, too. But could you go and get more people like that? Seems that if you really had a system committed to, you could get good teachers, and also quite possibly for no more money than we currently spend on teachers that aren't so good, because the things that seem to matter are not things

• 27:03Speaker 01

  that are in the pay formula right now. So basically, it seems like we go and replace a lot of teachers from, you know, better schools, and with master's degrees, and with more experience, with just enthusiastic, fun, young teachers. Yeah, maybe people haven't even finished high school. But this is something where almost as soon as you say, you think, well, the current system isn't going to go along with that. So again, this goes back to how the system does not seem worthy of trust because they don't seem interested in using taxpayer money wisely and efficiently.

• 27:34Marc Andreessen?

  A related question, I listened to your most recent Econ Talk podcast, and you articulate there, as well as you did here, kind of your deep skepticism of the entire concept of learning with respect to, let's just say, at least facts. I mean, that most people don't even seem to retain what they have, quote, "learned," unquote. But you use the term "internalizing," which is, in some sense, I don't know if it's like learning that sticks. Maybe you could define what you mean by "learning" versus "internalizing."

• 27:55Speaker 01

  I mean, you're assuming, learning normally would just say, like, if you can do it on a test, you've learned it. But then, "internalizing" is where you go and apply what you've learned on the test outside of the classroom. It is already mentioning scientists who don't apply science in anything except their scientific topic. Like physics students, who when you give them a video game that operates exactly on the basis of Newtonian physics, they don't do well, even if people have, in fact, learned the information. They're not likely to be using it in the real world. And this sort of goes back to the question, how do people get good in their jobs, given all of the things that I told you? The answer seems to be practicing that specific job. You know, the way that you get to be a good pilot is not by taking a critical reasoning class. It's by, you know, getting in the simulator, getting in the cockpit, and learning to deal with 1,000 different horrible scenarios.

• 28:39Speaker 01

  And then, you actually are a good pilot, although it doesn't mean that you're going to be good at driving a bulldozer.

• 28:43Marc Andreessen?

  So, "internalizing," in some sense, has to be task-specific, or at least very domain-specific?

• 28:47Speaker 01

  I mean, in practice, it almost always is. So, the main thing about, like, all this research, research never tells you what has to be. It tells you what is. So, when I say there's hardly any transfer of learning, and yet, if you look at my book, I couldn't have written this without transfer of learning. Like, you know, so, like, in the book, I'm taking all of these different areas of research and saying, you know what? This is relevant to signaling. You have to be looking at the world all the time. So, I'm like, wait, does this relate to signaling at all? I'd been working on it for a couple of years, and then Alex Tabarrok across the hall said, you know, Brian, signaling explains why students cheat. And I'm like, "Yeah, of course." "But it never occurred to me until you told me." "Yeah, the whole point, you cheat in order to impersonate a good student, "so you want to be treated like an A student even though you're not." "And the only reason it works is because most of the people getting those As didn't cheat."

• 29:29Speaker 01

  "But yeah, it's hard and it just requires this discipline and this attention and this curiosity." "And sort of like a breaking down of barriers between work and life." "So like Tyler Cowen, who I'm almost sure you know, "is the previous example of this." "There are people like this and they do add a lot of value." "Although, I mean, another thing you say is that they don't learn this in school." "Wherever they're actually in classrooms, "their minds are wandering so far from what they really need to know." "And again, it's not that the school is teaching this, "but you can go and do well in your classes and also think about the world simultaneously."

• 30:04Sonal?

  Coming from developmental cognitive psychology, "there was a big thing that was in vogue back then was this idea of case-based education, "which is sort of like basically more grounded ways of teaching "in a more applied, like a word problem instead of a math problem." "And that's basically an easy way to think about it." "And so one question I had is how much of this is tied up to the mode and delivery of education "than the education model being broken right now?" "Because then certain technologies like VR, "immersive experience games, "could give you the ability to learn these things in a much more grounded way "and that would be the education model." "So how much of this is basically about mode and delivery "than it is about the problem is that it's vocational school versus the current models for teaching?"

• 30:46Speaker 01

  There is a good amount of evidence on teaching techniques "that are more effective at both learning and retention." "The striking thing is that the existing system has almost no interest in this research." "If I were a venture capitalist, I would just throw them out of the room "because you're not even trying to do those, but you want more money?" "Look, how about you first get your own house in order, "go and apply knowledge that is already sitting right in front of you "that requires not an additional dime of spending "and just win me over by showing that you value my money?" And to me, this is a pretty obvious thing for taxpayers to say, but it's the problem just that we're not using good methods. I'd say it's deeper than that. The problem is there doesn't seem to be much interest in applying good methods, whether they're known

• 31:26Speaker 01

  or not. And I think this is because the signaling function of education is on an emotional level, at least, very deeply grasped. And so there's almost no pressure from anyone in the system to actually improve genuine learning, whether for parents, students, teachers, anybody.

• 31:41Marc Andreessen?

  So if Brian Kaplan were made the czar of education in the United States with power to implement a new system that would take effect in five or 10 years, what would be the new education system for the US that would be the best of balancing all the trade offs we've been talking about?

• 31:53Speaker 01

  I think the best would be separation of school and state and just get government out of education entirely. I realize this will seem crazy to most people, and I'm happy to admit this goes beyond the book. But a system where cut overall spending by a third and then reallocate, say, half of the money that we currently spend in junior and high school to vocational education. That seems like a very, very reasonable revision. Honestly, just to have average education level fall by about four years. So to turn to a world where high school serves the function that college does now and college serves the function that a graduate degree does now. And then when people are still in school, they're being prepared for a job if they don't want to go on to something that really requires

• 32:37Speaker 01

  a lot of technical training. So that's basically the rough sketch. I realize it is rough.

• 32:44Marc Andreessen?

  Got it. And then sort of same question, but for different places in the world. So there are, as you know, a lot of countries that are classified as developing economies where in the last 20 years, a lot of them have had very rapidly rising education rates. These countries now are building out high school systems. What would Brian Kaplan, education's czar of one of those countries do?

• 33:01Speaker 01

  So there is a pretty big literature on the effect of education on development. Again, almost everyone who works assumes education must be great for development. And then they go to the numbers and come away saying, huh, we can't find what we know is true. Right, and let's get a different data set. Let's try different statistical methods. Right, and in the end of it, still, the main result is that the measured effect of national education on national income is much smaller than the effect of personal education on personal income, which fits very nicely with the signaling story and does not fit really with any other story out there. There's a few fringe ones. A lot of people want to look at this and say, well, the reason why education isn't helping development is there's so many terrible schools and schools where the teachers don't even show up. Right, so Connaughton Pritchett has, you know, has done some great work on just the low

• 33:46Speaker 01

  quality of schools in the third world. But here's the key puzzle. Even in countries like India, where you have a terrible problem with teacher absenteeism, in the labor market, education pays the individual who has more degrees. So it seems to have, you know, quite good payoffs. This makes almost no sense from the point of view of skills, but why would schools that are so poor actually wind up paying off in the labor market? And again, I think signaling can explain why there's this disparity between the selfish payoff for the student and the crummy payoff for the country.

• 34:15Sonal?

  There is, though, an effect that's been shown in developing economies where the existence of high-quality universities is linked to the economic outcome. Is that not true?

• 34:23Speaker 01

  I'd say that not only is that very debated, I'd say it's even more debated than the other stuff. Sort of like the textbook view is that primary education's great, secondary education's good, and tertiary education, eh. Then if you go and look more closely, even that doesn't seem to be very well-supported. But this is very much people's intuition, is that the longer you go, the less directly useful it is. So again, the sort of the way you phrased it, if you say high quality at high-quality universities, so if you basically go and find the cream of the crop and say IIT is important for India.

• 34:52Sonal?

  Right, the IITs.

• 34:53Speaker 01

  That is a lot more plausible, but what normally happens? And normally, when third-world countries spend a lot of money on tertiary education, often you like you wind up getting something more like the Egyptian result where, I think the government employs like 80% of college graduates. And of course, and you need that in order to be considered for most of these government jobs. So, I mean, that's probably getting one of the worst cases where essentially the government said, "If you get a college degree, we will hire you, we've got guaranteed employment." And the result is they greatly expand the education system and then just siphon off almost all the talent into these government jobs that seem to add very little, or you might even go so far as to say that they are economically destructive and they're just interfering with the development that would have happened without them.

• 35:32Marc Andreessen?

  So, do you know the author of Werner Vinge?

• 35:34Speaker 01

  The Singularity, right?

• 35:35Marc Andreessen?

  The Singularity. So, one of his later novels is one of my favorite novels of all time called Rainbow's End that fewer people have read, but I really like it. It's got kind of two main themes. One is the end of the world and the other is education, and not in that order. And of course, I bring it up because Dr. Vinge is a very respected professor of computer science in the University of California system, so he's, you know, like you, he's somebody who's lived his life in the educational system. So, the conceit of Rainbow's End is a famous poet is in his 70s in the present day and gets Alzheimer's disease, but survives and comes basically out of the tunnel in 20 years when they find a cure for Alzheimer's and sort of reemerges as a healthy 90-year-old, but having basically lost life skills. I mean, in a very different world where there, you know, it's a completely different set of life skills required to succeed. And so, he gets sent back to high school, right?

• 36:16Speaker 01

  So, it's like an Adam Sandler movie?

• 36:17Marc Andreessen?

  Yeah, basically, right. It's like a very highbrow, highbrow Adam Sandler movie. So, it's, you know, 2038 or whatever, and he's sent back to high school and to his great dismay and alarm, the entire high school curriculum has been replaced by "learn how to use search engines" for four years. Basically, the theory being that by that point, the internet has become such an amazing intellectual prosthesis or augment, right, that it just, there's no point, kind of to your point, unquote, "learning something" because all possible knowledge is accessible through the correct search. And so, the only life skill that really matters from an intellectual standpoint is learning how to compose the correct search. Is there something in there that, as sort of the internet does become the fountain of knowledge, is there something in there that it might show the way to some new kind of way to think about education?

• 36:56Speaker 01

  I tend to think no, just because there's still so much judgment involved and there's so much background knowledge that you need in order to figure out what a good search would be.

• 37:04Marc Andreessen?

  But where would you get that background knowledge from other searches?

• 37:06Speaker 01

  So, in terms of the trade-off, how much do you need to have memorized versus how much can you have in the cloud? That's something where I can easily see technology has moved the bar a lot towards you just don't need to memorize as much. But in terms of the conceptual structures that you need in order to make sense of anything... That's something where so far I haven't seen any sign that technology is making those conceptual structures any less important. In a way, there's a bigger payoff to everything that's conceptual structures because it gives you an idea of what could even possibly be out there. Again, like for me, so much of writing this book involves searching Google Scholar. And yet, a lot of what I would do is start off with jargon that I already know as an economist, and then say, hmm, are there other fields that talk about this but use different words to describe it? And how would I find out what that is?

• 37:50Speaker 01

  In the end, a lot of times, I would just go and find a person, like psychologist Steve Cece. And I'd say, I'm having trouble finding stuff on, say, on forgetting in psychology. Say, oh, well, that's because we don't call it forgetting, we call it decay. But now I can go and type that in, and oh, there's tons of great stuff on forgetting, it's just they don't call it that. So can you imagine a search engine so divine that even this wouldn't be necessary? I guess. I don't see this as the natural end of where we're going, at least, but what do you think?

• 38:18Sonal?

  It sounds like it's actually validating the rainbow.

• 38:20Marc Andreessen?

  I need to find out what people call something, I'll do a search and find out.

• 38:23Speaker 01

  There has to be someone else who's thought about this very same question. So often, when you're doing research, you're asking a question no one else is really thinking to ask.

• 38:29Marc Andreessen?

  But how would you find that person? Through a search?

• 38:31Speaker 01

  I like the prosthesis, buddy. And like, the prosthesis has got to attach to the body at some point, or else you've got a real problem.

• 38:36Marc Andreessen?

  So I just wanted to touch real quick on a couple of your other books, which I think are equally fascinating. So one of your other books is called The Myth of the Rational Voter, and I recommend this book to people all the time. The response I always get is, duh. You know, have you met people? Have you seen who we elect? Obviously, voters are rational, what possibly makes this an interesting topic?

• 38:53Speaker 01

  I mean, is it duh for everyone but me, or is it duh for everybody?

• 38:56Marc Andreessen?

  Well, no, it's just, it seems like it's obvious, and therefore, it must be a shallow idea, whereas you actually turn it into a very deep idea, the idea that voters are not rational. So what's deep about the idea that voters aren't rational?

• 39:04Speaker 01

  Right. I would say that there is a very popular view that democracy is competition, and the people know what's good for them, and who are you to say that you know better than anybody else what's actually good for them or for the country? And in the book, do two things. So one is to say that, in terms of, you know, basic economic theory and common sense, there's a good reason to think that people's political views would be very poorly worked out compared to their practical views, which is that if your practical views are stupid, you suffer. If your political views are stupid, on the other hand, what happens to you? The same thing that would have happened to you if your views were smart. Because you're just one individual. So one individual can, you know, you can change the content of your shopping cart, but you can't actually, as one person, change government policy. So there's a fundamental difference between shopping and voting.

• 39:49Speaker 01

  With shopping, if you make a mistake, you pay the price. Whereas voting, if you make a mistake, you were just tipping the scales a little bit towards all of society having a worse outcome. So what I say is, you know, rather than thinking about democracy as a supermarket, think of it more about as like a common pool of water that is also our trash disposal system. And so people simultaneously are throwing in all of our electrical garbage, but it's also where we drink. And so you put your glass in and say, wow, I wish there was some other place that I could drink that didn't have the sludge in it. But then the other part of the book is actually trying to measure this. And in particular, there is a common view among social science that it doesn't really matter that most people don't know what's going on because errors balance out. So the people who think that immigration is better than it really is will balance out. People think that immigration is worse than it really is, and the median will still be

• 40:34Speaker 01

  fine, and this is what government is all about, is finding what the median person thinks, not what the extremes think. And so in the book, I go over a lot of evidence that the, that the popular misconceptions don't work this way. And usually, there are patterns to the errors, and there are some errors that are just much more emotionally appealing than other errors. So I talk a lot about anti-foreign bias, tendency to underestimate the social benefits of interacting with foreigners. Around the world, protectionism is very popular. And just in general, emotionally, it's just so pleasant for most people to go and blame foreigners for their problems rather than anything else that they're doing. So it's so much easier to say, you know what the problem with the government deficit is? It's foreign aid. Not Social Security, Medicare, the actual numerically enormous things, because those

• 41:19Speaker 01

  are spent on the elderly, who we like. Better to go and do, use like a moral causation, where you just blame everything on a very small number of people. So like in the book, I talk a lot about these misconceptions about economics, but... Also talk about misconceptions in other areas.

• 41:35Marc Andreessen?

  Right. Well, the other framework that I found very helpful in the book is the idea of rational irrationality. Maybe you could, maybe we could close on that. How could irrationality possibly be rational?

• 41:43Speaker 01

  If it responds to incentives is really all that I'm talking about. My favorite example, this is someone shoots their mouth off, makes an extravagant prediction. And then you say, okay, let's bet on that. And then suddenly, almost everyone tones down or just doesn't want to bet. Now, one possibility is people are just lying when they say this stuff. But normally, like whenever you're arguing with politics, like what incentive do they really have to lie? Seems more like they are sincere, but undisciplined. And what does the money do? The money gets them to raise the discipline level up to a much higher level. So the idea of putting your money where your mouth is, this is one where it's been around for a long time. But I think it captures this deeper insight that when there's no money on the table, people form their beliefs in a very emotional way.

• 42:28Speaker 01

  But this is not a fixed trait of human nature. It's not like they have to be emotional. Rather, they're selectively emotional and in ones where there's some real stakes, some skin in the game, this actually does change the way that people reason. It changes their probabilities. It gives them modesty. I'm a huge fan of Philip Tetlock and Dan Gardner's book, Super Forecasting. And again, just the habits of mind that they talk about, like who, what are some habits of people who have a good track record of prediction? Some of the main ones are, we are thinking in degrees. So avoiding hyperbole, not saying things are certain or impossible. When you've been wrong so many times in your life before, and yet it's so emotionally appealing. Saying it is certain, that feels so much better than saying it is 60% likely.

• 43:13Speaker 01

  Imagine someone pushing a war, saying, look, this war is a 50% chance of improving things, 30% chance of making no difference, 20% chance of making things worse. Let's go. Right? That's not a normal reaction. There's something broken in in human nature that we that we that we actually don't respond to that kind of person, even though that's the kind of person that really provides responsible and thoughtful leadership.

• 43:37Marc Andreessen?

  So I also carried away something, maybe even I would say human and sympathetic from that as well, which is, if I recall correctly, it's actually very difficult and in fact often painful to be rational. It takes a lot of work as human beings to learn all the facts and then to exhibit the kinds of habits of mind that you were talking about. And so we as human beings can't expect ourselves or our fellow human beings to be rational in all things at all times. And in fact, we should cut ourselves some slack. There are going to be many areas of life in which we are going to be rational and that has to be expected because we're human.

• 44:04Speaker 01

  I guess that I can totally see that reading. I read it very puritanically, like your sinners, stop doing this. Yeah. So I see it's true that it's hard to learn a lot of facts. It's not that hard to exercise common sense. And there are so many intellectual techniques that can be used that are not a lot of actual work. It's not beyond our capabilities as humans. And for me, openness to betting is one of the ways I try to pursue virtue is saying, look, I have a blanket announcement. If you think that I'm saying something ridiculous, let's form a bet on it. And like I'm sticking my neck out. I am always open to bets. And if I will say, okay, I misspoke, I exaggerated, I'm going to accept my failure and admit it. And that is the way you improve is by admitting when you're making mistakes.

• 44:46Marc Andreessen?

  That is a great place to end.

• 44:47Sonal?

  Austerity. Brian, thanks so much for joining the A6 and Z Podcast.

• 44:51Marc Andreessen?

  Brian, thank you so much for the time.

• 44:52Speaker 01

  Totally my pleasure.