Alex Tabarrok - Prizes, Prices, and Public Goods

Alex Tabarrok is a professor of economics at George Mason University and with Tyler Cowen a founder of the online education platform http://MRU.org. http://MRU.org I ask Alex Tabarrok about the Grand Innovation Prize, the Baumol effect, and Dominant Assurance Contracts. Watch on YouTube https://youtu.be/i75KvXVrtLw, or listen on Spotify https://open.spotify.com/episode/3XXAZWizp0oaqRB2QOQr4d?si=GQGabXD7SGy3xr_aUe2oQQ, Apple Podcasts https://podcasts.apple.com/us/podcast/9-alex-tabarrok-prizes-prices-and-public-goods/id1516093381?i=1000495352718, or any other podcast platform. Episode website here https://www.dwarkeshpatel.com/p/alex-tabarrok. Follow Alex on Twitter https://twitter.com/ATabarrok. Follow me on Twitter https://twitter.com/dwarkesh_sp for updates on future episodes. Alex Tabarrok's and Tyler Cowen's excellent blog: https://marginalrevolution.com/ Thanks for reading The Lunar Society! Subscribe to find out about future episodes! Timestamps: (00:00 https://www.youtube.com/watch?v=i75KvXVrtLw&t=0s) - Intro (00:34 https://www.youtube.com/watch?v=i75KvXVrtLw&t=34s) - Grand Innovation Prize (08:45 https://www.youtube.com/watch?v=i75KvXVrtLw&t=525s) - Prizes vs grants (14:10 https://www.youtube.com/watch?v=i75KvXVrtLw&t=850s) -Baumol effect (27:50 https://www.youtube.com/watch?v=i75KvXVrtLw&t=1670s) - On Bryan Caplan's case against education (31:35 https://www.youtube.com/watch?v=i75KvXVrtLw&t=1895s) - Scaling education online (48:50 https://www.youtube.com/watch?v=i75KvXVrtLw&t=2930s) - Declining research productivity (52:15 https://www.youtube.com/watch?v=i75KvXVrtLw&t=3135s) - Dominant Assurance Contracts (58:40 https://www.youtube.com/watch?v=i75KvXVrtLw&t=3520s) - Future of governance (1:04:05 https://www.youtube.com/watch?v=i75KvXVrtLw&t=3845s) - On Robin Hanson's Futarchy (1:06:02 https://www.youtube.com/watch?v=i75KvXVrtLw&t=3962s) - Beating Adam Smith (1:08:35 https://www.youtube.com/watch?v=i75KvXVrtLw&t=4115s) - Our Warfare-Welfare State (1:19:30 https://www.youtube.com/watch?v=i75KvXVrtLw&t=4770s) - The Great Stagnation vs The Innovation Renaissance (1:21:40 https://www.youtube.com/watch?v=i75KvXVrtLw&t=4900s) - Advice to 20 year olds Share https://api.substack.com/feed/podcast/%%share_url%% This is a public episode. If you would like to discuss this with other subscribers or get access to bonus episodes, visit www.dwarkesh.com https://www.dwarkesh.com?utm_medium=podcast&utm_campaign=CTA_1

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

• 0:00Speaker 00

  Look, I want a small government, but I want a government to be able to do what it's supposed to do at the time it's supposed to do it.

• 0:12Dwarkesh Patel?

  Okay, today I have the pleasure of speaking with Professor Alex Tabrock, who's the Bartley J. Madden Chair of Economics at the Mercado Center and a professor of economics at George Mason University. And of course, he's a co-author of the Popular Marginal Revolution blog with Professor Tyler Cowen, who I've also had the pleasure of talking on with the podcast. So Professor, thank you for coming on the podcast.

• 0:31Speaker 02

  It's great to be here.

• 0:33Dwarkesh Patel?

  Awesome. Okay, so first I want to ask you about the grand innovation prize. Can you explain what this is? And I'd like to ask you some more questions about it.

• 0:41Speaker 02

  Sure. I mean, the basic issue is, is that clearly, speed really matters at this point in time in the midst of the pandemic. We're already been too slow. We've been behind the virus every single step of the way. So we want to find a way of speeding up the incentives to produce a vaccine, or a diagnostic, or a therapeutic. And you might say, well, you know, don't the companies, for example, already have an incentive to be quick? And to some extent, they do, but not as much as we would like from social incentives. So think, for example, about a vaccine manufacturer. Typically, most vaccines fail, right? They're hard to

• 1:24Speaker 02

  produce, they're complex, and most of them fail. And what this means is that a vaccine manufacturer, they're not going to be willing to build a factory, to ramp up a factory, to get the doses flowing, right? Until the vaccine has been proven safe and effective, and it's going to be approved. Okay, so they're not going to start moving really until the vaccine is approved. So if you want them to move faster, you've got to give them bigger incentives. And there's a variety of ways of doing that. One is to have like a big prize, you know, a billion dollars to the first vaccine, which meets a set of criteria, you know, it's effective at, you know, 60%, 70%, has such and such safety criteria and so forth.

• 2:06Speaker 02

  Or you could pay directly for manufacturing costs in order to get the firm to build the factory, you can say, okay, we're gonna pay some of your costs. And there's pluses and minuses, but the basic idea of a prize or advanced market commitment, or advanced market purchase, or paying for at-risk capacity, is the firms don't have as strong an incentive to ramp up the vaccine quickly as we would like. So we want to give them some extra juice.

• 2:38Dwarkesh Patel?

  Gotcha. And the incentive is even weaker, because you can only sell somebody a vaccine once, if it works at least, right? Yeah. Okay, so I have a question about how do you get money to the people? I guess you could give grants for manufacturing, but if somebody has a great idea for creating a vaccine, there's a billion-dollar price to making that vaccine, but they don't have the initial funding to, you know, get that manufacturing or even testing up to par. What do you do about that?

• 3:03Speaker 02

  Right. So there's a tricky set of trade-offs, because on the one hand, the government really is not good at picking winners and losers. We know that from industrial policy, and that holds just as well for vaccine policy. So that pushes you towards just having a prize where you just, you know, all comers. And another advantage of that is, we really don't know what kind of vaccine is going to be the most successful. And there's a whole bunch of different types. You know, there's the traditional live attenuated vaccine, a killed-virus

• 3:40Speaker 02

  vaccine. There are some new vaccines using mRNA technology, some DNA vaccines. And maybe it's not going to be a vaccine at all. Maybe it's a therapeutic, right, which could serve the same kind of purpose. Or maybe there's some other innovation. And when you have a prize, you really are opening up the field to these, to the crazy ones, right? To the ones who wouldn't necessarily get through the NIH committees, right? And that's, of course, the classic longitude story of the longitude prize, which was won not by Newton, but by this watchmaker, clockmaker. However...

• 4:24Speaker 02

  We also then have the trade-off, which is the point that you mentioned, is that some firms may not have the capital. And for those firms, there's a greater argument for funding them upfront. And here, you know, there's no easy solution in a pandemic. I kind of think you want to go at all guns blazing, okay? You want to use almost all the tools that you have available to you. So what, I've been working on this problem with Michael Kramer, Nobel Prize winner, who is kind of famous for the advanced market commitment for the pneumococcus vaccine,

• 5:05Speaker 02

  which probably saved 700,000 lives, was given to millions of children. And in that, the primary tool was the promise of a fixed price, if you were to produce such a vaccine. We're more working on the COVID vaccine. We're more pushing towards paying upfront for a large fraction of the manufacturer's costs for producing a factory or for repurposing capacity. And it turns out it's cheaper in our context to do that in this way. We could talk more about that. But there are all these trade-offs, for sure. And right now, we're kind of pushing towards paying more for manufacturing capacity.

• 5:48Dwarkesh Patel?

  Yeah, that's very interesting. Why can't it be the case that financial instruments can be created so that, you know, VCs make their living off of these sort of like high-risk, high-rewards investments? Maybe they can like have a deal with a researcher that they'll fund a lot of, they'll fund their testing and manufacturing. And if the vaccine works, then they get half the return to the price.

• 6:10Speaker 02

  Sure, I mean, absolutely. When the markets are working well, I mean, that's what would happen. It just takes a long time to set these things up. And at every single stage, there are these information problems. So we say that, you know, the government is not good at picking winners and losers. But the truth actually is that the VCs aren't that great at it either, right? They have a comparative advantage. They have an advantage over the government. I mean, they have skin in the game. But For a VC to learn enough about vaccines and to figure out to, you know, you gotta be on, on, on presence to figure out which, which are the scams, you know, and which are a real possibility of working. This takes time.

• 6:53Speaker 02

  So I think, look, here, here's how I put it to a few people I've talked with. I've said, you know, I'm a sort of conservative or free market economist, right? So I rarely, if ever, say things like this. But what I've been telling people is now is the time to throw money at the problem. Okay. That's not something I usually say.

• 7:13Speaker 02

  But the costs of the COVID disaster, which are now running in the trillions, right, are so large that it's really gonna pay, as I said, to kind of advance on all fronts. And we're gonna lose money doing this. Okay, there's gonna be waste. We're gonna fund some vaccines which look great and then fail. The AstraZeneca vaccine, for example, this was early out of the gate. This is from Oxford. It looked great. They have sold, pre-ordered like 2 billion doses. So this is kind of the, really, the most promising vaccine. A lot of countries have bought into the AstraZeneca vaccine. And as you might know, last week, the trial, one of the clinical trials was shut down

• 7:58Speaker 02

  because they had a negative effect. They're not sure whether it was due to the vaccine or due to something else. But it's a danger and that, that could happen. I'm not saying the AstraZeneca vaccine is gonna fail. I hope it doesn't. I hope this turns out, you know, to be a false alarm. But that could happen at every stage in the process. So you really want to have a diverse set of vaccines in your portfolio. And some of them, you know, pay for manufacturing capacity. Some of them pay, have the price, the guaranteed price. We're gonna have some, you know, fast grants like Tyler has been given out. You know, take a lot of shots on goal because we don't know which shot on goal is actually going to be the one which succeeds.

• 8:44Dwarkesh Patel?

  Okay, so let me ask you about prizes generally then. So are you emphasizing prizes in this context because it's such an immediate and big problem that, you know, we should try out different funding mechanisms? Or do you think prizes generally are, are use, could be a useful solution in many fields where the problem space or the solution space is big?

• 9:01Speaker 02

  Yeah, exactly. So the latter. I mean, I think one advantage of the prize in the current context is it's easy to explain to Congress, okay? So like, just do something, okay? And this has been incredible frustration to me that Congress has been so lazy, so slow. They're so complacent. They're not doing anything. I mean, it took them ages to set up funding for testing, and then they set up the funding and they haven't spent the money. So it's just incredibly frustrating. And I thought a prize, because you can announce it even before you've gathered the funds or figured out who's going to be on the prize committee or anything like that. Just announce the prize. We're just going to give

• 9:46Speaker 02

  we have this pot of money. It'll be there, you know, a year or two from now. We'll figure out who gets it. That that could be done fairly quickly. And in fact, there is in the code, I think from Obama's time, every agency has the right to set up a $50 million prize, up to a $50 million prize. And they can be accumulated. So you could have, you know, two or three agencies or four agencies could right now set up $100 million, $200 million prize just by getting together. They haven't done that, but they could do that. So that was one thing, just the speed. The other point is the one which you just mentioned is that prizes are particularly good when sort of the experts have kind of failed.

• 10:28Speaker 02

  Usually the experts are right. Don't get me wrong. I'm not anti-expert. Usually the experts are the best people to go to. But then sometimes you have a problem where the experts have run against the wall and you need to take a very broad view and you need to get the crazies in and you need to get the out-of-the-box thinkers. And a prize could be very useful for that.

• 10:52Dwarkesh Patel?

  So why don't we have prizes in other fields? Like, why are grants the dominant way of funding research?

• 10:58Speaker 02

  It's a good question. And I don't really know the answer. Prizes were much more common in the 19th century. And then they kind of failed, failed off or trailed off in the 20th century. And they've seen a little bit of a resurgence in more recent decades. One reason perhaps is that the, you know, giving out money is kind of a powerful job.

• 11:20Speaker 02

  And if you're doing it, you know, the NIH or a committee, you know, the people you you you bring in, you give out the money to the people who are giving out the money. Um, and that's a very nice kind of job, very nice, can be a very nice job to have. Uh, and a prize. If it's set up correctly, where, you know, you have a fairly strict guidelines as to what meets it and what doesn't meet it. Um, and you only get it if you win. That kind of gives less power to the prize givers. And maybe that's part of it, but I'm not, I'm not really sure, actually. It's a puzzle. It's a puzzle as to why, uh, grants have been used much more than prizes.

• 11:58Dwarkesh Patel?

  Um, and but why don't we see evidence of the, a clear efficacy of prizes? Uh, because there doesn't seem to be a clear coordination problem here, like just an altruistic patron could just decide to like set up prizes for, not necessarily this problem, but just a lot of other problems. And if for many problems prizes are more effective, we should see that the results from these prizes, the innovations they create, are just much more effective than, uh, the grants that are afforded in that field. So why aren't we seeing like a clear evidence of the superiority of the prize?

• 12:24Speaker 02

  Well, we do have some evidence on this. Uh, for example, there are these, uh, Howard Hughes grants. Uh, so Howard Hughes actually tried to set up a tax dodge and, uh, ended up found founding the, you know, Howard Hughes Medical Centers, uh, which have lasted much longer than his, his tax dodge. So he had some benefits, uh, there. Um, and the distinguishing feature of these Howard Hughes, Howard Hughes grants is that they say, okay, here's a bunch of money. Come back to us in five years and tell us what you did with it. Okay. Um, you don't really have to come up with a plan or tell them what you're gonna do. They're sort of like a genius grant in some ways.

• 13:07Speaker 02

  And it turns out that those, uh, grants are much more effective. The researchers who get them, that research tends to be much higher cited, more patents and so forth. And now, of course, you have to control for the fact that the researchers who get these grants are different than the researchers who get other types of grants. But even when you do that, even when you set up some fine controls, it looks like giving money with less strings attached actually is more successful. We'll also see with my colleague, as you know, Tyler Cowen, has given out more than $20 million in COVID grants and given them out incredibly quickly.

• 13:48Speaker 02

  I think Tyler gave out $20 million faster than the NIH had given out a single grant. So we'll see how effective that have been. There've already been some good returns from that. So maybe there'll be a resurgence in prize giving, or fast grants, or prize giving, or however you want to call it.

• 14:09Dwarkesh Patel?

  Interesting. Okay, so now let me ask you about the Bommel effect.

• 14:13Speaker 02

  Yeah, Bommel, yeah. So we have these industries where every year, just the prices just seem to keep going up and up and up and up without fail. And education, both primary and secondary education, as well as university education, is one example where the prices sort of inexorably, it seems, go up. Healthcare is another one. But there's other ones as well. Those are the ones which are big and are most common. But another thing people have noticed, for example, is that fixing things, repairs, seem to go up a lot over time. So it's much

• 14:57Speaker 02

  more expensive to have a pair of shoes repaired today than it used to be. Auto repair has been going up in price over time relative to other goods and services. So much that, for example, I had to replace the blade on my lawnmower. I didn't have to replace it. The blade on my lawnmower got dull. Now, even 30 years ago, you'd probably go to a sharpener. But I didn't do that. I just went to Amazon and ordered a new blade, right? No one sharpens their blades anymore, you know, maybe for like a $200 knife, but not for a lawnmower, right? You just buy another one, which is what I did. So anyway, so the question is, why are these?

• 15:41Speaker 02

  Why is it that some sets of goods and services appears to be going up over time? And William Baumol is an economist. He has an explanation for that. Traditional explanations, of course, as you're probably familiar with, say, well, there's something wrong with the education industry. There's too much government, right? Or there's something wrong with healthcare. And in each case, there's like some special story. And the story that Baumol gives, which is why, one of the reasons I like it, it's like one theory to rule them all, okay? But that's kind of the background. So we have this set of industries. Prices appear to be going up year after year after year, and the question is why?

• 16:25Speaker 02

  What we did, Eric Allen and I, in this little book, is first of all, we looked at the kind of traditional story. Is it like unions, you know, is it regulation? And it just doesn't appear to fit. It doesn't appear to fit.

• 16:42Speaker 02

  So what is the Baumol, what is the Baumol story? I explained it different ways when people ask me. I think I'll explain it in a different way today. I think the way to think about it is to think, first of all, in terms of real goods and services. Just think about the barter economy or

• 17:00Speaker 02

  forget prices for a second and just think about kind of real goods and services, the real economy. And you have some sets of goods and services which are increasing in productivity every year, right? So computers, they're getting faster and they're getting more, you know, powerful every year. Manufacturing, right? We're able to manufacture a car using, you know, less steel than ever before and in a shorter amount of time, using less labor, you know, and so forth. So these goods are increasing in productivity. There's some other industries which, for whatever reason, are just slower increasing in productivity. I mean, you're bound to have

• 17:45Speaker 02

  some randomness or just some differences, right? Some, some industries for technological reasons are increasing in productivity faster than other industries. So if you think about these two different types of industries, one of which is the progressive sector, the other of which is the, the slower sector, the slower productivity sector. Doesn't have to be negative productivity, just it's growing more slowly in terms of productivity. So you get, uh, in the progressive sector, you get a lot more output per input over time. And in the stagnant sector, only very slowly does output grow per unit of input. Well, if you think about that, then it has to be the case that prices in the stagnant or less productive sector have to be going up.

• 18:29Speaker 02

  And the simple reason is because goods trade for one another. So if you can get a lot more goods in the progressive sector, well, that means if you want more of the stagnant sector, you have to give up a lot. Okay. Because you can get a lot, right? So the kind of the classic example is the symphony, right? Or the string quartet, right? And you think about 1826, the string quartet, it takes four people 40 minutes to do a Hyden string quartet. And you think about, you know, 2020, it takes four people, you know, 40 minutes to do a string quartet. So their productivity hasn't gone up at all, but what you can get, the opportunity cost of that has gone way up.

• 19:14Speaker 02

  So in, you know, 1820 to hire four people, you know, for 40 minutes, you are giving up, you know, I don't know, so like a bicycle, whatever you want to call it. I don't know, something, you know. But now if you give up, you know, four people for 40 minutes, that's worth a lot, you know? You can get, you know, the movies and a dinner out and, you know, two bicycles or whatever, right? So you're just giving up a lot more. Um, and that's really why the price has gone up is because. Of the progressive sector. So to put it the other way, it's not that there's a problem with the stagnant sector. I mean, it's not a direct problem in any case. It's not that there's something wrong, unions or, you know, regulation or something like that.

• 19:59Speaker 02

  What's really going on is something good. Something good about the progressive sector. That's why things are getting more expensive because you can, you have to give up a lot more. Because this sector is becoming much more productive. Let me just give you one other way of thinking about that, and that is. What this implies is that prices in the stagnant sector will start to go up less quickly when the progressive sector is slowing down. So this means that the prices in the stagnant sector go up much more quickly when you have a lot of growth in the productive sector.

• 20:43Speaker 02

  And now with the great stagnation, okay? As growth is slowing down, we're actually seeing like healthcare prices are going up less quickly. So when did healthcare prices really go up? The big increase in healthcare prices was in the 60s, 50s and 60s. And that's when healthcare prices really exploded. They grew much more quickly, like 10% a year. Okay, and that's when productivity was going up. So now productivity is slowing down and we're seeing prices are not going up so quickly in these stagnant sectors. So that's kind of tells you this is not a good thing, right? So higher prices could be a bad, could be a good thing in some ways.

• 21:24Speaker 02

  And, you know, less price inflation could be a bad thing.

• 21:29Dwarkesh Patel?

  Yeah, it's such an interesting explanation. It just changes all of the prevailing wisdom that you hear everywhere about why these, why these industries are getting so much more expensive. But I have a few questions about that. So, so as you laid out, one of the explanations of the bubble effect is that one of the main inputs into many of these sectors is labor. And if you can get more out of labor in one sector over time and not so much in the other sector, they're willing to pay more for labor in the progressive sector and that raises the price of labor everywhere. But how can labor be getting more expensive at a time where we keep hearing about wage stagnation?

• 22:03Speaker 02

  About what stagnation? Oh, wage stagnation. Oh, yeah, right. Yeah, yeah, yeah. Right. So, so first of all, we hear about wage stagnation. So this is something which Helen and I actually add to the Baumol effect in our book. Because Baumol was, exactly as you said, most focused on kind of labor in general. And we just give it a slight twist in our, our book. We say, look, it's skilled labor. It's skilled labor, which in particular has gone up in price because it's become, a skilled laborer is much more valuable in Silicon Valley, right, than they used to be.

• 22:45Speaker 02

  And so you think about healthcare, you think about education. These are sectors which involved a lot of skilled labor. Okay? So I have a PhD. I teach, you know, typically like 30 students. We'll talk about online education later. But let's say I teach 30 students a semester. That's a really expensive use of someone with a PhD, right? I'm teaching 30 students and, you know, I don't want to say like I could go to Silicon Valley and, you know, be rich or whatever, but someone like me could, right? That's, that is the trade off which we're talking about as a society. Not Alex Tabrock could, you know, suddenly move to Silicon Valley.

• 23:26Speaker 02

  But somebody who has the opportunity of getting an economics PhD, they can spend their lives teaching 30 students at a time, or they could go work for Uber, which a lot of economics PhDs do, or for Amazon, right? And so the opportunity cost of a PhD has gone way up because now they can be hired, you know, for Amazon or Uber. And that's why, but they're still doing exactly the same thing in the classroom. They're still just teaching 30 students. So that, that price has got to be going, has got to be going up.

• 23:58Dwarkesh Patel?

  Does the Balmule effect suggest that AI automation fears are overhyped? Because more and more of the economy is service-based. Because, I mean, that sector isn't growing more productive, so it's not growing smaller. And also, that the value of labor goes up over time as the sectors that are becoming more productive can produce more.

• 24:15Speaker 02

  Yeah, yes and no. So it's definitely true that the service sector has grown over time. And in general, you know, the stagnant sectors will grow over time because there's only so many cars you can have or want, really, probably, right? So even if, when cars get less expensive, people might have one or two, and the quality goes up somewhat, but there's kind of a limit to, you know, revenues in that sector. So the stagnant sectors do tend to grow. So education and healthcare have become a bigger part of our economy, which is kind of natural. Now, does this mean that AI is overblown?

• 24:59Speaker 02

  I think it means that we're not going to run out of work, that's true. But I do worry, and as we're already seeing, that certain types of labor, you know, can be overrun by automation, right? So, you know, less skilled labor, their wages have not been going up. And partly that's automation, partly that's trade. And that is only going to become more serious with remote work and as well with automation. So, yeah, I do worry about if you can't have,

• 25:43Speaker 02

  if you can't raise your education level, you can't raise the skill level of your workers, then you're in real trouble. You know, people are very flexible, that is true. But, you know, I think about horses, right?

• 26:00Speaker 02

  You know, the horse, when we introduced the automobile, it's not like horses all found alternative employment. Uh, they found the glue factory, right?

• 26:12Speaker 02

  And you know, horses are pretty flexible. I mean, you can do a lot of things with horses. But it's not like they all slotted into different areas of the economy like the, you know, I mean, we did have more, uh, you know, horses for recreation and so forth. Uh, but no, overall, the number of horses is down. And so I worry about less skilled, uh, labor.

• 26:36Dwarkesh Patel?

  Hmm. Uh, and how much of the great, you mentioned the great stagnation. How much of it can be explained by just the economy over time is more and more dominated by the sectors that are growing in productivity the least?

• 26:46Speaker 02

  Yeah, I think that's exactly right. I mean, as you shift to the service sector, you see a, uh, decline in your growth rate simply because you're buying more of the things which are growing more slowly. Um, and that's going to continue. Uh, it, it, it, it can be part of a naturally growing, um, e- economy. Uh, there's a new good book on this. Uh, you, you probably know it. I've forgotten. Uh, I should take up the title.

• 27:18Speaker 02

  The author is going to be mad at me for not, I'll come back to it.

• 27:24Speaker 02

  But there's an optimal stagnation. Okay? So it can be the good that, uh, we are growing more slowly in the sense that you have focused your spending on sectors which are growing more slowly. And that's kind of a natural change. Now, obviously what we would like is for all sectors to boom. Um, and maybe we could do something about that, um, but it's not going to be easy.

• 27:50Dwarkesh Patel?

  Uh, so speaking of one of these, uh, slow growth sectors, let's talk about education. So you write in, you write in the paper, um, to decrease the college wage premium and the relative price of goods and services that use college educated workers, we need to increase the educational attainment of the US workforce. Now, my first guest on the podcast was, uh, your colleague, Brian Kaplan. And he would argue probably the opposite, that he'd say, you're just increasing the cost of signaling if you make people have a higher educational attainment because they had a signal relative to the other counterparts who are now more highly educated. Um, uh, so how would you respond to that?

• 28:24Speaker 02

  Yeah, so um, there's clearly, Brian is clearly right, uh, that there's a lot of signaling involved, this particularly in college education. Um, that's not the only type of education that counts, for example. Um, in Germany and Austria and many of the European countries, they invest a lot more in apprenticeships and worker training, okay? Um, so the interesting thing is, is that we have this crazy system, um, in the United States where uh, the high school graduation rate is actually quite low compared to other countries. Uh, I think at the time I wrote my book, it was only like 75% for men.

• 29:07Speaker 02

  It's gone up since then, it's maybe like 80, 85%. But we still have a large chunk of our workforce which does not graduate high school. But then, of those who graduate high school, we send a huge proportion to college, way more than in Europe, in Germany, okay? So we're sending like 60, 70% of those who graduate high school then go on to college. And in Germany, it's more like, you know, 40%, I can't remember the exact figures, but it's something like that, okay? So it's like a completely crazy system and, and on, and on the other hand, in Germany, they're almost everybody graduates high school, like 97%, something like that, right? So I think we can rebalance here, uh, definitely more people should graduate high school. Um, more people should be highly skilled.

• 29:52Speaker 02

  Does that mean they should be going to college? No, because when a lot of these kids go to college and then they take, um, you know, I'm gonna upset some of my, uh, friends, then they go into journalism, right? Which is like a terrible field, or psychology, right? These are terrible fields to go into for, for jobs. Um, and yet, these fields, journalism and psychology in particular, have grown. The number of people, shockingly, going into computer science has been like flat for like 30 years. I mean, you would think, whatever field has grown more than computer science, you would think it would be booming. But actually, no, it's pretty flat. Um, and what increases we have seen...

• 30:36Speaker 02

  In people going into computer science has all been foreign students. Okay, so the foreign students, they don't, you know, if you're from South Korea or India, you don't come to the United States to do a journalism degree or psychology, right? You know, your parents aren't going to allow you to do that. You come to get some hard science, right? And so I think the US students should do more of that.

• 31:00Dwarkesh Patel?

  Yeah, I promise I'm not contributing to the problem. I'm right here at university studying computer science.

• 31:05Speaker 02

  Okay, so I didn't want to make any presumptions.

• 31:11Speaker 02

  But yeah, I'm glad, I'm glad you fit the correct, the correct, you know, modus operandi. You're doing, you're doing the right thing.

• 31:20Dwarkesh Patel?

  Yeah, yeah, it would have been quite a, it would be pretty funny if I was right here studying journalism and had to hear you say that.

• 31:28Dwarkesh Patel?

  Okay.

• 31:28Speaker 00

  You actually could be a good journalist.

• 31:32Speaker 00

  There might be room for you.

• 31:36Dwarkesh Patel?

  So let's talk about something that you have launched, Marginal Revolution Online University, which is trying to change up the system a little bit. And so you've written about online education and how it's going to, how these massive open online courses are going to select for the best teachers. What's going to happen to the rest of the people who are trying to teach right now or already teaching?

• 31:58Speaker 00

  Yeah, so I think the long run is pretty bad for teachers. You know, we always complained that, you know, look how poorly we teach our teachers, you know, we pay them less than our sports stars, right? You know, you know, basketball or something like that. And what's going to happen is we're going to pay our teachers more like basketball stars and baseball stars, but there's just going to be a lot fewer of them, right? Because when you teach online, you know, I said earlier, I teach 30 students, but with Marginal Revolution University, I can teach 300 students and I do. So hundreds of, hundreds of thousands of students.

• 32:34Speaker 00

  So like, why, you know, then there's, I mean, frankly, people like me, if not me, are going to drive a lot of people in the teaching business out of business. And that actually would be a good thing. Because, you know, why should you? The problem is, education has always been difficult to scale, right? It's such a one-on-one kind of business. And if you're able to scale education, actually, we're being forced to because of the pandemic. But if you're able to scale education, there are huge, huge, massive returns that are possible there.

• 33:15Speaker 00

  So whoever cracks that nut, and I don't think it's been cracked entirely.

• 33:23Speaker 00

  You know, I think we've gone a long way, Marshall Revolution University, but we're only the beginning. Whoever cracks that nut is going to do incredibly well because you go from one person teaching thirty students at a time to literally you could teach the entire world. And you tie, this goes back to the Bommel effect. If you can tie education to a progressive sector, right? Instead of tying it to a labor-intensive sector, you tie it to technology, which is a progressive sector. Then, you know, all bets are off, everything can change really, really quickly. So, artificial tutors. Artificial tutors are already as good in randomized controlled trials as real tutors, okay?

• 34:04Speaker 00

  And you, you, you homework assessment systems. I mean, so what a, an AI can do is look at hundreds of thousands of student responses on a test, for example. And then it can figure out, looking at, you've written some tests, and it'll see your pattern of errors. And it says, aha, okay, machine learning. I could see, the machine will be able to see what concept you're not grasping. And so instead of just repeating, you know, the lesson, okay? The artificial tutor can pinpoint and direct the student

• 34:45Speaker 00

  exactly to that piece of knowledge, oh, you're using the quadratic formula incorrectly. This is what you need to know, okay? It can direct them exactly to that piece of knowledge, which they need to unlock the key, the key to learning, so the students can advance much more quickly. So I think there's huge, huge possibilities for artificial intelligence and for

• 35:09Speaker 00

  You know, video, online learning and so forth.

• 35:12Dwarkesh Patel?

  So you give out these optimistic scenarios of how education is going to change. But a pessimist might say, well, listen, we've we've gotten these sort of predictions back since Edson said that the motion picture is gonna change education forever. And the same thing happened with the radio, with books, with television. So what's different this time? How is this is going to disrupt the system that's been around for a thousand years?

• 35:34Speaker 00

  Yeah, you're absolutely right. You would think that with books, right? You know, oh, you can get all this knowledge, you know, in this book. And then you can carry it around with you. You know, you can read it at night whenever you want to. You know, this is much better than having to talk with some professor guy. Yeah, the book has a lot of the advantages, which I claim for online education. So this makes me, I don't know, worried or humble or modest or or I could be wrong. It's for sure. You know, I do, I tend to think, though, that so, you know, it's it's it's bad to say this time is different.

• 36:15Speaker 00

  But it is interesting how

• 36:19Speaker 00

  a group of technologies come together, which seemingly small differences can make big effects. So I think, for example, about the Apple Newton. Okay, so the Newton was basically a cellphone. It was basically a portable computer, kind of a, but it was just it for it just it wasn't quite fast enough. It wasn't connected to the Internet enough. So it just the Newton never took off. And yet, really, just it's really just a small step from the Newton to the iPhone. And yet, the iPhone was huge. Okay, but technologically, there's very little difference between the Newton and the iPhone. The Newton was a failure. The iPhone was a huge success.

• 37:03Speaker 00

  So I think just in kind of the speed at which things can be done online.

• 37:11Speaker 00

  And just the the quality of the animations, the music, the artificial intelligence, all of these things advancing. You kind of reach a threshold and then you get and then things boom. And I can tell you that my students, you know, I teach a bunch of students online at George Mason University. A lot of them, surprising even to me, tell me that they prefer online. They prefer at least my class online. And you know, this makes sense because students have different methods of learning. And the kind of the classroom kind of sucks, right? I mean, you sometimes you can't hear the professor. You know, you have to be there at a certain time at a certain place.

• 37:56Speaker 00

  You can't pause, you feel you can't ask questions. With online, something you know, you didn't quite hear, you didn't understand, you just pause it, you know, and you rewind. Right? Or think about podcasts.

• 38:09Speaker 00

  You know, I think a lot of us are very familiar with listening to podcasts at, you know, 1.5 or 1.25 speed, right? Because then you just slow down when there's something new and you speed up when there's something not new. And that's like a huge advantage. Okay, so I can learn a lot more with the podcast. Just because I'm in control of the speed, you know. And so this means that online students can go at their own pace. So there's a huge number of advantages. And this idea of tying it to a progressive sector, I think is important as well. This is an example I give, which which we discovered purely by accident. Is that we captioned all of our videos in English, okay?

• 38:53Speaker 00

  And then we discovered, you know, that YouTube automatically takes the English captions and translates them into dozens, hundreds of other languages. And what this means is that every improvement in DeepMind, right? Every improvement in artificial intelligence automatically turns into an improvement in our product. Because periodically, you know, the captions, they get better without us having to do any work, right? Just the YouTube algorithms, you know, improves the captions over time. And so our product is becoming better even without us making any changes.

• 39:30Dwarkesh Patel?

  But I don't want to speak for Brian Kaplan here, but maybe somebody might say, what if you're just solving the wrong problem? That people aren't really giving up four years of their life to get a better education.

• 39:42Dwarkesh Patel?

  I might just be thinking for myself here, but your videos are much better as far as an economics course goes than my high school economics course. And, you know, I still had to attend my high school economics course and my college economics course. And so unless we can figure out a way for people to signal their intelligence and conscientiousness through online courses, people are not going to substitute this for what college provides.

• 40:05Speaker 00

  Right, so this is why I think that the online, the MOOCs, as it were, you know, Coursera and things like that, they're not going to replace universities, but universities are going to go online, right? So you still need that stamp. You still need, we haven't cracked that nut. Maybe we will, either, you know,

• 40:28Speaker 00

  there are some ways of doing it in some fields right now, like computer science. But you're right, we haven't cracked that nut. But that's why universities will be the ones who go online. They're not going to be, and not all of them are going to make the transition. You know, we're seeing this right now with the pandemic. A lot of universities are falling by the wayside. But the ones which can have a reputation,

• 40:50Speaker 00

  a high reputation, they're going to be able to go online and they're going to be able to expand their market. Now, some of them you don't want to, like Harvard does not want to expand its market. But a place like George Mason, where I teach, okay, has a decent reputation in the world, okay? And we'd be very happy to expand our market. We'd be happy to have tens of thousands of students all over the world. Georgia Tech has done this probably, as you know, better than anybody. Georgia Tech has the largest computer science program in master's program in the world, has something like 7,000 online students. The online students are treated exactly as the residential students. They're graded in exactly the same way.

• 41:33Speaker 00

  The professors don't even know which is online, which is not online. It is a hugely successful, graduates, I think, like 7% of all the computer science masters in the world. So it's a hugely successful program. And they're only able, it's a quarter of the price, less than a quarter of the price online than it is residential. So that, I think, is a... A trailblazer or a warning sign depending on your point of view. But Georgia Tech has taken a on a master's degree in computer science and successfully brought it online to the entire world. And that is a mark for the future.

• 42:14Dwarkesh Patel?

  Hmm. Uh, now tell me if this is too optimistic a scenario, but is it? Or if it's precluded by Bama's effect as well, but is it possible that the people who would have been teachers otherwise, but now are being outcompeted by the best teachers in the world? They will just end up being personal tutors and then we can get back to like a sort of older model of one-on-one instruction. Or is this gonna be too expensive?

• 42:34Speaker 00

  Yeah, some of that will will happen. So I think, uh, you know, what what is the real not to crack? Um, a lot of it, right? Is, uh, I'm gonna sound contradictory, because I'm gonna say psychology, right? Not that you need a degree in psychology, but, um, people are gonna need coaches, right? So, you know, life coaches, which people laugh at because only the rich and famous, you know, have a, you know, a life coach or whatever, right? But that is a large part of education, I think, is gonna be more like sports people, people in sports, they all have coaches. So why shouldn't you and I have coaches?

• 43:14Speaker 00

  So I think there will be a industry for life coaches, somebody to encourage you, somebody to try and map out your skills, uh, kind of like what the high school advisors are supposed to do, you know? You're this sort of person, you should go do this, you know, there'd be a lot more of that. So education will be debundled, right? The kind of traditional lecture part of it, sage on the stage or all that, that transmission of information. That part will go online and a many, a much smaller number of teachers will teach many more people.

• 43:53Speaker 00

  But the coaching part of it, uh, that the possibly the tutoring, uh, will be spread out, um, because it'll be much more hierarchical, um, actually. So, you know, one guy at the top will have this kind of hierarchy of TAs and things like that.

• 44:12Dwarkesh Patel?

  I mean, that's kind of how college works anyways, a little bit.

• 44:15Speaker 00

  A little bit already, yeah, but even more so, right? Like, we, you're absolutely, we have it already where, you know, Michael Sandel, you know, teaches, you know, 400 students or whatever in his philosophy class and has a bunch of TAs. But you multiply that by a factor of 10 or a factor by 100. And so I think the model will, which have already started, Tyler and I have already kind of started this, but, oops. The model will be kind of, you, you, you, you teach online, right? And then you show up at different places around the world for like guest appearances, okay? So I've been to India a couple of times. I go and I teach there and the old students all know me, right? It's kind of weird, right?

• 44:58Speaker 00

  But they're all, they're all Professor Tavarak, you know, they all, they're already familiar with me. They know how I teach and so forth. And then I just show up in the class. So I can kind of see a model like Paul Ehrdoz, you know, the famous mathematician who would just go around the world and co-author with different mathematicians, never had a fixed address. I can see some teachers doing that, just kind of going around the world. And they teach a few days in, in South Korea and a few days in India and kind of it builds up this community. Most of the teaching is online, but then, you know, you get to meet the person occasionally.

• 45:33Dwarkesh Patel?

  Ah, but with the Baumol effect, shouldn't I expect that? Because if education now becomes a progressive sector because of the substitute, then it's going to exasperate, you know, the increasing cost and things like TAs or coaches or whatever else that, you know.

• 45:49Speaker 00

  Yeah, so we're, so, yeah, we're, we're never gonna.

• 45:56Speaker 00

  As, so there are two effects going on at the same time, which Baumol nicely has both of them. And that is, you know, why is the stagnant sector becoming a higher price? Well, it's becoming a higher price because the progressive sector is, is growing more productive. But that means you're also getting richer. So there's a substitution effect and there's an income effect. Okay? So that's an interesting thing about education and healthcare.

• 46:24Speaker 00

  Is that even as the price has gone up, we're spending more on these goods, right? And that's really bizarre from just about the perspective of just about any other theory. Because if you think that the problem is a lack of negative productivity or regulation or something like that, right? Well, then, if education is becoming more expensive because of costs are going up, well, then you want to consume less, right? So then, why would you be consuming more when the price is going up? You need like two theories to explain this. One theory to explain why the price is going up. And then you need like another theory, well, we're becoming more credentialed or whatever, to explain why people are consuming more of this good.

• 47:07Speaker 00

  In Baumol, you're pushing out the production possibility frontier as it is becoming a more curved at the same time. So the price is going up, but you're also becoming richer. So Baumol explains both of these things very, very nicely. It's not contradictory at all in Baumol's theory that you would spend more on a good even as the price is going up. Because the reason the price is going up is in part because we're becoming more generally richer over time.

• 47:39Dwarkesh Patel?

  Right, and so it's becoming more affordable. But then you also point out in the paper that even though we have more teachers per capita now, math scores really have an increase.

• 47:47Speaker 00

  Right.

• 47:48Dwarkesh Patel?

  So shouldn't we expect like if teachers are going up, it's not just a cost per teacher, it's also the amount of teachers we have. Shouldn't that at least increase the outcomes we get?

• 47:56Speaker 00

  Yeah, so you're absolutely right. And if you were dealing with kind of a fixed set of students, I think that would happen. But we are, as I mentioned, you know, many, many more people are going to college than ever did before. And that just means like that the standards are going down and the students, the, I mean frankly, the students just need a lot more handholding today than they did in the past. Because in the past, I, I, I mean, you just had the elite, which was going to college, right? And these people could kind of take care of themselves. They came from wealthy families and could kind of take care of themselves.

• 48:39Speaker 00

  And now we're just having many, a much wider variety of students are coming to college. And they do need a lot more input. Wow. Now, now, there is another factor, uh, which is tied with the great stagnation, it, which is very worrying, is that we seem to be needing a lot more input per unit of output in, even in the progressive sectors. Like I mentioned, computers are becoming much more faster and so forth. But if you look at the number of researchers in the computer sector, like, yeah, yeah, kind of, you know, guys figuring out, I, I'm not talking about computer scientists. I'm, the guys who are working on the chips, right?

• 49:23Speaker 00

  Uh, the electrical engineers, I suppose, okay? You need many more electrical engineers to get the same growth rate today than you did in the past. Um, and the same thing, you know, I work a lot on pharmaceuticals. Um, I tend to, you know, focus on the FDA and the FDA kind of slows things down. But even putting aside that, the amount of research and development budget, which goes into creating a new pharmaceutical is much higher today than in the past. And it's not just a regulation effect. Uh, you also have just many more scientists, uh, it seems to get the same, you know, life expectancy increase from a pharmaceutical than you did, you know, in the past. It was more low hanging fruit in the past.

• 50:07Dwarkesh Patel?

  Right, yeah, as you know, and as Tyler Cowen wrote about, Nicholas Bloom and others wrote a paper where they talked about this, like, there's, we need eighteen times as many semiconductor engineers just to get the same Moore's law, oops, sorry about that, the same Moore's law doubling of transistors every two years.

• 50:23Speaker 00

  Yeah, exactly. It's very worrying. So, kind of, one hopes that maybe we will see kind of some quantum leap in some technology. You know, you kind of get a quantum leap, right, and then you get diminishing returns, and then you get another quantum leap. So, I'm hopeful that, you know, maybe it's quantum computers, I don't know, quantum leap, but, you know, hopefully we'll see some, some, you know, grow, some sector will explode. Otherwise, those facts which you just mentioned are kind of really disturbing.

• 50:57Dwarkesh Patel?

  Yeah, they mentioned that it was every thirteen years, you need to double the amount of researchers to get the same growth. So we could just have every pair of thirteen-year-olds just have four research kids.

• 51:05Speaker 00

  Yeah, yeah. And it's disturbing because, right, because you would have predicted that with many more researchers, researchers of public good, you know, ideas are free, they flow everywhere, that you would have had much, much more advance. And it just seems that we've only barely been able to keep pace, if that. You know, I am optimistic, sort of, you know, that with China and India becoming richer, that you get a lot more scientists and engineers in the world as a whole, and that will give us a boost. But so far, you know, it's not, it's not the huge gain that one would have hoped for.

• 51:47Dwarkesh Patel?

  Yeah, yeah. I just plugged those numbers into like, into some code, and it turns out that if, if the, if, you know, research productivity declines every thirteen years, or halves every thirteen years, you plateau at like forty percent increase or something like that. But if it doesn't, in the next century, you could get like six hundred percent growth, otherwise.

• 52:05Speaker 00

  Right, right. So there's a, so there's a small possibility of a very big gain if we get lucky, but the trend is not good.

• 52:13Dwarkesh Patel?

  Yeah. Oh, so let me ask now about dominant assurance contracts. Before I ask you some more questions about it, and we get in the weeds, do you want to explain what that is?

• 52:23Speaker 00

  Sure. So, so Dominant, it's kind of interesting. This is maybe is my most important paper, but it's kind of overlooked.

• 52:35Speaker 00

  Except recently it's grown a little bit, but this solves the public good problem. That's a little bit of an exaggeration. But, you know, Paul Samuelson said this problem was impossible to solve. And for a long time it looked like, you know, you couldn't solve this public good problem. And the public good problem, right, is that for a good which benefits everybody, non-rival, non-excludable, okay, there's going to be a free rider problem, okay? And so people won't want to contribute to it, even though it benefits everybody. People say, oh, I'm going to sit back, I'm going to let the other guy contribute to it, and then I'll benefit without having to pay the cost. And of course, when everyone does this, you don't get the public good at all, right?

• 53:17Speaker 00

  So how to produce public goods is a huge, huge problem. And it looked like this was impossible to solve. Now, I solved part of it with this dominant assurance contract, which is easier to explain today than when I wrote the paper, because when I, before I wrote the, I wrote the paper before Kickstarter, okay? So Kickstarter is now something almost quite similar, not quite the same, but similar. So in Kickstarter, you contribute towards the public good, and you only pay if enough other people contribute that you reach the threshold, right? You get over the, you get over the line, and then, then you have to pay.

• 54:00Speaker 00

  So Kickstarter solves one problem in that you're not worried that the other people, you know, just won't show up, and then your, your funds will be wasted, okay? Your funds are never wasted. Your funds only are taken from you if enough people show up to contribute towards the public good. Okay, that's an assurance contract. Here's the dominant assurance contract. It has just one simple twist. It says...

• 54:27Speaker 00

  If not, if you don't reach the threshold, then everybody who agreed to pay gets a refund bonus. Okay. So therefore, think about it. You're thinking about whether should I contribute to this project or not? Well, there's really only two cases you need to consider. One is if the project is successful, then I get the benefit of the public good. There's some benefit there. So that's a positive good to me. If the project is not successful, if not enough other people contribute, okay, then I want to contribute because I'll get the refund bonus and I won't have to do anything. Okay.

• 55:09Speaker 00

  So either way, you're actually better off contributing than not contributing. So what this means is that it turns an assurance contract into a dominant assurance contract, which means that it's now a dominant strategy. Okay. In some circumstances, if to contribute to the public good because you benefit either way, right? And so this means it doesn't solve all public good problems because it solves it for kind of a public good where you know the right size. So if you want to build a bridge, it's usually not too hard to figure out should be a two lane bridge or a four lane bridge. Okay. Or a lighthouse. Okay. You just have to figure out how much is it going to cost for the right size lighthouse?

• 55:54Speaker 00

  It solves the contribution problem. There are other public goods where you're not quite sure how much of it should you actually produce, like defense, like should it be 100 billion, 200 billion, 500 billion? How much do people actually want? This does not solve that, but it does solve the contribution problem. And what some colleagues and I, Tim Kason and Robert Zuberkis, what we've actually shown is this works in the lab. So we've run experiments and the dominant insurance contract is able to double the number of projects which are successful. So if you could put this on Kickstarter, you could probably double the number of successful projects. So it does actually work.

• 56:33Dwarkesh Patel?

  Compared to just a normal assurance market?

• 56:36Speaker 00

  Correct. Correct. So most projects on Kickstarter fail. We think some of them should fail. Some of them are just bad projects, but we think some of them fail because. The assurance contract doesn't solve all of these problems. But a dominant assurance contract, it would allow more good projects to succeed. That is, you have some project where the benefits are bigger than the costs. A dominant assurance contract will help you to get those good projects. It will help those projects be successful.

• 57:11Dwarkesh Patel?

  Yeah, it's such a brilliant idea and useful idea because like it takes the equilibrium from being towards being a free rider to actively contributing to the public good.

• 57:22Dwarkesh Patel?

  So what about projects, though, that are, you mentioned like a bridge, for example, right? But there's like a notorious problem of the inflation of the budget over time. So, you know, it was projected to be 10 million, but now we can't build it unless we get an extra 10 million. What do you do if that kind of thing happens?

• 57:39Speaker 00

  Well, one problem at a time, my friend.

• 57:46Speaker 00

  Yeah, I mean, that's going to happen whether it's a public good or not, right? So, you know, you're absolutely right that there is a tendency for these big dig projects to inflate in scale. And there's a, you know, there's a problem, especially in the United States, with why, you know, subways and construction and tunnels, why they're so expensive here compared to the rest of the world, right?

• 58:15Speaker 00

  And part of that problem is unions, okay?

• 58:20Speaker 00

  But there's a variety of reasons for that, unions and legalism and, you know, we've, you know, historical committees and we've, you know, we've just put on so many requirements to build that it's just become really, really difficult and expensive. So dominant insurance contract does not solve all problems.

• 58:41Dwarkesh Patel?

  So is the path for a libertarian-minded people like you and me to, instead of having to like convince a majority of people to adopt our ideas, to just like set up alternative institutions that are based on ideas like this?

• 58:53Speaker 00

  Yeah, so I mean, I think that's a very progressive way of thinking about it, right? Is that I've actually said that it could very well be the case that we need many more public goods, especially local public goods. And a dominant insurance contract is a way of bringing this out, of showing that this is true. So if we can allow, if we can create a regime under which not just dominant insurance contracts, but you know, Glenn Weil and Vitalik Buterin and a co-author, Zoe, I believe, you know, they have, they have, they have other schemes as well, right?

• 59:34Speaker 00

  So there are these mechanisms, some of which use the blockchains and which work online, some which don't work online, to produce public goods. And I would, creating a market mechanism, it sounds contradictory, but designing a market to produce public goods could change, you know, everything, right? And I think that would be a very, kind of a progressive way of thinking about things. Instead of saying, you know, the government's doing it wrong all the time, let's create an alternative institution which can do things better.

• 1:00:11Dwarkesh Patel?

  Yeah, yeah. The optimistic scenario here is, so Eliezer Yudowsky wrote a book called Inadequate Equilibrium, as I'm sure you know, and then he detailed three ways in which like a civilization can fail. Asymmetric information, which, and you wrote an article with Tyler Cowen on how that's becoming a smaller problem. Decision maker is not a beneficiary, which is not a problem because of dominant insurance contracts. And also, inoptimal Nash equilibrium, which is also not a problem because of dominant insurance contracts. Is this, so is this just like the future? Is this, is this how we solve a bunch of the problems, coordination problem, whatever else this is, there are?

• 1:00:41Speaker 00

  Well, well, I guess we'll see. Try to get the paper, you never know until the paper gets out into the world, right? So hopefully, I do think that...

• 1:00:56Speaker 00

  You know, democracy itself is a mechanism. It's a mechanism to both collect and aggregate preferences and to kind of exclude dictatorial. It's to improve governments. It's a governance mechanism, right? And we've now had, you know, 200, 250 years of experience. And we know it works well in some cases and less well in other cases. It's not a great preference aggregation mechanism. It is actually a pretty good way of limiting government. There are things which democratic governments don't do. Democratic governments do not starve their own citizens.

• 1:01:36Speaker 00

  That's a very low bar, but it's a low bar which many governments fail to meet, right? So democracy has a lot of benefits, but it's not the end of the story either. There are different types of democracies. There's different ways of shaping preference, different mechanisms. And particularly as we go online and as we live online, much more. I think many more of these mechanisms are going to become viable. And the nice thing about online is that it satisfies the taboo conditions. Taboo conditions are local public goods, where that is, you can move from world to world to world to world very easily at low cost online.

• 1:02:20Speaker 00

  So it's much less costly to set up a new type of government. You know, every new blockchain project has got its own governance system. Most of them are terrible, most of them are going to fail. But when before in human history have we had as many experiments with governance mechanisms as we are today? It's really quite extraordinary, like Tezos has got its mechanism and, you know, Ethereum has got its mechanism. And so we're doing much more experimentation in how to organize collectively, right? This is Glen Wall and his Radical Exchange project.

• 1:02:59Speaker 00

  So the collective organization is an unsolved problem, right? How do we act collectively to solve public good problems without becoming dictatorial, without becoming...

• 1:03:15Speaker 00

  You know, without falling under rent-seeking, you know, without falling under all of the problems of collective action, right? If we can crack that nut, that's a huge thing. And the plethora of online worlds and experiments is, I think, one way which we're gonna do this, and it hasn't been done. All right, the last time we did this was transitioning to democracy. Basically, after World War II. You know, World War II, I think there was like eight democracies in the world at the end of the world, at the end of the war, and now we have hundreds. So we've got many experiments in democracy. Some of them have worked well, some of them haven't.

• 1:03:57Speaker 00

  But now with online worlds, we can have thousands of experiments with new types of governance mechanisms.

• 1:04:04Dwarkesh Patel?

  That's so fascinating. By the way, what do you think of Robin Hanson's idea of a futarchy, where prediction markets decide which policies

• 1:04:10Speaker 00

  Right, right. So, you know, Robin's ideas are remarkable. You know, I often, you know, tell my students, look, in the whole history of the world, okay, you can kind of say, there's really only been like maybe four types of government. There's, you know, there's the monarch, dictator, there's rule by the aristocrats, oligarchy, right? And democracy. So there was, Aristotle understood all three of those. Then you had like maybe kind of some Rothbardian anarchism, okay? Rothbardian and David Friedman-style anarchism. So that maybe make four kind of governance systems. And then Robin has his, a new one, which is incredibly rare, right?

• 1:04:55Speaker 00

  So that makes five, which is futarchy. It's an entirely original idea for governance by futures markets. And he's created a lot of powerful arguments in favor of it. So I am hugely, I'm a huge promoter of futarchy to run experiments with it. I would love to see many more experiments with futarchy, with dominant assurance contracts, with kind of some of Glenn Wild's radical exchange ideas. You know, I want to see a lot more experimentation in these big ideas than we've had in the past. And online worlds may be one way of doing that.

• 1:05:33Dwarkesh Patel?

  Well, can we consider your idea of dominant assurance contracts as a sixth former government with like actual consent of the governed?

• 1:05:39Speaker 00

  It's pretty good. Yeah, it's pretty good. I wouldn't, I wouldn't quite put it up there with, you know, monarchy or democracy, but it's a mechanism. It's a mechanism. I'm not sure I can run the whole government with dominant insurance contracts. So, so I'm going to give Robin his, his kudos for creating the fifth, the fifth governance system. It's pretty, it's pretty amazing.

• 1:06:02Dwarkesh Patel?

  Yeah, yeah. So let me ask you about the, the different ways in which you've been creating content. So you write papers, you write books, you've written a textbook with Tyler Cowen, and of course you did the online videos with Marginal Revolution University. Which have you found to be the most effective way of communicating information to your audience?

• 1:06:22Speaker 00

  Well, it's pretty amazing when I get emails from students like around the world from, you know, India and Pakistan and, you know, South Korea and so forth. And, you know, they, they say, you know, "Thank you, professor." You know, "I was, I was able to pass my class because of you." And of course, I don't know who these people are, right? But so it's amazing that I have students that I, I've never met, I never, probably never will meet. Sometimes when I do travel, I get to meet some of these students, which is great, right? So this idea that a teacher can, you know, teach thousands and hundreds of students, thousands of students, millions of students, is pretty remarkable. And I'll tell you kind of a secret, Tyler, that, you know, between Tyler

• 1:07:07Speaker 00

  and I, our goal is to teach more people economics than anyone has ever taught in the entire history of the world. So we are right now, we're behind Adam Smith, okay? We're behind, we're behind Marshall, okay? We're behind Mill, and we're catching up to Manku, okay? So I think we might overtake Manku, which would make me very, very happy. So, and that's like possible. It's, I'm not saying it's gonna happen, but that's like, that's a conceivable dream that Tyler and I could teach more people economics than anybody else in the history of the entire world. That is

• 1:07:49Speaker 00

  actually possible. I don't know what's gonna, I'm not saying it's gonna happen, but we're climbing the ranks.

• 1:07:55Dwarkesh Patel?

  I don't know, haven't, have you not already done this? Because I don't know how many copies of The Wealth of Nations have sold and much less how many have been read. But like, I can imagine that Marshall Revolution University has had more views than there have been reads of The Wealth of Nations.

• 1:08:11Speaker 00

  Wealth of Nations has been around for a while. It was pretty best seller in its time. So I think we have a while to catch up with Adam Smith. Indeed, the world has, the world would do, would do itself a favor by learning it's Adam Smith, which it still hasn't done. So Adam Smith still has some lessons to teach. So we're catching up, but we're not there yet.

• 1:08:32Dwarkesh Patel?

  Yeah, also lessons that can be learned through your videos as well, though.

• 1:08:36Speaker 00

  True.

• 1:08:36Dwarkesh Patel?

  Yeah, so, okay, so to close out this interview, which has been really fascinating. Oh, actually, before I ask the final question, I want to ask, so you have all these ideas, dominance insurance contracts, prizes for increasing innovation during pandemics, tons of other ideas, right? Veriolation from Robyn Hansen. Is it frustrating to be an economist and you're putting all these ideas out there, but like, very few of them are adopted by the government and you know how useful they would be? How, what is it like to be an economist, putting out all these ideas out there?

• 1:09:06Speaker 00

  Yeah, it's frustrating at times, for sure. Like I just said, you know, the world does not cut up with Adam Smith, let alone with dominant insurance contracts. I do think that in economics, we are very fortunate that the skills which we are taught as a profession are widely applicable to a large range of fields. They're very, they're very general and widely applicable to a large range of fields. And we do have credibility in policy and with the administration. So while I don't feel that, you know, you know, I'm super listened to or whatever, but

• 1:09:47Speaker 00

  I, you know, I found myself talking with people at the White House and the Council of Economic Advisers, me and Michael Kramer. So I don't know, you know, Operation Warp Speed, I don't know how much they listened. But maybe it's, you know, in part being in Washington, I feel I'm certainly not, you know, at the heart of things, but I, you do feel in Washington that you're sort of close to the pedestals of power in some ways.

• 1:10:22Speaker 00

  You know, Koest said that that an economist could earn his entire lifetime wages just by stopping one bad idea.

• 1:10:34Speaker 00

  So I guess I feel I've done that. And that's probably maybe that's maybe that's enough to justify.

• 1:10:39Dwarkesh Patel?

  I find that very comforting that you're being consulted with. Because all I see is Paul Romer just getting incredibly upset at how little his advice is listened to. And then I feel bad that if other advice isn't listened to as well.

• 1:10:51Speaker 00

  No, absolutely. I mean, and, you know, Romer has been pounding on testing right from the beginning. And he's been absolutely right. And

• 1:11:01Speaker 00

  yeah, I've been incredibly frustrated talking with people in Congress. And they're just and they're always like, well, nothing's going to happen in this bill. But maybe in the next bill, it's I don't understand it. It does seem that

• 1:11:15Speaker 00

  we it does seem something is wrong, to be frank. It seems that something is wrong. And we haven't solved that problem. But it, yeah, I hate to end on a negative, on a negative note. But yeah, there's something wrong.

• 1:11:31Dwarkesh Patel?

  Well, can you expand on that?

• 1:11:33Speaker 00

  Well,

• 1:11:35Speaker 00

  look, this is a part of what, you know, Tyler called state capacity libertarianism. Okay. And what is that? This is something that I've written about. I didn't use that term, but I've talked a lot about is, look, I want a small government. But I want a government to be able to do what it's supposed to do at the time it's supposed to do it. Okay. And even I was shocked, like the CDC, their entire reason to Aethra, right? Their entire reason for existing is to stop a pandemic. And yet they completely failed. They botched the first test. Okay. And then the FDA came in and said, oh, private companies, you cannot use your testing. You have to apply to us to do, you know, to get approval from us to do your test.

• 1:12:19Speaker 00

  And that slowed everything down at the beginning of the pandemic until the virus got ahead of us, right? And it just seems like, you know, in the past, the government might have been smaller, but it was able to do things, sometimes bad things for sure. But the way we ramped up for World War II, you know, is a remark was a remarkable achievement. And it just seems that, and it's not all Trump's fault, okay? Trump is terrible, whatever, you know, you blame him as much as you want, okay? But you cannot blame the failures of the CDC and the FDA all at Trump's doorstep. You know, Congress has completely failed as well, right? Congress is the one supposed to be passing the law.

• 1:13:04Speaker 00

  Where's Congress's testing plan? Where is Congress's vaccine plan? Why is it that the, the only good thing that has come out of the administration on the pandemic is Operation Warp Speed? You know, why didn't Congress do that? That's actually Congress's job, okay? So Congress has completely failed as well. And there's a lackadaisical attitude, there's a complacent attitude, which, given that some of them are getting the virus, I, I completely fail to understand, you know? Like Boris Johnson, you know, he actually got the damn thing. You'd think he would want to solve this problem. And so it, it seems that not only are we not getting the government that we paid for, right? We're getting less, okay?

• 1:13:47Speaker 00

  I at least want to get, you know, something for what, for what I pay for it. And, you know, pandemic assistance or pandemic responding to a pandemic, responding to a war, they're high on my list. Like, I'm worried now, like, the way I would never was before, that, you know, we would lose a war with China or something like that, okay? You know, that never even occurred to me, that the United States, you know, the greatest military superpower in the history of the entire world, right? By far the largest economy, highest per capita, you know, incomes and so forth. To think that we might, you know, lose a war.

• 1:14:30Speaker 00

  It would never have occurred to me, but now I just think things are so dysfunctional that anything is possible. And it's, well, I am worried. Yeah, I'm worried. And it's not just about the pandemic, but I think that is just the canary in the coal mine, which tells us that this government is not working and it's not all gonna be. The danger is that with a Biden administration, you know, the liberals will be so happy that all thought of fundamental change will end, which is what happened, for example, under Obama. You know, under Bush, there was a big anti-war effort, right? Okay, you know, big anti-war protests and movement.

• 1:15:14Speaker 00

  And Obama came into power and that all went away and the war kept going, right? So, you know, we're still in Afghanistan, we're still in Iraq, still, you know, bombing countries around the world. So, Obama did not solve the problem, but the protests went away. And if Biden is elected, that's not gonna solve the problem of complacency. That's not gonna solve the problem that the government is riven with legalism and bureaucracy and inefficiency and slowness, but it may take some of the force for change, take the window of the sails, and that is a concern.

• 1:15:51Dwarkesh Patel?

  Wow, okay. What's a way we make our government more effective? Is there some solution?

• 1:15:58Speaker 00

  You know, I don't have a solution, but one thing which I think is gonna be helpful is actually competition with other countries, such as China and India. You know, like what really made us, why did we go to the moon? Okay? We went to the moon because we were competing with the Soviet Union, because they kicked our butts by putting somebody into space and made us look like fools, made us look like technological laggards. So we got our act together and we devoted a huge amount of money to innovation, right? To NASA, a huge amount of the budget, like 12% at the height, 12% of the US budget was going to NASA, which, you know, you can complain about whatever, okay?

• 1:16:43Speaker 00

  But it was going to innovation. So one of the things that I've said, you know, my book, Launching the Innovation Renaissance, is what we have today is a warfare-welfare state. Okay, we do two things. We invest a lot in the military and we invest a lot in various forms of welfare, by which I'm including, you know, Medicare, Medicaid, and redistribution and whatever. And I'm not even gonna say it's bad or whatever, okay? But what I would like to see us be is an innovation state. And as the warfare and the welfare state have grown as a share of the budget, the innovation state has shrunk. So we're investing much less in research and development, federal dollars, in research and development than we used to.

• 1:17:24Speaker 00

  And I would like to see it to shift us away from warfare and welfare and towards innovation. And when we see China start to kick our butts, where China's developing new pharmaceuticals, China's developing more official intelligence, China's developing genetic engineering, like what are we gonna do when the first Chinese with, you know, super IQs of, you know, 160 plus, you know, start marching out of their factories, okay? You know, hopefully that will kick us into gear. And we'll say, okay, we've gotta, we've gotta, you know, respond. So, you know, I'm not one, I'm much more about cooperation than I am about international competition.

• 1:18:07Speaker 00

  I don't think we're at war, literally, with other countries. I think our interests actually align. But in terms of igniting the passions of a nation towards innovation and away from complacency, competition in the sense of the Olympics, a good kind of competition, if we can keep it, if we can keep it to that kind of competition and not to, you know, blowing each other up. That kind of competition, I think, would do, would do a lot of good in reigniting America to become kind of a, the leading nation of the world, not just in terms of innovation, but also in terms of freedom and liberty and, you know,

• 1:18:51Speaker 00

  racial and sexual equality, all of these things which, in many respects, not, not all by any means, but in many respects, we've led in the past. And to kind of reignite that, you know, to say like, bring all people from Hong Kong, let's bring them to the United States. Let's bring the Uighurs to the United States, okay? Let us be that shining beacon on the hill. That's what I would like to see the United States be.

• 1:19:15Dwarkesh Patel?

  Yeah, yeah, I had the pleasure of interviewing Caleb Wadney, who made a similar point as you, that the path dependence of technologies are important, and so it's important that it happens in a free country. Now that you mentioned launching the innovation, do you mind if I ask you a question about the book?

• 1:19:26Speaker 00

  Sure. If your audience can take it, I can.

• 1:19:31Dwarkesh Patel?

  So you explained that both the supply of innovation in terms of global talent will be increased because of globalization, and also the demand, because we'll have a global market. So then, how do we explain the great stagnation, which happened during the same period as the time of greatest globalization?

• 1:19:49Speaker 00

  Right. Yeah, so this, so all of these trends should be very, very positive, right? Because, you know, the way I like to put it is that as China becomes richer, they're all going to be looking for a cure for cancer too, right? Okay, this is a problem which affects everybody. And so you have a much larger market, a much bigger R and D market. And you already see in China, leading technologies, Tencent and so forth, Baidu and things like that. So they're already investing a lot in these technologies, which are public goods. Artificial intelligence is another one. So I think we have a lot to gain by

• 1:20:28Speaker 00

  this increased budget devoted to R and D. What I didn't, what I didn't realize, I suppose, is that there is this countervailing wind, which we now know from the Bloom and Van Rhenan work, we talked a little bit about earlier, that it does seem there were a bunch of low-hanging fruit, and we are having to devote more and more resources to get the same output. Now, whether that's a fundamental technological factor, whether it's a regulatory factor, whether that's a temporary factor in response to these quantum leaps in technology, like we got electricity, we got the internal combustion engine, and it takes 50 years to kind of grind everything out we can out of them. Now we've got computers, it'll take

• 1:21:13Speaker 00

  some time to do that. Maybe biology is the next one, or as I said, quantum computers, I don't know what it's going to be. But there may be a next technology where, again, that gives us kind of a leap. Who knows? What I can say is that the globalization has been good for research and development. We would have been much worse off if we didn't have that.

• 1:21:39Dwarkesh Patel?

  Okay, good. And the final question, which I'll ask all my guests is, what advice would you give to a 20-year-old, or to yourself when you were 20, or just to a generic 20-year-old?

• 1:21:47Speaker 00

  Yeah, well, I have a generic 20-year-old. I got two of them.

• 1:21:53Speaker 00

  So, look, the world is changing faster, so probably my advice is becoming, you know, the advice of older people is becoming less useful over time. You know, it's a, don't play those video games. And now, like, video games, you know, you could be, you know, you can earn a lot of money playing video games.

• 1:22:12Dwarkesh Patel?

  Right.

• 1:22:12Speaker 00

  So I guess I got that one wrong. I should have told my kids, you know, learn how to play this video game, be a sports star, be an esports star. Look, the return of the skill is going up, so education is, continues to be extremely important. Get an education in a, in a sector which is complementary to technology, okay? So you don't want to be competing against technology, you want to be racing with the machine, to use Eric Bryngdolphensen's and McAfee's term, you want to be racing with the machine. So if

• 1:22:52Speaker 00

  you can get educated in a sector like electrical engineering, computer science, or economics as well, but in a sector which is complementary to technology, data, data science, so if you're able to, like, what is increasing in the world faster than data? So if you were able to analyze data, that is going to be a incredibly beneficial skill, because that's, we're getting a lot more of it, right? And extracting meaning from data is very difficult, and making it accessible to human beings. So if you have some skills which come out of economics, you know, causal inference, but also out of data science, machine learning,

• 1:23:35Speaker 00

  and so forth, if you have some skills to extract information from data, that's going to be very valuable in the world going, going forward. So it's not, these are the areas where I have some sort of expertise in, it's not all that. You know, marketing is going to continue to be important, oddly enough, right? You would think that that's going to go, no, that's marketing, things like that is going to be important. Design, right? So what is Apple? Apple is just a great design company, right? Just an absolutely fabulous design company. So, and that's complementary to technology, right? So you want to take the technology and find a way of putting it in human hands,

• 1:24:16Speaker 00

  which is creates a delightful, satisfying experience, right? That's very airy-fairy, and yet that's an incredibly valuable skills. It's not just hard science, but if you have an artistic impulse, then I would say that's fine. That's also going to be valuable if you could combine it with a technological field, right? So an artistic impulse in design, that is going to take you much, much further than in like just poetry, let's say, okay? So you want to combine these things.

• 1:24:54Dwarkesh Patel?

  Oh, okay. Very good advice. And thank you so much for being on. It's an incredible privilege to get to ask questions to somebody of your expertise. So thank you so much for your time.

• 1:25:04Speaker 00

  Oh, great, great being here. And you ask great questions, so thank you very much.