An Interview with Tyler Cowen and Paul Krugman

Thank you all for coming. About 2 weeks ago, Paul and I already did an interview which you can read or listen to online. This is going to be a bit more like a conversation. I just thought I'd start by throwing it to Paul. Paul, what do you feel like talking about?

Because we have We have no moderator. We're interviewing each other, which is more or less—

yeah, it's called talking.

So actually, we each interview ourselves. That'd probably be— anyway, never mind. Okay, wow. So I thought I'd start— we're not here to argue, but I thought it might be interesting to find—

absolutely—

talk about an area where we actually did have some disagreements once upon a time. So you and I were both very involved in I now hear people saying about the old days of econoblogging.

Sure.

Which, you know, that was back when people actually had 600-word thoughts as opposed to 280 characters. But there was a big debate over the nature of the problems we were having in the aftermath of the financial crisis about high unemployment. And you and I were pretty much on opposite sides there. I was a very much straight Keynesian, this is just not enough demand. You were pretty big on the structural, you know, that there's been a big change. A lot of the workforce just doesn't have the right skills to be employed. Here we are finally, you know, unemployment back below 4%. Have your views changed? Have you? And if they have, what do you think actually went on?

I agree the demand side has been vindicated. So my view at the time was maybe 50/50 demand-supply. And now I'm probably more like 70/30 demand-supply. I think how long the recovery has taken is surprising, and it implies there was a coordination problem behind the aggregate demand problem, and that wages really just haven't gone up for non-managerial workers is itself an indication there's something wrong with labor markets. That's not just demand because nominal GDP is way above now. Its pre-crash peak. So I still think it's 70/30. I always viewed you as stressing too much demand and not worried about supply at all, but I've definitely revised my view toward demand having been more important than I thought.

Okay, yeah, that puts you way out of line with the rest of the profession because my—

I don't think so.

No, no, the rest of the profession, my, my general view has been that nobody ever admits having been wrong about anything.

Oh, okay.

So by actually, by actually changing your views, uh, So, no, that's interesting. It's funny, I would have said that the absence of, 'cause it's not as if any significant group has had big wage increases. Highly educated have not done particularly well, so I would have said that the notion that we had a skill shortage just doesn't fit, still doesn't fit the data. So I guess I've changed my view. I thought it was 90% demand, now I think it's 100%.

—Okay, so we've moved in the same direction at least. —Yeah, that's interesting. —Have you ever wondered, by the way, if every class should be taught like this? I know it costs 2x, but you have 2 people and they talk to each other and they disagree. And rather than just teaching, well, x is true, students get a feeling for what's at stake and it somehow becomes more vivid.

—Boy, when I was in grad school, we had econometrics was jointly taught by by— people probably don't know— but Bob Hall and Jerry Hausman. I can imagine. Jerry Hausman talks at about 180 words a minute and Bob Hall at about 20 words a minute. And so we had a student skit where the two of them proved the Gauss-Markov theorem simultaneously, starting together and ending together, but with the Jerry Hausman character saying about 9 times as many words as the— anyway. But no, I mean, it would be kind of— I think actually it would— I have had a couple of courses that were joint taught, and it actually kind of was interesting. I'm not sure it was the best pedagogical method, but I think it does liven things up.

Let me tell you why I don't think it's 100% demand, and maybe you could clarify your view. When I look at data, say, on earthquake shocks in Japan or other negative real shocks, they might be small, they're not the main problem, but the economy seems to respond classically to negative real shocks, whereas the 100% demand models could even imply the negative real shocks would be good for the real economy. —and it doesn't seem they are. Do you see what I'm saying? Oh, OK.

No, there is the argument that says that a natural disaster could actually be positive because it could— because of the spending that it induces. Right. And you don't really see that, but it's not clear. That's actually not a bad— that's actually a good story, although I don't think I buy it because I think in practice the wealth destruction from natural disasters depresses demand as well. And it takes a whole lot of government spending or whatever to make up for that. Actually, the one place where I would— I'm a little confused. What I would not have expected, but I'm not sure that it invalidates the demand story of the crisis, but I don't know if you looked at this, but how far below, how far we are below where we were expected to be. If you look at projections, CBO projections of GDP in 2018, we're about 10% below them now. So we're as far below to previous pre-crisis expectations now as we were at the depths of the recession, even though the economy looks like it's at full employment, and that's a, I don't know, do you have a story about how that happened?

Oh, but that I think's a productivity growth problem, that ongoing revisions, productivity growth has been systematically lower than we had expected. It's not inflationary because credit growth becomes lower too, and we're stuck here for now, and I think that is mostly a supply-side problem, though demand and supply-side factors interact and they each make each other worse, right?

We can go on, 'cause I would actually say that the problem is you see that kind of shortfall in every country, but the size of the shortfall is very different, and it's basically one for one with how deep the recession was. But a lot of— So Greece, if you believe the estimates, Greece has suffered a 25% decline in productive capacity relative to where, and is it just a coincidence that Greece, which also had the disastrous recession, suffered that? So that's, anyway, it's an interesting, there's a lot that's still unresolved. I'm glad to see that, that you, you do think that, that some— as I say, it's, it's a very worthy thing to see you making some revision and, and views, because that's been one, one of the big surprises, I think, of the past 10 years has been how few of our colleagues are willing to do that.

I agree. Let me see if I can get you a little more optimistic on crypto, right? So I share a lot of your, a lot of your skepticism. I don't own any Bitcoin or any other crypto asset. Okay, my modal forecast is that none of it goes anywhere, but is there not some chance that we can use blockchain to develop some regime of self-enforcing smart contracts referring only to the digital world, enabling micropayments, helping people establish their identities, lowering transactions costs? And they're saying a 10 or 15% chance this is actually a major breakthrough, more in governance than currency. Like, forget about it as a currency. Think of it as an alternative to law or corporations. It's another way of making decisions. Yeah, so I would—

when I want to be a crypto-skeptic skeptic— so if people don't know, I mean, that the whole— so I'm, I'm— yeah, I've been basically of the view that Bitcoin is nonsense, and a little bit less over the top than Nouriel Roubini, who's Tyler Cowen is actually very entertaining on this, but I get invited to crypto conferences as the designated enemy, because I'm polite and I say very politely this is nonsense.

I'm amazed you go, by the way.

They pay me in real money. But the argument that— my problem— there are two things here. One is that Blockchain is technically ingenious. It's not totally clear what problem it solves.

But— But think of transactions.

But think of it as a way of creating— one way of thinking is a way of creating a digitally, you know, something out there that is digitally verified, doesn't require any extra party. But then it's not— so there's two problems. One is it's actually— it remains quite expensive to do stuff that way. It's not actually a cheap technology. And the other, which I used to say that it was untethered to reality, and then someone came out with this version called Tether of cryptocurrency, so.

Which is the most dangerous version of crypto.

But there's a lack of an anchor to that, and so if people decide that the value of one, certainly if people decide the value of Bitcoin is zero, it's zero. That's it, it just disappears.

But that's true of fine art also, right? —still a store of value.

No, but my example is that actually gold, you know, gold has very limited real-world usefulness relative to the value out there, but it persists because for centuries and centuries, thousands of years actually, people have thought it has value. And there's a lot of other things that are really pretty bad assets. You know, the ratio of US currency in circulation to GDP. I've been interested, because if you had asked I think most people, if you asked me if I'd been in 1960, except I was 7 years old, but if I'd been me in 1960, I would have said, well, cash will be steadily declining as transactions become more sophisticated, which is what you do see, say, in Sweden. But in fact, the dollar value of currency relative to GDP has been rising in the United States.

In the U.S., but that's traded—

That's all— Foreigners. Well, it's all $100 bills., and so it's hard to think of anything that's a sort of less productive asset than a vault full of $100 bills sitting somewhere.

There are now more $100 bills than $1 bills, believe it or not.

Yeah, it's all of the, so everything under $50 bills has actually been declining relative to GDP, but that is saying that there is a demand for anonymous assets that are, and if some one cryptocurrency comes to fill that role, it can indeed just hang there in midair, and that I think is the case for thinking. Now, I don't think that's actually gonna happen, among other things 'cause there are so many people out there trying to market rival cryptocurrencies. The blockchain itself as a way of, you know, independent of currency, that's, it's not entirely clear what problem it solves.

But again, there are a lot of things we might want to transact for— so micropayments for media stories, or online proving your identity, which could be quite difficult. Like, oh, send me my password, two-factor authentication. Oh, but then Facebook gives away your phone number. It seems to me there's a lot of problems with transacting. I'm not saying the blockchain can solve them now, but if proof of stake could displace proof of work, which is energy costly, and you get some other advances— it's all basically 10 years old. A lot of major innovations, even the internet, can take decades. So, I'll give it a 15 to 20% chance that it's important. And is your number zero or what number will you give it?

I would give it 5% just on the general grounds that there's always a significant chance that my basic model of how the world works is wrong. And it could be, but I have to say it's actually interesting because I do go to these few of these crypto conferences and listen to people try to make the case for why blockchain is useful. Are uniformly really bad. Unconvincing, I agree. You know, it's this, that, this way you can go to the pharmacy and be sure that the drug you're getting is the one that was prescribed. Well, that's not really usually a problem. The real problem is that your doctor prescribing the wrong drug, and that's, blockchain won't ever help you with that. So, but maybe, I mean, it could, it's interesting. And the other thing is I think you do need, motives don't really decide truth, but looking at somebody's motive for advocating something may help to, as a kind of screening device, how seriously to take them. And crypto is this sort of combination of techno-babble and libertarian fantasy that does not inspire confidence in the people who go for it.

I've made no money in crypto, and I have young students who I think are millionaires, and they don't take me seriously on crypto because I've made no money on it. Whether that's a rational inference or not, you all can judge, but—

Yeah. Education. Let's try— let's talk education.

I have a colleague, Bryan Caplan. He thinks 80% of higher ed is signaling. You don't teach people things, but you certify them. Now, my view is about 35% of higher ed is signaling, which is significant but not massively high. That some of it— MBA programs, people are jumping through hoops. You're simply proving that you're smart in some way without learning very much. In some MBA programs, none of those represented in this room, of course. What's your view on signaling in higher education?

I think it varies a great deal by school and by type of school, by level. I think I spent, all right, I spent 15 years at Princeton. Princeton is, although there's a lot of fine teaching going on, it wouldn't matter if in many ways if all Princeton taught was ancient Greek because it's mostly about validation of social class. If the time comes when, I actually wrote an article about this a long time ago, if the time comes when all the things that we think are highly skilled can in fact be done by artificial intelligence, the elite schools will do fine because they will be there as badges of social class. And dating clubs. Yes, definitely. I'm now at City University of New York, and the impression I get is that that's a lot of real education going on. It's not a very prestigious thing even from the schools that are— leave aside the Graduate Center, but at the schools that are really very good like Hunter or Baruch, are, I don't think that those carry the kind of credentials weight except that you went to school, but there's a lot of real education going on and there are huge, Raj Chetty has this amazing stuff about which schools are engines of upward social mobility, and the elite schools are not because although if a poor kid gets into Princeton, great things happen, but not very many do, whereas places like some of the UC schools, Some of the CUNY schools are huge engines because they take students from poor backgrounds and move them up, and I believe there's a lot of real education that goes on. So I'm actually relatively— I'm willing to buy the more than 50% is signaling for elite schools, but not at all for the kind of place where I now sit.

What else are you thinking about these days?

Oh, well, let's talk— can we talk, since we happen to be timed for the Nobel Prize? Sure, Bill Nordhaus.

I thought they were two great picks. They both showed economics matters for the world. They show that growth and environmental sustainability are two sides of the same coin. I love that aspect of it. Both are super smart people and have also been willing to change their minds on lots of things. Yeah, and for those who don't know, I'm—

I have a personal connection. Bill Nordhaus was my original mentor. I was a junior in in school and took a course from Bill Nordhaus and Tjalling Koopmans on the economics of exhaustible resources. And apparently I struck Bill as being a bright, if very nerdy, kid and went to work as his research assistant in the summer of 1973, which was the greatest apprenticeship anyone's ever had. And it's— it is— It actually tells you one— there's nothing quite like having a mentor, a personal mentor in economics or any other field. It was a huge thing. Little known, do you know about Bill Nordhaus' thesis?

Well, which one? His dissertation thesis? Dissertation thesis. I don't think so.

Oh, it was actually Paul Romer before Paul Romer. Really? He had a dissertation that was all about endogenous innovation in which people—

Is this the price of light work?

No, no. It was about, he had models of an economy in which people invest in order to get patents, and the patents, it was all, it was actually all the stuff that Paul Romer was gonna do except that there weren't any monopolistic competition models available at the time, so Bill did it in this kind of awkward marriage of micro and then a macro framework that didn't really grow out of the micro, although it was plausibly linked, but in effect all of the ideas that became the Romer growth model were laid out in Bill Nordhaus' thesis. I'm not sure the Nobel Committee knew about that, but it's actually funny.

Yeah, it wasn't cited.

Yeah, I'll go back and look at that. Yeah, it's an amazing thing because I remember reading that and thinking this is great, but where's the general equilibrium? And because You couldn't at that point given the tools available, but it was actually a pretty amazing thing.

There's a video of Paul Romer giving a talk, I think from the early '90s, where he goes through Nordhaus's work on the price of light falling so dramatically, and he relates it to his own work. But what made Nordhaus such a good mentor, and how much of mentoring is simply someone believing in you as opposed to them guiding you?

Oh no, no, this was— and this actually comes very much back to what Robin was saying. It wasn't them— him believing. It was— it wasn't him saying you're great and promoting. It was him putting me to work, saying, "Okay." I remember the first batch of stuff I came, I was immured in the geology library because he was working on energy at the time and I was digging up all of this Bureau of Mines stuff on energy resources and technologies for alternative energy production. I came back with some stuff and he was just saying, "Okay." You've given me some papers that I could read and so on, but you need to figure this stuff out. You need to bring me a way in which this work and give it to me in a form I can use, which was actually the perfect thing. This was saying, learn, you're not my library assistant. You're supposed to be working as an economist. And of course, I got to watch, also watch him work it out, because he came into that that anybody's ever seen. It's this— the Allocation of Energy Resources, 1973 paper. And it's a— he had, at the beginning of that summer, he had not figured out at all how to do this. And the way in which he over time figured out how to take all of this data and put it together into an intellectually coherent and numerically workable computationally workable framework was incredible to have a ringside seat in that. And it was in effect like working out— worked out exercises in a textbook. You were getting to see the process by which this stuff is actually done, although at a higher level than we hope— than you're going to try and put on Econ 101 students.

Now if you think of Paul Romer's increasing returns models, and of course your increasing returns models. They're broadly similar, but is there some core difference between the two in terms of testable implications, and what, what assumptions would that difference come from?

Okay, actually, let me ask you, what do you, what do you think are the testable implications of Romer?

That as market size increases, you should have more innovation, which doesn't seem to be as true as any of us might have thought because the global economy is much larger, productivity growth still appears sluggish, or in a pharma, new drugs slow down. I didn't expect those things. And that rates of growth actually accelerate over time, which at least the 1990 Romer paper mentions as evidence in favor of his view, which maybe it was at that time, but it certainly has not been since then. And I think of Romer as actually having stepped back and saying that he himself is not sure how much those are empirically true papers as opposed to, you know, models.

That's what I was gonna ask. What predictions from it do you think are right or surprising?

I think regionally and sectorally specific, the Romer model and some of your models too help explain something like Silicon Valley or the growth of some parts of China, but not like the world as a whole, as a large macro picture. How do you do actually—

how do you do growth? I don't know how your textbook does growth. I haven't looked at that chapter, but the long-run growth—

institutions, property rights, but also inventiveness. We talk about culture, but growth is tricky, as you know. Robert Solow has pointed out, another indirectly of your mentors, what economists understand about growth changes. It's hard to predict which countries are going to take off. You feel you know growth miracles when you see them, and, and you can identify ex post what they did right, but China has surprised us all if only with the persistence of its growth. To have thought post-1979 they're gonna do much better was a normal obvious prediction, but that it would be so many years above 7%, to me is still a shock.

Yeah, 'cause I would, what I would say is that what's remarkable is that, 'cause I remember the excitement over new growth theory, And yet in practice what people mostly do is Solow. Sure, and we—

Solow's the main thing we teach. And we, yeah.

So do you know my favorite Bob Solow quote about country growth rates?

I'm not sure, the one about sociology?

Yeah, he said, it was actually about Britain's lagging performance, but it applies to lots of things, that most such discussions end in a blaze of amateur sociology.

You know, Jeff Sachs has a review of the Asumoglu paper Robinson book, and he says if you go back to 1960 and even knowing which countries did well, try to predict which ones are going to do well using markers, it's very, very hard to do. Yes. And it's certainly even harder ex ante. So people think, you know, Philippines or then Ceylon were going to be the big winners. That was a common view. It seems plausible at the time. It certainly has not come to pass.

I'm actually old enough to remember when people used to write earnest tracks explaining that Confucian culture was fundamentally incompatible with economic growth, which doesn't look too good in the modern world.

Yeah. And what is it about growth you think is so difficult to explain? Because it is the single most important topic in economics. Yeah, that's what—

that's— yeah, that's— I mean, the thing I would say is that and I'd like to know your view on it, is that fundamentally what it comes— in the end, we just do not know very much about what total factor productivity— we don't know very much about what causes productivity growth to rise or fall. We don't know very much. Actually, we don't even— the truth of the matter, actually, leaving aside growth, just look at level differences among countries. Because you have huge, still, A huge difference is the basic technology is available everywhere. Everyone has access to the book of blueprints, we used to say. And yet you have vast differences among countries. And at a fundamental level, we don't know why some countries do much worse than others at making use of that available technology. I think what you would want— the trouble is, I think economics, which suggests that people take advantage of opportunities, and if there's a better technology, people should use it, has a very hard time coping with the evident fact that in many cases they don't. The countries, or at least they don't seem to be able to find a way to take advantage of these possibilities. Whatever it is that makes workers in— in the United States, you know, 10 times as productive as workers in a lot of the emerging world still. The fact that countries don't find a way to close that gap, and would seem to be such a— so much gain if they can do it, remains a deep puzzle.

I would give a slightly different, but I think not incompatible answer. You know, people love to debate big government versus small government, but when I look at the data, it seems to me quality of governance is all-important. So Brazil is a pretty big government, but it's very poorly run, and so no one wants to defend it. Denmark maybe has a government that would appear too costly to support, but they manage just fine. And quality of governance can change actually fairly quickly in some cases. And we as economists don't really have good theories of the quality of governance. Now, I think this leads back to your TFT. If you have poor quality of governance, if you know that a Parenti-Prescott model of monopoly. You put up barriers to keep your domestic monopolies. You keep out TFP because TFP threatens your domestic monopolies. And that's why the poorer countries don't adopt better technologies. They're not unwilling to. They're very happy to use cell phones when they can do that in standalone fashion. But they have poor quality of government, governance, domestic monopolies. TFP is kept out, but we can't explain governance very well.

Yeah, I would actually— how much travel do you do in developing countries? Oh, lots and lots. Okay, sorry, condolences.

But— 100 countries maybe, I'm not sure.

Yeah, I don't know how it strikes you, but what always strikes me is actually the high technology is all there. And it's the little things like actually managing to have roads that aren't— that are passable, that are where things fall apart. And as you say, it's very hard to explain. Explain why that is so different among countries.

But there's some high technology. There's not nearly as much, and quality of management is much lower. There's lower trust, which is a kind of governance, so you cannot delegate so easily. Everything goes through rigid hierarchies, even in the private sector. So even using the same technology, an Indian factory can be one-tenth as productive as, say, a Western factory in some, but not all cases.

Anyway, actually, I wanna change subject. Sure, yeah. When we did our talk, you were asking, and it was a good question, about how do we think about these tech companies that are nothing like the theory of the firm that we teach pretty much in our textbooks with zero marginal cost, network externalities. So I was doing a little more thinking about that. It relates to— I'll give you my concern. So if you look at the most valuable companies in the world, most of them now are companies that are nothing like the theory of the free market. There's no U-shaped average cost curve, and there's either zero marginal cost or at least huge marginal cost is way below price. They also, they're subject to huge Externalities of some form are central to the business, network externalities, and in a lot of cases they actually make their money not really by selling what they sell, but by selling, you know, so Google is an advertising platform and so on. That's one sector. Then if you ask about where are people employed, where is money spent, well, 17% of GDP is spent on healthcare., which is a— Kenneth Arrow told us 55 years ago— is a sector which none of the things we assume about what's supposed to make markets work right applies, and it's all asymmetric information of various kinds, moral hazard. The Lemons problems are— all of the institutional structures are built about trying somehow to deal with those. So between those, we're already talking about a very large part of our economy, and in both cases, it's not that, well, there's a market, but then there's a market failure. It's like there's all, it's all market failure. There's almost none. Are we teaching economics wrong is my question? Should we, I don't actually, I'm not planning to do this to the next edition of Krugman and Wells, but the question is, Are we underemphasizing the extent to which a large part of our economy looks nothing like supply and demand and the, the Viner model of the firm? What, and if so, what would you do about it?

Depends who the we is. Now I have not looked through your last edition. In our previous edition, we put in a whole new chapter on network goods where we teach people how to analyze Google and Facebook, telling them it's not You know, your grandpa's story of monopoly or market failure.

We actually had that in first edition and no one seemed to actually teach it, so we kind of dropped it.

Bring it back. It's very important. You know, healthcare, a lot of books including ours, we cover under adverse selection and moral hazard, but even that's tricky because Arrow's model of asymmetric information implies we'll spend too little.. And that's not the case in the US. We spent too much and in the wrong ways. So there's some public choice problem where we've multiplied lobbies who all want a lot spent on them. And as you know, we spend much more than Western European nations. Yeah. And some of it is return on innovation, but a lot of it really is just waste. And I think public choice angle is essential to understanding healthcare, not just market failure. It's like you've market and government failure pushing in the same direction, multiplying each other, at least in this country. It's funny because I would have said—

I'm not sure that I would have said that Arrow's conclusion was we spend too little. I think it was that insurance markets would fail, but he was also— and one of the things actually that people have missed about Arrow is that almost all of the focus— if people haven't read this, it's amazing. Among other things, it's amazing because it's Kenneth Arrow, Tyler Cowen writes this paper about healthcare, which you don't think of. And this is the guy of fixed point theorems and Arrow-Debreu model, and he has this deeply insightful paper with not a single equation, not a hint of explicit math in it, but it's a great, great paper. But HIPAA's— so when we talk about it now, largely involves the failures of insurance markets. We talk about adverse selection where If insurance companies can't sort, then only sick people will buy insurance. And we talk about moral hazard, which is because insurers pay for things, it's third-party payment, you're going to practice flat-of-the-curve medicines, will do too much. And that's actually an arrow as well. But a lot of what he talks about is the problem of you're depending critically upon your doctor. And the doctor is necessarily giving you advice on something that you is you didn't go spend all those years in med school. And you don't— you can't do comparison shopping. And you can do second opinions, but that's really not a very effective way of choosing it. So a lot of it is all about how we have to have institutions that allow you to trust people. Literally trust your life to people who know more than you do and might have self-interested reasons for not giving you the best advice or giving you the right treatment. And that's kind of all dropped out of this. And, you know, what— if I were— one argument about the US healthcare system is that actually we just started to treat medicine as a commodity. And it's because it's so not suitable to be treated as a commodity We've strayed from the ethos that healthcare providers are supposed to be better and nobler people. You know, we have TV shows about heroic doctors. We don't have TV shows about heroic middle managers, and there's a reason, because we need our doctors to be a bit heroic, but we kind of lost that in the monetization of the whole thing.

But I see parts of healthcare markets which work well. So Concierge Medicine, which of course most people cannot afford, But it does seem to overcome asymmetric information. You pay a large sum of money, a small club of doctors takes care of you, they seem to do a good job.

My understanding is that it's actually not at all clear that concierge medicine is good treatment. It's very nice hotel— hospital— you can see Freudian slip— they're very nice hospital rooms, but it's not at all clear that concierge medicine is actually giving you better treatment.

Well, but medicine itself is problematic, right? The marginal return on medicine is not quite as high as we might like to think. So there's a lot of randomized control trials, there's a positive effect for medicine, but you have to spend a lot, and even then a lot of what you spend doesn't get you anywhere. Well, but also you look at heart surgery in Indian hospitals where they repeat the same procedure thousands of times in a row. It seems to be cheap, it seems to have a pretty high success rate. That's done in the market. Again, it's not describing the whole sector. But there seem to be pockets which indicate the future of entrepreneurial medicine is maybe brighter than you're painting it to be.

Well, it's a— all of these things tend— well, we can go on. I think most of these apps— the Indian hospitals is a better one than the usual ones. The usual ones are things like laser eye surgery, which are— atypical of medicine because they're optional, uh, timing is not a problem, there is in fact time to shop, and it's a simple procedure that, that you can, you can ask, you can do compare— you actually can do comparison shopping.

And even though I won't let them do to me, by the way, I don't blame the market, it's just— no, I don't— it's highly risky.

Uh, I've kind of decided that glasses make me look more professorial, so the hell with it.

So, yeah, but I think there's a chance that 30 years from now, US healthcare markets will have mostly beaten the cost inflation problem. Some out of the last 5 years, right, cost inflation has been much lower and it will look a lot better compared to its competitors, even though right now in many ways we have the worst of all possible worlds.

That's right. We even have higher public spending than does France. Yeah. They, even though, yeah, we have the most privatized system there is and yet we actually spend more public money. It's actually, do you do like, so I actually teach about comparative welfare states. That's my very limited teaching now is focused on that. And do you do any work on that?

Have you ever looked? Never done research on it, but I follow it from a distance.

Yeah, well, I'm a consumer of the research, not a researcher. But one of the things I find is really striking is that you might think that some countries are good at running government social safety net programs and some are bad. And so you would look and say, well, you know what, if you look at healthcare, the US clearly does an abysmal job of the whole thing and we have incredibly high costs and pretty poor results.— and you could start to have all kinds of political economy reasons, interest groups, why the U.S. does that so badly. Except that we do retirement pretty well. Social Security is a smooth, clean system with minimal administrative costs. It's not very generous, but it does provide a basic floor for everyone. France has miraculously good healthcare. Costs half of what ours does. The results are better. Wait times are minimal. We don't know quite how they do it, but you run— their retirement system is a total mess. And it's all very specific to the programs.

And at the risk of incurring Bob Solow's wrath, I do think it comes back to culture. Oh, that's my point.

I don't think it does. I think the point is that if the French were culturally good at running welfare state programs, they'd have good healthcare as well and good retirement. And if we were bad at it, we would have bad healthcare and bad retirement. And in fact, see, I'm big on the path to A lot depends on whether you get a reasonable program in place at the window of opportunity when you have a chance to shape it. And if you screw up at that stage, then it's very, very hard to fix it because then you have stakes. Then you have— if the United States had had single-payer healthcare when Harry Truman tried to put it in in 1947—

Or even Nixon, right?

Or even— well, Nixon was trying for something that looked a lot more like Obamacare actually. A little bit to the left of Obama here, but that tells you something about it. But the— and we define France, Mitterrand made some really bad moves in the early '80s and every subsequent French government has been crippled by that, whereas FDR set up a good system in 1935. But that to me does suggest maybe it's culture because the U.S.

is a culture of consumption, low savings rates for some time now. And single-payer healthcare, whether one likes it or not, it's not really geared toward the consumer mentality in the sense of wanting all the hospital rooms to be well furnished and people— American healthcare spending is well explained by our general level of consumption expenditures, just like we spend more on snowmobiles. It actually all fits onto the same line. No.

Oh well, I would— No, I agree. Every American, but every American Actually, I have my Medicare card now. So every American— although, yeah. So I'm on a single-payer system, although not really. But anyway, we do have a single-payer system. And it's just our Medicare is a lot more lavish than the Canadian system, which is also called Medicare.

But it works OK, but it's fiscally unsound. So that's in tune with this culture of consumption.

Anyway, yeah. Well, I guess I would— actually, what?

Like the average person pulls out of Medicare $300,000 to $400,000 more than they put in. Right. Now, we could sustain that if we were fiscally virtuous on all other things, but we're not. Yeah.

Well, actually, that's a good— I'm just bouncing from things, you know, as things come up. Why do you think— I mean, if you actually look at determinants of national savings rates, as best I can make out, the best single explanatory factor is whether you speak English. Do you have any thoughts about what's— why are savings rates so different among advanced countries?

Explaining almost anything across countries is very hard, as we've been indicating. If you look at the time series of the U.S., my unvalidated hypothesis is the U.S. savings rate really drops in the early '80s, which is when you're entering the CEA. Right, my fault. And partly it's because of healthcare costs, that what people expected to save is somewhat eaten up by healthcare cost inflation. And for us, savings is the last thing we do because we're culturally oriented towards consumption. And since then we've had a pretty low savings rate. And you know, the last year savings was back up to like 7%, which surprised me, but that's coinciding with healthcare cost inflation being somewhat tame again.

Except that most people don't actually— well, that's not true. I was going to say that people don't see the healthcare costs, but in fact, no, that's— the amount that people pay out of pocket has actually been rising quite a lot. Sure. Because health insurance, especially employer-based insurance, is having much larger employee out-of-pocket expenses, deductibles, so that's cutting the wrong way, right?

The actual cost of healthcare is going up. But coverage has gone up as you've noted.

Coverage has gone up. —but I don't think the people who are covered, who are basically low-income working-age adults, are—it can't be them doing a lot of that increase in savings. Though actually we don't know that, do we? But I find that a little implausible.

Savings rates are oddly hard to explain, I find, either across countries or in the time series. What is that? And you would think it's a pretty simple thing. Well, the price effects are often not that strong, right? Right. And then you're left with a lot of mush. And then too many variables are changing. Right.

What do we explain well in economics? I think— No, I mean, it's a good question. I mean, this is something, because actually one of the things we want in teaching is we want— you can't, certainly can't have a textbook be a series of mysteries. Yes. It has to be—

Only Agatha Christie can do that. Right.

No, it's a fantasy.

Even then, someone is always inventing something, right?

I saw my tweet about the stock market last night. I was able to say, why is this happening? I have no idea. Why is this happening? I have no idea. And then ended up, but neither does anybody else. But that's not how to write a textbook. So we have to have success stories. Economics really does explain this.

And what are your prime examples? I mean, the first example I give is the best argument for economics is talking to someone who doesn't know any. However, whatever bad shape you may think things are in, when you hear people who just either don't know stuff or they're lying, you think a lot more of economics. But that said, changes to relative prices do mostly seem to have the predicted impacts. I know there's a debate on the minimum wage, but for the most part, if you put a sales tax on something, people end up buying less of it. So basically demand. Demand, yeah, downward sloping demand.

Downward sloping demand, yeah. We've found it's actually not that easy to get good examples of that even, 'cause you don't get that many natural experiments, but once in a while you do get something like, well, I think we really like congestion charges 'cause there you get an abrupt— But they work, right?

They work, exactly. Singapore's done it and they work. People in Singapore think they have traffic jams, but they don't. The other big truth, I think, is that significant negative shocks to aggregate demand are disastrous. Yeah. Even when they're offset, which they should be, but it's very hard to offset them completely, And that's remarkably well validated, as you've pointed out yourself, better validated than much of micro. And just those two facts, law of demand and negative shocks to aggregate demand, and then some game theory, which hardly ever predicts well but gives you categories for organizing your thoughts. That to me is saying a lot for economics, going back to the people who don't know any.

Yeah, well, I mean, yeah, I mean, so it sounds to me like you agree with I mean, I've been saying that events following the financial crisis have actually been great in terms of validating economics. Nobody's— well, some people saw the crisis coming, but every one of them saw 5 other crises coming that didn't happen. But the actual way that things worked, the recession, the fact that big expansion of the that the monetary base was not inflationary, that— there were a whole series of things that people who knew something like ISLM predicted that people who didn't think that way considered wildly implausible turned out pretty much the way the textbooks said they should have, and that it was actually a glorious period for economics.

You can teach most of it in Economics 101. Yeah. I like AD-AS better than ISLM, but that's a secondary point. Yeah. Negative shocks to aggregate demand and what follows from that. Yeah, I would have said shadow banking system, but I think you also have a high level of uncertainty, not policy uncertainty the way a lot of Republicans argued, but actual uncertainty like what's the shadow banking system, how's it gonna work from here on in, can we fix it, and I still think we have some of that and in some ways actually should have some more than we do. And that led to a big risk premium where people simply had their core assumptions so upended that it made the recovery quite slow.

Okay, actually did— I don't know if you— did you follow their— you probably did. There was this kind of debate between me and Dean Baker and on the other hand this guy named— what's his name? Bernanke, Ben Bernanke. Yeah, yeah, sure. Over how important the credit crisis was crunch was in the Great Recession. Do you have a— where I'm actually joined, somewhat to my own surprise, but I've come to the view that it didn't matter that much, that even if Lehman had been rescued, that the shape of what happened wouldn't have been very different.

I think it's overdetermined, so maybe I agree with both you and Ben. I didn't quite agree with the way you put it, but if something's overdetermined, You have 7 hammers banging on the nail.

Well, yeah, my question there is, and maybe you have a view on this, I think I may have heard it from you already here, but let's try it again. The financial disruption, the credit crunch, was pretty brief. It really, credit markets only really went kablooey in the fall of 2018, and they were pretty much back to normal by most measures 2008, and they were pretty much back to normal by the summer of 2009.

If you look at flows in the numbers, I agree, but I think if you had a deeper measure of real credit, letters of credit, how much is actually there in terms of implicit guarantees, I think it was worse than the numbers made it look, but I agree we don't know that for sure.

Okay, 'cause the normal financial stress indices, there are about 5 of those out there, They all do this, and— but then we had with all measured spreads, most of these are spreads that we're looking at, all spreads pretty much back to normal, but we still had a very slow protracted recovery. And you think that was— what's your view on what that was about?

Residential investment being very slow, rates of household formation being sluggish.— and those are investments of a sort. People just being more cautious, taking fewer commitments, taking less risk. For smaller businesses, I think it was more difficult in a way not reflected in aggregate credit flows. And you add all those things up, your smorgasbord, but back then—

If people don't know this, by the way, yeah, I've been saying that the next recession will have no one big cause. It'll be a smorgasbord recession. It'll be a bunch of things all coming together.

The slow recovery was a bit of that too.

Okay, see, I would have said that it was just, you had a housing bust, took $7 trillion of home equity off people's balance sheets, and it took 4 points of GDP off residential investment, and that would have happened even if we'd managed to avoid the whole crunch in the shadow banking system. I agree that's the single biggest factor.

Yeah.

Yeah. Were you paying any attention? I was just curious.

To housing before, you know, I looked at it, I thought, oh, maybe there's a bubble, but this will be like, you know, the first Bush, you know, the bubble will pop, mild recession, we'll get over it. I wasn't really— I didn't think it would matter. Okay. Was my view until it became obvious. So I think I was slow to see that it would matter. Even when I should have known.

I thought it would matter. I thought it was going to be nasty, but I had no idea how nasty it would turn out to be. And that actually is a— that weighs against my claim that it was all the housing bubble. Because if I saw the housing bubble, shouldn't I have concluded that we would have the kind of horrific slump we did? But I think there's a— actually, I think there's a problem. Throw this out. I guess this is a little bit— I think we all suffer from a failure of— I don't know if of imagination, of intellectual nerve, that even if our models say that something really drastic and say that where we'll be a year from now is going to be very, very different from where we are now, we don't really want to believe it. Believe it. It's really hard to stick your neck out, and the safe thing is always to say, "Well, yeah, there may be something, but—" Let me ask you a question about shadow banking, looking forward.

Right now, everything seems under control, but you could imagine— indeed, I think you would expect shadow banking to grow relative to deposit-insured banking because deposit-insured banking is only going to grow so much. Don't we eventually reach a point again where shadow banking is too large So, it's— one doesn't want the government to guarantee at all. Then there's too much moral hazard. But what do we do? Do we capitalize banks at 40%? And can we get to that without inducing a big recession?

But the thing is, and I'll say that you— we can impose capital requirements on the banks, but imposing capital requirements on things that are not conventional banks—

Of course, you push the problem somewhere else.

Well, no, but it's also— it's very— Who is covered by that? Yeah. Yeah, actually, so I have a view. I actually want to get— I mean, what would you do? Because I'll tell you what my problem is, but I want to hear your view on this first.

I know I have no good answer. I know something about off-balance sheet risk. I know something about the ability to substitute out of the regulated sector. In principle, I favor slowly easing up capital requirements on banks, but I don't actually have a great deal of confidence in that prescription. Yeah.

Okay, my— I thought Dodd-Frank was a very intelligent way to deal with it, which was— but it relied heavily on intelligent use of discretion, that it created this— financial committee, basically Treasury and Fed and a few others, that would identify systemically important financial institutions, SIFIs, and impose extra requirements. And the question is, what's the definition of a SIFI? And it really wasn't defined. It was the pornography. You'll know it when you see it. And that actually, I thought, that was something that would work fine as long as you had well-intentioned, reasonable, smart people in charge of the relevant agencies. And then, although this is not a political talk, obviously we are where we are, and so that's not a very good assumption about how things are going to work under current conditions. And I haven't been able to come up with an alternative. The thing about regulation of conventional banks is it's mechanical. You don't need to have brilliant people running the FDIC for the FDIC to do its job. But the trouble is with Dodd-Frank is you kind of do need to have pretty good people running agencies for it to do its job, and that, as it turns out, is not at all something we can count on.

But say you had the 2007 team, who were pretty good people, with the counsel of Dodd-Frank in 2007. I don't think things would have been much different, if at all. I'm not opposed to having such a counsel. I'm just not sure it gets you very far.

Well, it's an interesting question. I think they would have— Probably, although there's a question, would they even have been aware? Would they have realized? I mean, I would never have thought of AIG as a source of massive systemic risk, which it turned out to be. And funny thing, it was just one office of AIG in London, which went crazy, but turned out to endanger the whole system. Resolution authority, I'm too deep in the weeds, but the ability to close up a complicated financial institution institution, to bail out the functioning parts of it without giving a lot of money away to undeserving people. That's a big thing. And that was very much lacking in 2009. I was even a little bit in on those discussions. And you would say, we should be seizing at least one or two of these things just as a— to encourage the others. And they were saying, we don't have legal authority.

How are you going to do that? Am I allowed to ask you a politics question? Not an endorsement at all. I know that's different. The Senate right now has not in every way distinguished itself. So why don't Democrats in the Senate states they're losing simply run candidates who are at the median voter in that state and half the time win? Now, those candidates would be to the right of where you would like them to be, but if it's really the case that things are so dire now, as you've argued, should we not just pour all of the energies into having moderate Democrats, like, take 6, 7 states that right now they're clearly losing in?

I thought they kind of do that.

Am I allowed to ask you this?

I actually would have just said, I would have thought, you know, Heitkamp is a pretty conservative, pretty conservative for a Democrat. Manchin is very conservative for a Democrat. He is, clearly. And Heitkamp is as well. If you actually look at her position So I think they actually do that. Now the problem is that what a candidate's actual positions are and how that candidate is perceived. I actually do believe that you could run the same people that are, I'm trying to think of an example, but you could run a very conservative, Democrat, and it wouldn't stop his or her opponent from running ads accusing them of being in league with MS-13. I mean, it's just—

But it works for Manchin, right? So, right, 8 more Manchins.

Yeah, but it's been around for a long time. And I think there are a bunch— anyway, it's an interesting— actually, and also you're presuming a level of control, you know, presuming the existence of a smoke-filled room that does this stuff. You know what Will Rogers said?

I miss the smoke-filled rooms these days.

You know what Will Rogers said? What? I'm not a member of any organized political party. I'm a Democrat. Oh yeah. So the— what do you think about Trump going off on the Fed?

I think he wants a scapegoat. I don't think he wants to actually take on the Fed's independence. The Fed being independent gives him that scapegoat. He's worried about losing the House.. And the actual picks to the Fed board, in my view, have been fine. I don't agree with all of them on every issue, but they're all serious and professional, and I'm pretty happy with the Fed.

They've actually been pretty hard money, which is kind of odd given they're harder money than I am.

But you— I would never say like, oh, Trump has wrecked the Fed. They're the same kind of picks you would have gotten under, say, a Mitt Romney or any number of other Republican presidents. In some ways, possibly better.

What's your view on trade conflict now that we know? So we have avoided the catastrophic breakdown of NAFTA. We changed the name. Maybe we changed the name. The documents still say NAFTA, but apparently—

People are gonna call it NAFTA, I hope.

Oh, first question. When you first heard USMCA, what did it make you think of?

Marine Corps, of course. I thought Marine Corps.

Yeah. Which just shows that we're boring guys, 'cause several of my colleagues said it made them think of the village people. But that's old stuff too now, right? Of course it's old stuff. No, I didn't say it makes us old, it just makes us stodgy. But anyway, yeah.

China trade war is permanent in my view. You think we're, that's a— Start of a new Cold War, more or less forever for the rest of our lives. And it will be bad in ways we cannot foresee or easily predict. And the rest I think will trickle away. And there are some ways in which the new NAFTA is worse than the old, but they're tiny. And 3 years from now no one will be worried about that. Yeah. What's your view on Brexit? You know, if it's simply a customs union, we can all live with that, but there's no majority in Parliament for the normal sane way of doing Brexit. What's going to happen and how bad will it be?

And if I know— all right, there I can say I have no— it's— I mean, Brexit is interesting. And by the way, that this is a— it's kind of a teachable moment here because Estimates of the cost of, people have actually asked me, so I have back of the envelope estimates of the cost of Brexit. Which I agree with, by the way. Yeah, it's 2% plus or minus 3, but anyway.

But that's assuming they do it the normal way where they conclude a new agreement with the EU.

Well, if they actually get a customs union, then it's less, I think, because, but anyway, the question is why are the cost estimates Is Brexit comparable to cost estimates for a global trade war, even though it's, even for Britain, it's a much smaller thing? And the answer is that Brexit raises the, actually raises the resource cost of doing trade. Tariffs don't increase the resource cost. They deter trade, but they don't, but you get the actual cost of the tariff back in revenue, whereas Brexit is, now if they get the customs union, this doesn't happen, but if they don't, if they have, even if it's with very low tariffs, if you have to go through the process of basically clearing customs for goods that are coming across from the continent to Britain, that's imposing a lot of additional costs that ordinary trade policy doesn't do.

Yeah, it's a lot of infrastructure they need to build up. It'll take them a long time.

If you've heard about the M26? Yeah. If people haven't, there's a motorway. Britain is, they're doing contingency planning to take the M26, which is a big highway that runs down to Kent, runs down to Dover, I guess, and to convert it temporarily into a parking lot for trucks. Not temporarily. Well, but in the event of a no-deal Brexit, Brexit, so there is no customs union, they think there will be huge queues of trucks waiting to clear customs. And so they're setting aside a big stretch of highway to, for basically 30 miles of trucks waiting in line to clear the border.

If we have an orderly Brexit, the main uncertainty in my view is how much will FDI into the UK go up to offset Brexit. I don't believe anyone's estimate on that one way or another. I think what the true costs will be depend mostly on that variable. And I don't think anyone knows.

And what happens to the City of London, which is the City of London not being the city, but the financial district? How much erosion of the British financial industry will take place? And that's a hard one because, again, that's all network externalities. Financial districts are nothing but a peculiar kind of information externality made flesh., and we have no idea how really how bad it'll be, how much stuff will switch to Frankfurt or Paris.

And these days, are you a China optimist, pessimist, somewhere in between? Boy, I don't know.

I mean, here's one of these things where I'm probably, I mean, have you been in, I don't know what you think.

I thought for years they're gonna have a crack-up through an excess capacity recession,, you know, get through it okay because they respond swiftly, but really have a year or two of, you know, 0 to 2% growth and it would look disastrous and then return. And that has not happened at all, so I've been totally wrong.

Yeah, actually I'm in the same place except I thought it was gonna last longer. I've been looking at this totally unsustainable investment share of GDP saying that has to end and it'll be nasty and not sure what they'll do then.— and it keeps not happening. Does that mean that we're wrong about the fundamentals? Do you think we were wrong about the fundamentals, or do you think it's just the guy who jumped off the Empire State Building and passes the 20th floor and they say, "How are you doing?" and he says, "Fine so far"?

I think I underestimated the rate at which they could convert human capital into useful output, and if you do that at a high enough rate, you can overcome all sorts of sins. The sins we saw were real, but their ability to mobilize human capital and educate people and get them into good jobs has been phenomenal. You know, you can credit private sector, public sector, testing system, whatever, but they've done it. And you see it in the Chinese students who come to this country too. Except that— so just, I underestimated human capital and how good a job they were doing at it.

But isn't this a little bit of a confusion of demand side and supply side? I think they're a supply-side economy.

All right. For now.

We actually were about at the end. Were we supposed to take questions? Maybe not. But, okay, question. Couldn't you have said the same thing about Japan in 1989? 'Cause Japan, actually Japan is still very impressive in a lot of ways, but they were incredibly impressive in the '80s and—

Their demographics hit earlier and they didn't generate quality companies at the rate they had been doing so, whereas the rate of China generating quality companies really accelerated. And I think that's a big difference.

Actually, I was a little shocked looking at the most capitalized companies globally and they are, I think like 8 of the top 10 are basically network externality companies, but one of them is Alibaba.

Yeah, and that's not just a rent-seeking company, it's real stuff.

No, I know. It's funny because you never think of them at all, but it's— and actually, I believe there are no other non-US companies in that group, so it's all— maybe Tencent is pretty close, but that's— it's weird. It's kind of a— That sounds right. It's South African origins, but basically a Chinese company now. Yeah. And so sort of the top 10 companies in the world are 8 US and 2 Chinese, and that's actually— that's pretty wild. It is. More to come. It's, um, yeah.

Why don't we just thank each other and thank all of you for listening? Yes, thank you. My pleasure.