Jesús Fernández-Villaverde on the Quandary of Global Demographic Decline

Jesús Fernández-Villaverde is a professor of economics at the University of Pennsylvania. Jesús returns to the show to discuss his rise on X, how to frame global demographic decline, the three accelerants of demographic decline, the role of housing in family size, how AI will play a role in global demographics, what we know about AGI, the question of dollar dominance, and much more. 

Subscribe to David's new Substack: Macroeconomic Policy Nexus

Read the full episode transcript:

This episode was recorded on February 20th, 2026

Note: While transcripts are lightly edited, they are not rigorously proofed for accuracy. If you notice an error, please reach out to [email protected]. 

David Beckworth: Hi Macro Musings listeners, this is your host David Beckworth with some very exciting news. We recently announced that Macro Musings has full-length video of each episode going forward. The full-length videos are posted on our YouTube channel @MacroMusingsDavidBeckworth. There is a link to that channel in the show notes, and it would mean the world to me if you subscribed and shared it with others. Not only will that channel have full length videos, but it will also have some fun behind the scenes content. Additionally, the full-length video from each episode will also be posted on our X account @Macro_Musings. So, make sure you are following that account as well. We are so happy to give you even more macro bang for your buck. Now on to the show. 

Welcome to Macro Musings, where each week we pull back the curtain and take a closer look at the most important macroeconomic issues of the past, present, and future. I am your host, David Beckworth, a senior research fellow with the Mercatus Center at George Mason University. I’m glad you decided to join us. 

Our guest today is Jesús Fernández-Villaverde. Jesús is a professor of economics at the University of Pennsylvania and is well known for being an expert on quantitative macroeconomics, economic history, and AI. He is also a leading voice on demographic decline, and joins us today to discuss this topic as well as AI, dollar dominance, and the future of the macroeconomics discipline. Jesús, welcome back to the podcast.

Jesús Fernández-Villaverde: Thank you for having me again.

Beckworth: It’s great to have you on. As I mentioned in that introduction, you do a lot. You have this wide breadth of knowledge. You’re an expert in many areas. I mentioned this, I believe, the first time you came on—this is your third time, I believe—I mentioned this: You are the one person who gives me doubts about the notion of comparative advantage. Comparative advantage, the logic is sound. Mathematically, it’s clear. You do everything from economic history—in fact, you have a book on global economic history, you taught it—quantitative macro, DSGE models, quantum computing, you’ve worked on that. You’ve worked on markets and the history of thought. You’ve done a lot. How do you respond to that? Do you defy comparative advantage?

Fernández-Villaverde: I think that my wife will have a different view because she knows that I’m unable to switch a bulb when it goes off. If you let me live alone for a week, I will probably starve to death. If you think about the range of abilities of a human in a wider way, I’m thinking I would be awful at most of them.

Beckworth: Fair enough. Let me put it this way: What we see of you in the public—and now you are very visible on X, or formerly Twitter, close to 50,000 followers—you definitely are engaged on many different issues. You’re engaged on demographics, on the discipline of economics, on macroeconomics, on modeling AI. It just seems like you have something thoughtful and informed to say on all these topics. That’s why I’m delighted to get you on. Now, you’ve been very busy on X, right?

Fernández-Villaverde: Yes. I didn’t use X until around one year ago. I didn’t like that much the old X, where you have only 140 or 280 characters. It’s a format that is difficult for me. Then, around one year ago, I actually published a book in Spanish on public finances in Spain. My editor told me, “Unless you have an X account and you promote the book a little bit, you are not going to sell many copies.” I thought, “Conditional that I wrote the book, I might as well try to get a little bit of an impact,” and so I started tweeting in Spanish.

I realized that if you pay like $80 or something like that a year—I don’t remember exact amount—you can write as much as you want, and that’s a format I felt much more comfortable with. It was a little bit like going back to the times of blogging in the early 2010s, and I blogged quite a bit at that time in Spanish only. Then I started writing a few more blogs in English, a few more posts in English, and very quickly I started to get some traction. Then, one day I posted my slides from a talk I gave at the CEPR on demographics and therefore in fertility, and I woke up the next day and I have over 1 million visits. I was like, “This is crazy.”

Beckworth: Fantastic.

Fernández-Villaverde: Absolutely crazy. Anyway, I have been keeping doing that. What I find is very nice is I’m using it a lot as a scratch for ideas. For instance, I posted this week three posts on sequential risk about saving and retirement. Even if it sounds completely random, it’s because I’m writing a book on the future of the welfare state, and I wanted to make some points about redistribution across generations, et cetera. I figured out, “Guess what? Let me write something on X. Let’s see how people react to it, what people find interesting, what people don’t, what are the confusions.” I’m thinking about X as a way to think aloud, get early feedback, and recycle everything.

Beckworth: See that right there? You are writing another book on the future of the welfare state. You are, again, becoming an expert in a new area. In fact, you told me, I believe on a previous show that if you really want to know a topic, you write a paper or a book about it, and then you become the expert on it, and that’s what you’re doing.

Fernández-Villaverde: Exactly. It actually links with the issue on fertility, if you want to cycle into that back later, is how are we going to finance the welfare state in a world with very low fertility? I started looking a lot into the big two items in government spending right now, which is, Social Security benefits and health expenditure for elderly people. I started thinking very seriously about Social Security, and exactly how Social Security interacts, for instance, with issues like the fiscal policy, but also with immigration, which at this moment is probably the most important topic of discussion in a lot of countries in Western Europe. I figured it out, “I may as well write a book about it.” That’s the way. It’s the best way to think about it.

Often, when students come to my office and they say, “Oh, I want to write a paper on “x”. I’m going to read these 50 papers,” I say, “No, don’t read anything. Reading is overrated. Try to do it yourself. Once you have written your model, then you go back and read what other people have done, so you can check that you are not repeating yourself. Yes, this means that sometimes you are going to rediscover old arguments, but that’s okay because in that way, you actually interiorize very, very deeply what you are doing more than learning and reproducing.” I think that the idea of learning by doing yourself is extremely important.

Demographics

Beckworth: Yes. Let’s jump into demographics because that is one of the big reasons I wanted to talk to you. You’ve done a lot of conversations about this, but hopefully we can run with this and apply it to some of our growth theories, as you mentioned, implications for the welfare state, but also things like r-star, and just trend growth in general. You mentioned a minute ago, too, about this presentation you gave in London at the CEPR, the slides that blew up on X, a lot of people watched them or looked at them. Very fascinating discussion you have.

Maybe we can begin just by giving us the lay of the land, like what are the big data points? What are the stylized facts about demographics as this world now faces them?

Fernández-Villaverde: The most, I think, striking observation is that humanity is already below replacement rate. Replacement rate is the number of children that a woman has on average during her lifetime that we need to keep population constant. That, at the planet level, is around 2.2. At this moment, the planet is at 2.17, 2.15. Look, I’m being very careful. I’m not saying the Western countries. I’m not saying the rich economies. I’m saying the whole planet.

The reason I always start with that observation is because when you are talking about fertility decline, most people are thinking about the US, most people are thinking about Japan, most people are thinking about South Korea. What most people are not thinking is about Jamaica. Most people probably have never looked at the demographics in Jamaica. Jamaica is not a big island, it’s a small population, but to me is very interesting because it’s a sign of how widespread the declining population is worldwide. At this moment, fertility in Jamaica is probably around 1.4. To put it in perspective, in the US, it’s 1.6. The fertility rate in Jamaica is well below the US.

Now, let me give you a second, I think, very striking fact, which is the following: Go out there in your university and randomly stop the first 10 people that you see. I say in a university because you are going to randomly stop professors, graduate students, et cetera. Ask them, “I’m going to give you two population groups. The first population group is non-Hispanic whites in the United States,” because people tend to think about non-Hispanic whites as the people who are having the lower fertility, which, by the way, is not true anymore, but stay with me, “and Mexicans in Mexico. Which of the two groups do you think has a higher fertility rate?”

I bet you that none of the 10 persons is going to tell you that non-Hispanic whites in the US. That just shows you how incredibly fast fertility has been falling all across the world—in Mexico, in Iran, in Turkey, in Barbados, in Vietnam, in Indonesia. You name it. You pick randomly, you pick a map of the world and you randomly throw a dart, it’s going to fall in a place where fertility has been falling very, very fast.

Now, the third point that I made in my London presentation is everyone comes back to me saying, “Oh, but sub-Saharan Africa.” Then, what I tell about Africa are two things. Thing number one is the data is probably very bad. Second, most likely—and again, this is due to the fact we are not absolutely sure because the data is so bad—but most likely fertility is falling incredibly fast in Africa as well. The problem is that it started very high. The comparison I like to put is the following. Imagine that you drop a football from a 20th floor high-rise and now the football is in the 15th floor, and you say, “Oh, but the football is still very high.” I’d say, “No, no, man, this football is falling and it’s falling in an accelerating way.”

My view of the world is that in 2040, 2045, we are going to see articles in the newspaper talking about how incredibly low fertility is in Africa. It’s the future, so I may be wrong, but everything that I say as of 2026 is pointing out that we are moving in that direction. I think that people are really getting a lot of these things wrong because they keep thinking that this issue of fertility is about South Korea, it’s about people in New England. No, it’s about the whole planet, and it’s going to change a lot of things in ways that we have not really fully appreciated.

Beckworth: Let me draw from your presentation. I’m looking at a slide where you compare several countries. This is for 2024, but you look at the fertility rates, and the US is at the top, 1.6. Then you go down the list. I’m going to list some countries for a particular reason: Tunisia, 1.56. Turkey, 1.48. Iran, 1.44. Now, what do those countries all have in common? They’re Muslim countries. You often think that certain religious groups, Mormons, maybe some of the religious Jews in Israel, and the Muslims, they’re going to have higher fertility rates, but what your evidence is showing, that’s not the case. Even these cultural groups, religious groups, they’re also seeing these declines.

Jesús, help me make sense of this. Why is this so widespread? Why is it so pervasive? Then, why now? Is this just part of a regular trend? Because you did show the long-term trajectory, and it looks around 2017 or that period, there’s a deceleration even more so. What’s going on that even these religious groups are being affected by it?

Fernández-Villaverde: Let’s take this and divide it in two parts. Part one is, how do we go from a fertility of five children per woman to a fertility of, let’s say, around two? There’s the standard model in economics, the quality-quantity tradeoff of Barro-Becker. As an economy grows, as we modernize, as getting more and more education is a good idea, families decide endogenously to have much fewer children. If you interview me in 2010, 2013, that’s the story I would have told you. The story fits the data very well until 2013.

You are absolutely right, that something odd happens in some moment between 2013 to 2017, which is this standard story of modernization and people having fewer children because women are getting more educated, and because you want to educate your children more, and because the move from rural areas into urban areas accelerates. Suddenly, we are in a situation where you have countries like Chile with fertility rates of 1.0. You have countries like China with fertility rates of 0.89, which is absolutely unprecedented. It has never happened.

Beckworth: Wow, wow, wow.

Fernández-Villaverde: I think there are three possible theories out there. Unfortunately, until a little bit more time passes by, we are not going to be able to distinguish them. The problem over there is that it’s very, very important to distinguish between completed fertility and the total fertility rate. Let me not get too technical, but the total fertility rate is the number of children that a woman will have if she behaves as other women at this moment in the population. Basically, we look at American women that are 20, American women that are 21, American women that are 22, et cetera, and we build a synthetic woman.

Now, completed fertility is the fertility that a 20-year-old woman today will have at the end of her fertile years. The problem of completed fertility is that, by definition, you need to wait 25 years. Right now, we are in 2026, we usually make the cutoff at 45. Yes, a few women have children after 45, but it’s such a small number that it doesn’t matter. We are now looking at the completed fertility of women that were born in 1981. What some people are saying is maybe that you are just having a timing delay, that people are having children later. Now, I don’t think this is a lot of the story. I think it’s just a very, very small part of the story, but until we have completed fertility, we are not going to be able to be sure. That’s just an important warning. 

Second important warning, because every time I give a talk, every time I post on X, someone says, “Oh, I went to the World Bank data, and I don’t have the same numbers that you do.” The answer is, yes, because those numbers are wrong. All the numbers I cite—and in fact, the World Bank has asked me to give them a seminar and try to explain why the database they use is not correct—all the numbers I use are from the national statistical agencies of every country. I spent an incredible amount of time going to the national statistics of Turkey, getting the national statistics from Turkey, talking with demographers in Turkey, and being sure that these numbers are the correct ones. People just think—this is a little bit of a naive view of the world—that because some international agency has compiled some data, and it’s in some Excel file, that data is correct. That’s not true.

What are the main three hypotheses on the table? I can’t tell you how much weight I give to each of them. The first one is housing prices. Housing prices have skyrocketed all across the planet. This probably has a lot to do with something that is close and dear to your heart, which is a very low r-star. Think about a house price as just the present discounted value of the flow of housing services you get from it. When r-star, for whatever reason is very low, it means that net present discounted value is very high. Unless you can produce a lot of new houses and lower the price of housing services, what you are going to have is just a very simple valuation effect.

I have a paper. I just sent it back to the journal. If the referee is listening to us, we did everything you asked us to do. We actually argue that this valuation effect is a first-order consideration. All across the planet, housing prices are more expensive than ever. That has a big impact on people getting married or entering into long-term relationships, and that has a very large impact on fertility. Now, why is this important? Because remember, I told you that we went from five to two with the standard Barro-Becker. Now we are thinking about two to one, or 1.5 to one. In some sense, it’s a lot, but in some sense, it’s also less. I don’t need to explain that much. I only need to explain that you went from having two children to having one child. It’s not such a dramatic change in life.

I think about my own brother, Carlos. The housing prices in Madrid are very expensive. It’s very difficult for him to get a house with one extra bedroom, so he decides to stop at one kid instead of having a second one. I think there is a lot of evidence of that. For instance, within the US, states that have cheaper and more affordable housing are having higher fertility than states that are having less affordable housing. Again, we are not talking here about dramatic differences, but differences that are enough to account for quite a bit of the difference.

The second hypothesis is something that has to do with cultural norms. Everyone’s favorite argument is cellphones. You basically have every teenager in the planet looking at their cellphone, and instead of going on a date, they prefer to play with their cellphone. It’s hard to have a lot of children if you don’t go out on dates. How much do I believe that story? Personally, not that much, but I think I believe a little bit a related story, which is a story about change in social norms. The change in social norms that I think matters actually helps to explain Mexico or Chile.

Imagine that you are in a society where gender norms are a little bit unbalanced. For instance, a society where you have that a lot of men still prefer a very unbalanced power within the marriage, but the women have already been socialized into much more equalitarian gender roles because they have had access to social media. You have a husband who is expecting to get home at 7:00 p.m. coming from the office and that the wife has the food on the table, and you sit down and you have your dinner, and then you go and watch a football game on TV and your wife cleans up, but the wife has watched whatever TV show where this is not happening.

I think this is particularly important in Latin America and in the Muslim world, where you have a lot of men who are still thinking that my wife is going to work for me, and where a lot of the younger women say, “No, I don’t want to do that.” While in the Western world where there is a little bit more of gender equality—and all this is relative, of course, I want to be very careful—this is happening less. That’s why we may end up in the very funny situation where in the middle run, the fertility in Latin America and in the Muslim world is going to be considerably lower than the fertility in the rich countries.

Beckworth: Amazing.

Fernández-Villaverde: Yes. The third possibility is that the education weapons race is tighter than ever. By education weapons race, what I mean is, think of an economy that has a lot of positional jobs. What do I mean by a positional job? There are only—I don’t know how many partners there is at Goldman Sachs, but let’s suppose there are only 200 partners at Goldman Sachs. If 200 people go to get PhDs in finance or MBAs from top business schools, it’s not that difficult to become a partner. If you are in a world where there are 10,000 or 20,000 people with PhDs in finance or MBAs from top places, it’s very difficult to become a partner at Goldman Sachs. That’s why I call positional jobs, they depend on where you are in the world ranking.

The evidence is that the reward for these positional jobs is now higher than ever. That becoming an endowed chair at MIT in economics is way better in terms of lifetime income, in terms of social prestige, in terms of social influence, than being an endowed chair at MIT in 1960. Parents are involved in this weapons race, where they want their kid to get into the top high school because that will ensure that the kid goes to the top school, which means that the kid will go to the top graduate school, which will ensure that the kid will go to McKinsey, which will ensure that yada, yada, yada, yada.

This is particularly important in East Asia, which is a very hierarchical society. In the US, let’s say you are a super smart kid, you don’t make it into Harvard, it’s okay, you go, let’s say, to Northwestern, and your life chances are not going to be that different. In fact, Alan Krueger has a very nice paper where he argues the life differences are not that important. In Korea, if you don’t make it into what they call the sky universities, like the top three universities, you are never really going to get to the very top of society.

The payoff of being 100% sure that your kid gets into one of these universities is so humongous that you don’t want to have more than one kid or two kids, because otherwise you don’t have the time and the effort to get them through that. I think it’s a combination of these three things. Just to recap, since this was a little bit of a long answer: very, very high housing prices, unbalanced gender norms, and an education weapons race.

Beckworth: How much weight do you put on the housing argument? Is that the most important one in your view or not?

Fernández-Villaverde: I think it may depend a little bit from country to country. In the case of the United States, I think it’s the most important.

Beckworth: Okay. It’s interesting to hear that. To the extent it is important around the world, again, to varying degrees, situations are unique and different, but housing is this thing that plagues humanity. I recently had on Mike Bird, he has a book called The Land Trap, and this whole long history of real estate and housing and who owns it, who doesn’t, who has access to it. One of the key tensions with land as wealth is it’s also a service; you need housing as a consumption good, you got to live in a house, but people also want to use it as an asset to save for retirement.

Which of those two ends of the scale win out? Right now, maybe it’s leaning more toward a vehicle for saving, but there’s this inevitable tension that arises. Policy people debate this. It’s just so striking that it could also be an important factor weighing in on demographics. Ed Leamer had this famous article where he said, “Housing is the business cycle.” Maybe housing is the fertility cycle. Maybe it’s other things as well. It’s a huge important factor.

Fernández-Villaverde: Exactly. Look, the fertility rate in the US right now is 1.6, but the replacement rate for the US is around 2.1. 1.6 is too low. If we could increase it to 1.8, 1.9, this is still below replacement rate. This is not asking for the population to continue growing. This is just saying that instead of going down a cliff, we are just going to have a gentle decline. I think that if we move the dial in terms of housing policy, it’s going to be relatively easy to go from 1.6 to 1.8, 1.9, but you need to move the dial. You really need to be sure that young people have access to houses, and that you incentivize early partnership formation. This is very important.

Now, if you tell me, “Is this as important in China?” I don’t think so. I think that in China, the reason they are at 0.9, a big chunk of it is unequal gender norms. That is something that China needs to work much more. Yes, it’s a very country-specific answer.

Beckworth: That’s interesting. If you had to ask someone to sit down, what are the biggest economic problems faced in the United States, or even Europe, they’d say, “Oh, population, housing.” What you’re suggesting is, a lot of this can be solved by, first, addressing housing in these advanced economies because then on the margin, it will affect population and some of the other issues as well.

Fernández-Villaverde: Exactly. In fact, at the risk of moving away from the core topic, it is my view that a lot of the unhappiness in the political process right now, and a lot of people voting for parties—which I don’t want to use the word “populist” because it’s so loaded—but maybe political options that promise too much, given that what can be delivered, is precisely because of housing. That you are 26, you are 27, you are finding that the way to move up in life is difficult for you because of housing. You go on Sunday to vote and you don’t vote for someone who is very reasonable, you vote for someone who is proposing something radical and different. There is even a political economy aspect to it that is absolutely fundamental.

More in general, I think that a society that does not offer young people opportunities is a society that has problems in the very, very long run that go beyond the headline inflation. Don’t get me wrong, David. I’m the first one who knows that you need to get the inflation next quarter right. I’m just trying to say that in macro, we often confuse the urgent with the essential.

Beckworth: Absolutely. No, you’re absolutely right. Macroeconomics is both business cycle and long-run growth, and long-run growth is far more important. Wasn’t it Bob Lucas who said, “Once you start thinking about economic growth, it’s hard to think about anything else”?

Fernández-Villaverde: Oh, yes.

Beckworth: He was somebody who worked on business cycles and all those issues.

Fernández-Villaverde: Can I tell you a super nerdy story?

Beckworth: Sure.

Fernández-Villaverde: When I was an undergraduate, I read that. He wrote that “On the Mechanisms of Economic Growth,” which I think is 1986. I read that sentence, and I liked it so much that I actually went to a local Kinko’s, the equivalent of Kinko’s in Spain, I printed it out in very glossy paper, I framed it and I put it in my bedroom. The next time my mom came to visit and saw that, she was like, “What? You’re a teenager. You need to have a photograph of a model or a soccer player, and you have a quote of Bob Lucas? What is wrong with you?”

Beckworth: Wow. Hey, so your fate was written early on, and when you saw that article, your destiny was set for you that you would be an economist and deal with these very issues. Here you are, many decades later, wrestling exactly with the issues that Bob Lucas laid out many years ago.

Fernández-Villaverde: Yes. Exactly.

Beckworth: Let’s go to the consequences. Let me lay out how I think through them, and you tell me which is the most important way to do it. Maybe there’s another way. I often think about this, Jesús, through both the supply side and the demand side. On the supply side, I think of Paul Romer’s endogenous growth models: You get more people, you get more ideas. I always push back against the degrowthers, the people who think about, “Oh, we’re destroying the environment.” I always say, “No, no, no. More people, more ideas, more solutions to environmental problems.”

Then, there’s extensions. Chad Jones has a semi-endogenous growth. Yes, there are diminishing returns to more people, but the idea is, is humanity a brain or a stomach? View them as a brain. That’s, I think, one way to look at, if we have fewer people, fewer ideas, less innovation, lower growth, but there’s also the demand side. This goes back to Adam Smith and his idea, the division of labor is limited by the size of the markets. There’s been a lot of follow-up work.

This is a practical example of this. I live outside Nashville. Listeners know I work in DC. I live in Nashville. I commute back and forth. I live on the outskirts of Nashville. If I want something—some specialist, some special doctor, even a contractor—I have to go into the city. They don’t have as much in a smaller place. In a world with fewer people, you’re not going to have specialists and productivity gains and all that good stuff. I think there’s both demand sides, supply sides, and there’s other things as well. How do you frame this, and what do you think is the most important consequence going forward?

Fernández-Villaverde: I could talk about this for hours, so let me try to be very, very careful and select the most important. First of all, I fully agree with this idea that people generate ideas, and that ideas are good, even for the environment. Every time I give this plenary or keynote talk, very first question, someone raises the hand and says, “Oh, but this is going to be great for the environment,” and I say, “No. No.” First of all, because as we will discuss in a moment, when population collapses, the fiscal situation becomes extraordinarily bad.

The first thing that is going to go out in any budget when you are in a really, really difficult situation is everything related with environment. You are going to have a lot of 75-year-old people saying, “Oh, I want to be sure that my Social Security is paid. I want to be sure that Medicare is paid.” Guess what? “All those money to keep national parks, what do I care? I’m going to be dead anyway, so no.” If you care about the environment, you want to have a prosperous society.

Let me give you a very concrete example. Where does the environmental movement in the US start? In California in the 1960s because it’s the first time we have a generation that has pretty much everything. They have a house. They have two cars. They have a TV. Then they go to the Santa Barbara area, and they see an oil spill, and they say, “Why do I need a third car. I’d rather have a beautiful view.” Prosperity is the only foundation we can have for environmental sustainability. No one cares about climate change when you are poor.

Climate change, decarbonization, net zero, recycling, sustainability is an upper-middle-class desire. The best way we have to get this desire satisfied is to have as many upper-middle-class people as we can. Second point. What I appreciate a lot about your point regarding Smith is not just for things like medical doctors; it’s even for the life of a community. Last summer, I was in England—my wife and I love England, so we spent a lot of time at the Cotswolds. We were in this absolutely gorgeous village, and we were talking with the owner of the cottage that we have rented. She was mentioning that because of population decline, the local pub in the village just closed. This is a total disaster for the village. Because remember, in an English village, the pub is not the place where you go to drink, it’s really the social club. It’s where you go to meet your friends. It’s where you go to build community links, even if you are just getting a soda, you are not getting a beer. When the community is too small and you just don’t have enough clients to keep the pub going, you lose an enormous resource. 

Now, let me give you a second example that I just became aware of recently here in Philadelphia. Philadelphia, like many other cities in the East Coast, is a city that has had tremendous decreases in fertility over the last 15 years. One of the really interesting things, and again, many people don’t realize that, is that drops in fertility in the US over the last 10, 15 years have been concentrated on the very lower-income people, and a lot of the lower-income people live in inner cities, like in some areas of Philadelphia. That means that the Philadelphia school district has been forced to close a lot of primary schools. This is a drama, not only for the kids that now need to be bused far away from where they live, but because the gym of the local primary school is also the community center.

Suddenly, you cannot have a Saturday night event for the local community because the local primary school has closed. There is nothing artificial intelligence can do to solve that. You can have the best version of Claude that can write the most beautiful code you can imagine. You can have the best version of ChatGPT that can write the most eloquent emails replying to your boss, but you don’t have a pub to go on Saturday night to meet with friends. You don’t have a gym to organize a bake sale to raise revenue for the local park. This is the type of social institutions and social network that we tend to underestimate in economics fundamentally, but that at a very basic level are absolutely essential for the life of a healthy community.

Beckworth: That’s a great point. The importance of civic life, of having sports leagues for your kids, having pubs, having churches. There’s towns around here where all those things get smaller, fewer, and then in the future they get eliminated. Let me give a personal anecdote to this. As I mentioned, I live in Nashville, on the outskirts. You may have recalled a few weeks ago there was a big ice storm that came through here, so power was shut off, lots of trees fell. What made it really bad was the ice was so thick on many trees, lots of trees here in central and middle Tennessee. Power was out for a long time. I joined a crew from my local church. We took chainsaws. Economists with a chainsaw, be careful.

Fernández-Villaverde: Insurance problems.

Beckworth: Yes. We went out and helped clear trees from people’s yards. What’s interesting is—this hit me as you were talking about this—there were a number of older people who couldn’t do anything with the trees in their yards, so it was important that we went and helped them as a community. Just imagine a future where it’s not just populations declining but it’s more older people, the age pyramid has inverted, and we have another ice storm and there aren’t as many younger able-bodied men to go out and clear those trees. They’re stuck without power, without food, a lack of community, just even having contacts.

Some of these people have been in their homes for a week without power, and someone can come knock on their door, check on them. That just paints a picture of how dark and dire this could get if taken to the limit.

Fernández-Villaverde: Exactly. We see that in Japan. In Japan, there are all these absolutely heartbreaking stories from time to time in the newspaper about this poor fellow who died in his village and no one checked on him for two years. One day, a utility crew needs to go inside the house for whatever reason, and they see this body dead over there for two years. 

Again, I say that before, but I want to emphasize. When I say this is a problem, I’m not advocating that we come back to explosive population growth. I’m just saying, look, if you have maybe a little bit of a gentle decline in population, and that may have some good consequences for the environment, that’s fine with me. I don’t have any problem with that. What I don’t think most people understand is the type of social collapse that you are going to have when your fertility rate is 0.9.

Again, we tend to underestimate that in economics a lot. I love economics. I’m a big defender of mainstream economics, but I’m also aware we have some blind spots. One of the blind spots is exactly what you were saying. We do not appreciate that for a lot of people, once you have a house, a TV, and a car, having the local church group where you go and have coffee and donuts with a friend is way more important than having a second car.

Artificial Intelligence

Beckworth: Yes. Absolutely. For the sake of time, let’s move on to AI because you touched on this. Many people, including myself, have wondered, will AI at all provide some offset? Now, we just described reasons why clearly it will not. Even if you have a robot who moves as agilely as a human, it looks like a human, it’s not going to be the same thing as having contact, community, the local engagement. Let’s put that aside and acknowledge that it’ll never be that. What about other things like having AI, maybe robots taking care of certain tasks and jobs should we have a declining labor force?

Fernández-Villaverde: First of all, yes. I’m very happy. I’m a big believer in artificial intelligence. I write papers. I amaze myself at the type of things that artificial intelligence can do. I just look back at my own research, how my life, my workflow has absolutely changed. My wife and I make the joke that ChatGPT now runs our lives. We go to a restaurant and my wife had a little bit of high cholesterol, so the doctor asked her to be a little bit more careful. Now, we go to the restaurant, we make a photograph of the menu, and we ask ChatGPT to decide an order that is good with cholesterol control. Even ChatGPT now is deciding how we order in restaurants. Maybe robots have already taken over us.

On the other hand, we need to be careful with two considerations. Consideration number one is the following: Do I believe that the United States, Japan, Sweden are going to have an economy that is flexible enough to introduce robots and get a lot of the job done? Yes. Why do you think that Mexico or Spain are going to have the same flexibility? This is not about the supply of technologies. This is going to be about the political ability to implement those policies.

Let me tell you what is going to happen. Waymo. I love Waymo. Whenever I go to San Francisco, I love because it’s so clean and it’s so neat and you get in and you don’t have a driver and you don’t need to give chitchat. The other day, I took a Lyft. I went to Washington University in Saint Louis to give a talk. When I was going back to the airport, I get into this Lyft and, man, did the guy stink. It was so smelly. I couldn’t wait to get to the airport.

Are we going to have Waymo? The technology is there. Well, you are going to have many cities that are going to say, “No, this destroys jobs and this is capitalist,” et cetera. You’re going to have many countries that are going to impose barriers to the adoption of that technology. Even if the technology is there, it is not obvious that countries can take advantage of it.

Second, a lot of what this artificial intelligence and robots are going to be able to do is going to be great to increase welfare, but it’s not going to show up anywhere in national income and product accounts. Because it’s not going to show up anywhere in national income and product accounts, it’s going to be very difficult to tax it.

Again, coming back to the example before. Now that I need to be a little bit careful with my cholesterol, or my wife needs to be a little bit careful with her cholesterol, it turns out to be the case that you have an artificial intelligence app that is helping you to make food decisions every day. Where is that showing up in GDP? Nowhere. How are you going to tax that? How are you going to generate the resources from this extremely productive economy in terms of welfare to pay for Social Security benefits?

People tend to forget we need to tax. This is a transfer, and the transfer needs to be financed in some way, and it needs to be with real taxes. Economists always say, “Oh, I can’t devise a lump-sum transfer.” That works great when you are doing first-year graduate school and you want to prove Ricardian equivalence in the prelim, but that’s not how the real world works.

How are we going to do that? How I’m going to tax David, that on Saturday morning, you are going to take your kid to play baseball and you are going to check on Claude, whatever thing you need to do for your kid to do warm-up exercises, and they are going to be much better warm-up exercises than what the coach could give to you.

Those two considerations. One, even if the technology exists, can we implement it? Is the political process going to allow to implement it? Second, a lot of the gains from this technology are going to be nonmeasurable and, hence, very difficult to tax.

Beckworth: Yes, a lot to process there. Let me go in a little different direction here. Let’s put aside economics for a minute. Since you are such a widely read person—again, history, economics, philosophy—let’s get even a little bit deeper. We’re going beyond the normal purview of the show, but let’s do it. Let’s have some fun, Jesús. What about the idea of AGI? Do you think AI truly becomes sentient, conscious? What would that mean for us, I guess, in terms of, what are we? I’m a Christian. I believe I’m made in the image of God. What would this even mean? There’d be a big ontological shock, I guess, for some of us if this did happen.

Fernández-Villaverde: I tend to be on the side of those who think that that’s not going to happen. That we are going to probably reach relatively soon the concave part of artificial intelligence. That the immense transformation of the next 10, 20 years is not going to be that the next version of Claude is sentient or much better. It’s that it’s much easier to use, much cheaper, much faster, and we are going to completely restructure society around that, and that we are just seeing the beginning of that happening.

This is a little bit like the famous paper of Paul David about electricity and how it really takes a generation to adopt electricity and really reshape our societies. Now, having said that, and being very open about the fact that I don’t think we are going to have anything that looks like sentient artificial intelligence, the last 10 years have been surprising for everyone. We will need to think a little bit about that.

I teach this one-semester course on artificial intelligence in economics. The first day, I talk a little bit about some of these issues. Then, of course, the big discussion is about, what does it even mean to have a conscience? The problem is that even philosophers themselves cannot agree. There are very operational definitions, like the famous example of the Chinese room by John Searle.

John Searle was this very famous philosopher at Berkeley. He devised this thought experiment where you have a room with two small holes, one on the left and one on the right. One person on the left hole is putting a ball with a Chinese character. There is a fellow inside the room with a gigantic dictionary. For every character that you put on the left, there is an arrow that tells you which character you need to pick up and return. The book is so gigantic that it can always do that. Is the fellow inside the room actually understanding Chinese or is just checking a very large dictionary?

If you take a very operational definition of what knowing Chinese means or what understanding Chinese means, the fellow understands Chinese because it can keep a conversation. This is, at this moment, what large language models are. They just have a gigantic dictionary where for every input, they can produce an output. Now, there is a different interpretation that says that there is some threshold effect where you can actually start talking about conscience.

I don’t know. Again, as I was saying before, I’m a little bit skeptical that that’s going to happen. I don’t think we are going to need to have Vatican Council III where we decide if we need to baptize or not Claude in the next 25 years. In that sense, as you say, it may be my own personal biases than my deep understanding of this technology. On the other hand, I really believe that we are really already seeing a little bit of the concave part of the large language model. It’s just that what we are still seeing is the convex part of us understanding what you can do with large language models, which is a slightly different thing.

Beckworth: No, for sure. The whole notion of what is consciousness is a big question still in philosophy, so we shouldn’t pretend that we’ll understand it, the two of us, let alone AI as well. Let’s step back from AI being sentient to maybe a more modest but still important question. This is one I got from Noah Smith, and I think I sent you the article.

Again, nothing to do with consciousness or sentience, but he argues that these AIs will get so capable, they’ll do so many of our tasks, that we will lose some of our agency when it comes to creative thinking or creative tasks. Yes, maybe they make us better off economically. Yes, maybe they’re not conscious, but you know what? They’re going to be doing so much for us that we would normally do, work hard on, be creative with, we’re going to feel disempowered. And the big turning points in history may be shaped by AI as much as humans. What is your thinking on that?

Fernández-Villaverde: There is something to it. It’s very interesting. There is ancients—now we are going to jump all the way back to Greece and Egypt. Ancients were very concerned that the invention of writing will be the end of memory. That before writing existed, you had to memorize everything, and thus, think about the Odyssey by Homer. You have these bards that will go from village to village in classical Greece, reciting the Odyssey. Because there was no cheap writing—there was already writing, but it was so expensive to write down, and it was so costly to learn it, very, very few people had it.

Then when somehow this gets democratized, people say, “No, but we are not going to remember that anymore.” It is true. I don’t think there is a single person in the planet right now that remembers the Odyssey by heart. I can probably only quote the first five or six lines at most. That has a loss. You don’t really fully appreciate a poem until you are able to recite it to yourself, to say it again and again, to appreciate the cadence of the poem, the inner structure. That’s something that we have lost.

On the other hand, how are we going to run a modern financial system if I cannot record how much money you have in your checking account? To get them completely out of the mainstream of this podcast in previous episodes, not only the ancients worried about that, but a lot of the great sociologists in the late 19th century had similar feelings about modern technology.

For instance, Max Weber—which by the way, Max Weber always thought of himself as an economist, not as a sociologist. His chair, he had two professorships in his life. Both of them were in economics, not in sociology. Max Weber talks a lot of times about this enchantment of modernity. It was translated by Talcott Parsons into English, perhaps a little bit in a loose way, as the iron cage of modernity.

What Max Weber highlighted was that everything that machines and modern technology do for us is great, exactly as you say, because you can get into your car and go to your doctor appointment in downtown Nashville, as you were mentioning before, but we also lose the pleasure of going for a walk, and as Henry Thoreau will remind us, appreciate the beauty of Walden Pond.

Part of the challenge of our future education system will precisely be helping us cope with this iron cage of artificial intelligence. That’s actually why I have been trying to talk with the leadership at Penn and try to tell them, education needs to be 100% redesigned around artificial intelligence. It’s absolutely crazy not to embrace artificial intelligence, but it’s also absolutely fundamental to think what are going to be some of the most basic questions that we need to get out of an education.

For instance, issues of meaning. What I think that Noah is trying to talk about is a very fundamental issue of meaning. What is life about? What is learning about? What is research about? Why do we even want to write papers in economics? Have you thought about that for a second? Why do we want to write papers in economics? Well, at the end of the day, it’s because we want to understand society, but also because we want to build a better society.

If we go back to these very basic principles of why we are engaged into economics, the fact that AI is going to help us to write better papers will give a meaning back into what we are doing. Yes, I fully agree with Noah that we are going to get into an iron cage of artificial intelligence, and that the challenge for a society, will be to be able to find meaning within that iron cage and being able to handle the disenchantment of modernity and hyper-rationalization. That’s why going back to the classics and reading both Homer and Max Weber is more important than they ever were.

Currency Dominance

Beckworth: Fantastic. Great thoughts. We could go on, on this topic for a long time, as you mentioned. In the time we have left, let’s switch gears, go back to macro. I want to end on a paper that you just finished. The reason I want to go to this paper because it’s something near and dear to my heart. This paper is called “International Currency Dominance.”

It deals with an issue that a lot of us are thinking about right now, dollar dominance. How robust is it? Will it fade and go away like the British pound, which was also the international reserve currency up through World War I? Are we locked into the dollar despite all the challenging economic policies that we’ve seen in the US? Will we still have a dollar a decade, two decades from now? What say you, Jesús?

Fernández-Villaverde: That’s why I wrote the paper. I was actually having this conversation with a couple of friends and I said, “Well, let’s write a paper about it.” What do we get out of the paper? We get that the answer, at least through the lenses of the model that we build, is probably yes. There are two forces within the paper that says that the answer is probably yes.

The first one is that there is a lot of momentum in dominance. Yes, the US dollar eventually overtook the British pound, but it really took 60, 70 years. The United Kingdom became a smaller economy than the US at some moment in the late 19th century. The United Kingdom got bankrupt after two World Wars. The United Kingdom had a terrible macroeconomic policy in the 1950s and 1960s. Yet, the British pound still played a relatively important role until the mid-1960s. It takes a lot of effort to lose your position as dominance.

Basically, in the paper, the way we model this idea is that there is basically a coordination device. We have one of these search and matching models of money. It’s in the new monetarist tradition of Lagos-Wright. In this type of model, moving away from one equilibrium or one coordination is really very difficult.

The second thing that we try to argue in the paper is what is the alternative. When the British pound starts to lose some attractiveness, the situation is particularly dangerous because the US dollar is there. You have a good currency, issue with a country with good rule of law, a country that has political stability, a country that allows you to invest, et cetera. It’s very easy to start using dollars.

Who can be the rival of the dollar right now? The renminbi? Now, China. Do you really want to bet on China not expropriating you? Do you really want to put your eggs on the basket of the Communist Party of China? That’s a much harder proposition. There is no rule of law in China. There is really no commitment device on part of the Chinese government that they are not going to change the rules of the game. In addition to it, and that goes also to some of the papers that I have written lately about the economic future of China, I’m relatively pessimistic about China in the middle run. That has to do, in fact, a lot with fertility, just tying down with everything we said before. When China has a fertility of 0.89, the future of China doesn’t look that great in the middle run. 

I think that unless the US really pursues very misguided policies, the most likely scenario is that in 10, 20 years, the US dollar will still be the dominant currency, maybe a little bit less than it is now. The biggest danger then, according to the lenses of the model, will be that a rival appears. That suddenly some other major great power becomes a much more reliable alternative. Imagine there is a democratic revolution in China, and suddenly China becomes a gigantic Japan, let’s say, with maybe bad fertility, but a democratic government with rule of law, with open financial markets, then suddenly the Chinese currency will become a much, much stronger challenger to the US dollar.

Beckworth: Check out the paper, listeners, and great points, Jesús. Three comments. Number one, I like this point about the pound losing its status was not a discontinuous jump. It was a gradual process. It’s important to keep that context in mind. It takes time to undo the network effect, the path dependency of an established network. Also, I’m glad to hear this paper uses monetary search models. You mentioned Lagos-Wright. That’s another thing I recall from a previous conversation, Jesús. You do these search models, including the monetary search model. That’s fantastic that you did that. 

Fernández-Villaverde: The way I think about models is that you want to pick the right model for the question at hand. If I’m going to write a model about what is going to happen with, let’s say, changing the Taylor rule, I probably want to use a New-Keynesian model, but if I want to build a model to think about the dollar dominance, I think that a New-Keynesian model is silly, and I want to use a different model.

I think that is one of the hardest things, actually, to try to explain to graduate students that the models that we use need to be designed for the question that you have at hand, because reality is extremely complex. There are thousands of different moving parts on it, and what you want to do is write a model that highlights what you think are the mechanisms that are really at the very core of what you want to understand. That’s why economics is hard, and that’s why research is hard.

Beckworth: Yes. Listeners out there, take all the models, see which one applies, but don’t forget the monetary search models. I’m a big fan of those. Final point on your paper, and I’m very sympathetic to your findings, it’s going to be hard to break the dollar’s hold on the status of reserve currency. I think one thing going forward that’s going to add to that is dollar-based stablecoins. I’ve mentioned this before on the podcast. 

Early on, I was a little concerned about what was happening with the trade war undermining the dollar’s role. Look, this administration will go, its time on the scene will fade soon. What will last will be more likely the effects of the GENIUS Act and US dollar-based stablecoins, which should extend the footprint of the dollar. The trade war may not be something that lasts as long. I think that’s something to keep in mind, that some of these policies may not seem as important now, may be far more important down the road as the dollar network continues to grow.

Fernández-Villaverde: Yes, I couldn’t agree more. I teach economic history to the undergrads, and what I tell them is that a lot of the policies that really matter in the long run are the type of legislation that got enacted in a completely obscure way on December 15 before Congress went home for Christmas, but that set up the right institutional detail.

Why does Delaware still gather so much of the corporate business of the United States? Naive people will tell you because Delaware has good tax rates. When you actually look at the tax rates of Delaware, they are not that good in comparison with many other states. What Delaware has that is fantastic is that it has a very well-designed corporate law system and has judges that actually understand the system and can apply it in a reliable way.

I talk with a lot of my friends in the business. I went to business school as an undergrad, not economics. Some of them have moved into doing great things in life. What they constantly, constantly tell me is, “At the end of the day, I don’t care if the corporate tax rate is 30% or 35%. Of course, I prefer 30%. No one wants to pay an extra 5%, but that’s not what is pivotal in my decision of doing something or another. What is absolutely pivotal is that I can be sure that if I go to a judge because we have a dispute in our contract, the judge is going to decide in a predictable way.”

That’s why it’s so important to defend the rule of law because what businesses and market economies need is the rule of law that makes sure that we understand the rules of the game. That’s a point that Hayek makes a lot in The Constitution of Liberty in 1956 and that people tend to forget when they read Hayek. The predictability of the legal system is a first-order consideration.

Beckworth: Completely agree, and great book by Hayek. On that high note, our time is up. Our guest today has been Jesús Fernández-Villaverde. Jesús, thank you so much for joining us again on the podcast.

Fernández-Villaverde: Thank you for having me one more time.

Beckworth: Macro Musings is produced by the Mercatus Center at George Mason University. Dive deeper into our research at mercatus.org/monetarypolicy. You can subscribe to the show on Apple Podcasts, Spotify, or your favorite podcast app. If you like this podcast, please consider giving us a rating and leaving a review. This helps other thoughtful people like you find the show. Find me on Twitter @DavidBeckworth, and follow the show @Macro_Musings.