From Iceland’s math classrooms to Berkeley’s macroeconomics, Professor Jón Steinsson traces how price rigidity and the Phillips curve’s slope shape debates between Keynesian and neoclassical views. This conversation spans his long-term coauthorship with Professor Emi Nakamura, the microdata behind “Five Facts About Prices,” and why economic history is essential for understanding rare, economy-defining events.
Elliot: Could you introduce yourself and your work a little bit?
Jón Steinsson: Sure. My name is Jón Steinsson, and I’m a professor here in the department. I’ve been at Berkeley for seven years now. Before that, I taught at Columbia for 10 years. I did my PhD at Harvard and was an undergrad at Princeton. I specialize in macroeconomics, in particular in monetary economics and other kinds of related things, business cycles, and so on. And I teach 101B, which is the quantitative version of intermediate macroeconomics in the undergraduate program. And then at the graduate level, I teach first-year PhD students.
Richard: We want to ask about your upbringing in Iceland. What really drew you to economics and specifically macroeconomics?
Jón Steinsson: That’s a good question. Well, in Iceland, you know, I was always relatively good at math. So I was on a math track in high school, and most of the people that were in my high school class, which was the kind of most mathematical track of what you might call the top high school in Iceland. Although if there were Icelandic people here, they would maybe dispute that. Most of the kids in that class ended up in engineering or the sciences, and that was what I was gonna do. Economics is not taught at the high school level in Iceland, and so I wasn’t exposed to it, and I didn’t really know what it was, and so there was no sense in which I could have aspired to be an economist because I didn’t really know what that was when I was in high school. So I ended up going to Princeton. Now, most people in Iceland go to the University of Iceland, so that’s what most of my friends did, but I went to Princeton, and at the beginning, I was an electrical engineer. But I had read this book by this Icelandic economist-professor. I think it was around that time. And I really liked this book. It was a book of newspaper articles, like a collection of op-eds that this person had written. And I found this to be very interesting. And then, you know, in America, in a liberal arts kind of university setting, you’re allowed to choose courses from wherever. And so I was reading through the course catalog at Princeton and decided, hey, this subject seems interesting. I’ll take a class in economics. And so I ended up taking a class in economics in my first semester. Then I kind of got exposed to it and really liked it, and pretty quickly decided to transfer into economics. So that’s how it started for me. I mean, I guess I had an upbringing where my parents talked a lot about politics and history. And so, I had some background in kinds of things that you might call social science, although not economics. And so I had an interest in those things. You know, many people come to economics either from physics, and I had a lot of physics background, or from history. And you know, there’s a sense in which I had some exposure to history because my parents were interested in it, and my parents talked about politics a lot, and so on. So I had kind of exposure to the things that usually draw you to economics, and then when I found economics, yeah, it seemed like the perfect fit for me.
Elliot: What’s it like doing research with your wife, Professor Nakamura? Do you guys specialize in different parts of the process, and how do you manage that relationship?
Jón Steinsson: Yeah, so as you are alluding to, most of my papers, or almost all of my papers, are co-authored with Emi, who’s my wife. We met while we were undergraduates at Princeton, and we went to graduate school together. And we’ve worked at the same place and worked together since. Now originally, when we were in grad school, we weren’t working together. We were not even exactly in the same field. Emi was more of an IO economist. I was more of a kind of macro and international economist, and then over time, we started working together, and our first projects were kind of on the intersection of IO and macro. Now, I guess, why did we come to work together? I think there are kind of maybe two reasons. One is that when you are in a relationship with somebody, you have to get along with them. And so we had obviously worked very hard to get along with each other. And there’s a sense in which you can leverage that. If you have a co-author relationship, you also have to get along with your co-author, right? And so, we had figured out how to get along with each other, and why not leverage that in this other sphere? So that’s one thing. Also, when you have a co-author, it’s also extremely important that your co-author cares. So there’s this free rider problem in a co-author relationship. So whenever you start a co-author relationship, you might put in some effort, and you’re worried that the person you’re working with is not going to put in as much effort. They’re going to free ride on your effort. So that’s like an obvious problem that arises in any kind of collaboration. Now, Emmy and I, of course, have this deep, kind of bigger relationship. We’re married to each other, we have children, and so this notion that there’s going to be some free rider thing is much less obvious because you know, we just have such a deep relationship that I think that is less of an issue. Now, of course, anybody who works for a long time with somebody as a co-author or in a family, there are periods of time when one person is working more, and the other person is working less, but because you know that you’re going to have this long-term relationship, it averages out over time. And so that’s a major benefit of the situation we have, where we have this long-term co-author relationship and a more general relationship. So that’s a big benefit of working together, and probably one of the reasons why we find it very efficient to work together. The other thing is just like, which I think, you know, some people don’t appreciate fully, but like, working in teams is just really, really important. And so if you can find somebody you work well with, it’s just like really, really good. I remember when I was in college, and I read this interview of Bono, the lead singer of U2. And by that time, it’s around 2000, U2 had been a band for 20 years, right? And Bono had at that point become like a public figure, and he was going to solve all the world’s problems and so on. And there’s this issue: does Bono continue to be the lead singer of U2, or has he become bigger? Does he stop that and start doing something else? And I remember thinking about this, and I’m like, this U2 thing, obviously, this is an incredibly successful relationship he has with these other three musicians. And it’s very special, very few people in the whole world ever get that good a relationship going. And I’m thinking to myself, does he give that up? Obviously, he didn’t. U2 basically still exists, it’s still made music for 15 years afterwards, and so on. And, I think there are many stories, especially in music—obviously, the Beatles broke up, and that was a tragedy for productivity. But you read about many bands, the lead guy in Queen went on his own, and that didn’t work out either. Like, there are many people who I think seem not to appreciate how important the team aspect of what they’re doing is. And I always think about that interview with Bono, I’m thinking about that interview, and thinking about John Lennon and Paul McCartney, yes, they were way better when they were together than when they were apart. And so I’ve always thought that this was incredibly important that I had found this relationship, this co-author relationship, and I better not screw it up. This is really very valuable. And so yeah, I try not to screw it up.
Richard: Thank you for that. And on your co-author relationship with Professor Nakamura, we have a lot to ask about your research, but being respectful of your time, we want to ask about your paper, “Five Facts About Prices,” which seems to be your most cited paper. And we were curious as to why you think it was so influential.
Jón Steinsson: Yeah, so that was like our big, our first kind of hit paper. And so, the topic of that paper is about how often prices change and how rigid prices are. And price rigidity is the core assumption in a Keynesian model. And so when you’re thinking about macroeconomics, whether you want to think about the world as being basically neoclassical or basically Keynesian, how often prices change or how rigid prices are, it’s really the key distinction between those two classes of models. So there’s a sense in which this thing that we’re studying in that paper is really core to whether you want to think about the economy as the markets just work and the government doesn’t have to do anything—that’s the neoclassical view—or markets sometimes don’t work very well, and monitoring fiscal policy plays an important role in stabilizing the economy—that’s the other view. Now, a few years before we wrote that paper, there was pretty modest evidence on price rigidity. It’s the key assumption that distinguishes these two models, but there is modest evidence on it from innovative papers, and pretty small amounts of data. And then these guys, Bils and Klenow, they realized in a sense that there’s way more data on this that exists, and nobody has actually used. So they wrote a paper that was published in 2004, if I remember correctly. It’s the first paper that uses the underlying data that exists at the Bureau of Labor Statistics on prices. Because the Bureau of Labor Statistics publishes inflation data, so they gather a lot of price data. So there exists all this price data at the Bureau of Labor Statistics, and you can use it to figure out how often prices change, and nobody had done it. Bils and Klenow are the first people to do it. But in that original paper, they didn’t get access to the underlying microdata. They just got a table of summary statistics, and their paper is based on that table. And I remember, we were at a seminar together, Emi and I, at the macro seminar at Harvard, where we were grad students, and we were listening to—I can’t remember if it was Bils or Klenow—one of them was giving this paper, and we thought it was super interesting. But at the same time, it’s one of these papers that moves the frontier forward a lot. But at the same time, there’s so much more you could do. Like this is just scratching the surface. And so, we look at each other across the room, and we’re like, yeah, we better work on this. And so we got access to the underlying microdata. Our original idea had to do with actually working on the producer currency prices, because they had worked on the consumer prices. Our original idea had to do with seeing whether price rigidity at the producer level was similar to that at the consumer level. Now that analysis is in the paper, but the paper eventually became about something else. It became about the notion that a lot of the price changes that you do see are temporary sales. And so the price falls and then comes back to its original level. And two weeks later, it’s the same as before, but you had these two price changes. So does that mean prices are very rigid or very flexible? Because they do change a lot—it’s just that they’re always changing back to the original value. And so thinking about how much flexibility those price changes actually create was what our paper was about to some extent.
Elliot: On the topic of Keynesianism, you wrote a paper in 2022 about small slopes for the Phillips curve. Could you talk a little bit about that?
Jón Steinsson: Yeah, so there’s a sense in which that paper is the kind of last paper in the line of work that started with the “Five Facts” paper. The “Five Facts” paper uses data back to 1987. But the most interesting period is actually before that, in the 70’s, when there was high inflation in the U.S. And the data for that didn’t exist, at least in digital form, at the time. But we came to realize that it existed on microfilm at the BLS. So we did this big project. And this is before LLMs; this is before OCR software was good. We’re in the Stone Ages of doing work that today would be much simpler. But we have these millions of documents, over a million pages, that need to be converted from microfilm into machine-readable form. It was a big, big project that took years of work to do. And we create this data set that goes back to the 70’s. And then we wrote one paper about it, called the “Elusive Costs of Inflation.” That was the first paper that we wrote using that extended data series. But then, one of the things that we knew would be of interest using this new data was to create state price indices. The BLS doesn’t publish state price indices. And there’s a lot of demand in academia for state price indices. And we thought, well, we’ve digitized this data, why don’t we create state price indices? So, that was one of the components of what we did in the paper you mentioned, but then what it came to be about is using those state price indices to estimate the slope of the Phillips curve. The slope of the Phillips curve is one of the most important pieces of any business cycle model. It’s very contentious. Again, this relates to this kind of war between people who believe in neoclassical views and people who believe in more Keynesian views. If the slope of the Phillips curve is flat, then the world is very Keynesian. If the slope is very steep, then the world is more neoclassical. And earlier work had tried to estimate the slope using aggregate data. And it’s very difficult to estimate the slope of the Phillips curve using aggregate data, basically because you have all these confounds that, in particular, inflation expectations are moving around—it’s hard to control for inflation expectations—there are supply shocks—it’s hard to control for the supply shocks. And we had the idea, why don’t we try to estimate the slope of the Phillips curve using this state-level data, so in the cross-section of states? And that’s what the paper was about. One of the major benefits of doing it at the state level is that inflation expectations are, at least in the long run, national. So then you can use a time fixed effect to control for long-run inflation expectations. That was a major benefit. Another major benefit is that you have 50 times more data because you have 50 states instead of one country. And then a third benefit is that in order to deal with the supply shocks, working in the cross-sectional data basically provides you a way to come up with an instrumental variable that allows you to estimate the Phillips curve better. So those are the three benefits. The cost is that the slope of the Phillips curve in the cross-section might be different than the slope of the Phillips curve in the aggregate. And so we had to talk a lot about that in the paper. How is the slope that we’re estimating in the cross-section similar or potentially different from the slope in the aggregate?
Richard: We have one last question for you, time permitting. It’s about your post on X in support of recent Nobel prizes in economics being awarded to recipients in the field of economic history. And we want to ask you about your opinion on why it’s so important. Why do you think it’s so vital to have the economics departments preserve the study of economic history?
Jón Steinsson: I think that by and large, the economics profession has made a lot of good choices over the last 50 years in terms of the methodology that we have adopted. So there are several things. The field became much more mathematical. I think that actually is a strength of the field. It’s much more precise because of the mathematics, for example. We also had what’s called the revolution of identification, where our ability to do causal inference is much better than it was 50 years ago. So those two, for example, are major methodological advances that are a source of huge strength. There are other parts of social sciences, for example, that have not adopted mathematics as much or have not adopted the revolution of identification as much, and I think those fields, to some extent, suffer from that. So those methodological choices on our part have been an enormous strength. However, one of the things that has become less important in terms of our methodology over the last 50 years is historical analysis. And I think that allowing economic history to atrophy and die in certain economics departments has been a mistake. And that’s a place where we made a methodological choice that was not good. And why do I say that? Well, the main reason I say that is that when you look at really, really big questions in economics, and when really big things happen, things that happen rarely—big financial crises, hyperinflations, or even the trade war that we’re having now—when these kinds of things happen, and you try to figure out what you should think of these things, you can do it in theory, but when you think about it, typically, the most convincing evidence is historical because these things don’t happen that often. And the big events that are most convincing may have happened far in the past. So, a field that doesn’t make use of evidence further back than a few decades is really throwing away potentially an enormous amount of good evidence. So I think we as a field are at risk of that by letting economic history atrophy. Now, the Berkeley economics department is one of the few economics departments that retains a strength in economic history, and I hope that that continues. And those posts on X were trying to encourage other departments to also see it this way, that they should reevaluate the notion that they let economic history die, because when the big questions come along, often it’s answers from economic history that turn out to be the most convincing. You know, people should think about that.
