Economic Models and Monetary Policymaking Transcript
Jekyll Island Conference
November 5, 2010
Charles Davidson: Welcome to the podcast of the Jekyll Island conference sponsored by the Federal Reserve Bank of Atlanta and Rutgers University. We're in Jekyll Island, Georgia, at the Jekyll Island Club Hotel with conference panelists who gave presentations on model's and monetary policy analysis. Moderator Jeffrey Lacker, president of the Federal Reserve Bank of Richmond, speaks with presenter Lawrence Christiano, professor of finance at Northwestern University, and panel discussant Thomas Sargent, professor of economics at New York University.
Jeffrey Lacker: Larry, your presentation was about the journey over the last couple of decades and the conduct of policy and macroeconometric policy analysis. I know at the beginning of that period there were these large models and very small models. Those large models got thrown out—thrown under the dustbin of history—although there might be different views about whether that is warranted or not. But you've been building larger and larger models with more and more features, more and more frictions in them, and I understand those models have been actually put into use; models based on dynamic stochastic general equilibrium principles and all. I wonder if you'd say a word contrasting this generation of models with what was going on and what sort of got tossed aside in the early '80s.
Lawrence Christiano: Well, the initial set of models, I would say, were models that were not well founded on economic reasoning. It was a previous generation's approach to macroeconomics, and in the '70s a revolution occurred in macroeconomics and we joined the rest of economics. Tom Sargent was one of the leaders of that. And the previous models belonged to this old approach to macroeconomics, and the new ones were an attempt to place the models on sound economic reasoning.
Thomas Sargent: Can I take a cut at that? The old models had many virtues. They were designed to become more and more realistic and statistically to describe the past data, so the spirit of them is—they were models in the Keynesian tradition that put the evolution of time and uncertainty at the center of things, and following Keynes, they said that the way the people form expectations about the future—the people inside the model: the consumer, the investors, people building houses, buying houses—an important determinant of their behavior was what they thought the future would be; what they thought the future asset prices would be. And Keynes was one of the people that emphasized the importance of expectations. He didn't have a theory of expectations; he said they would move around. He used the term "animal spirits"—that something activated somebody to become optimistic or pessimistic—what that meant was that expectations were something you really wanted to know about.
So people built models of house prices, for example, in which people's expectations were really important, or stock prices. Then several people, one of them is John Muth, scratched their head and said, "Wait a minute. Our models themselves have the property that they're actually—when we're done with the model, we can forecast prices. And if our theory's right, we can forecast prices really well. By God, we can forecast them better than the people in the model." And then famous people bought policy and said, "Yeah, we can, and we should exploit that fact." The whole design of policy is to take into account that with the models, we know better what's happening in the future than the people.
So then answer the question that Muth asks, "If you're so smart, then why aren't you rich?" because generally you can make money by forecasting the future. So then the questions is, "Well, wait a minute. Why don't we build models in which the people in the model can forecast at least as well as the economists?" Which is probably true, if you know any economists. So doing that took a little while to figure out how to do it. But when you do it, then you build something called rational expectations models, and that's just a fancy word for the people in the model care more about forecasting the stuff in the model than we do, because it's affecting their lives. So we built models that have the discipline of that. So they've eliminated that cheap way of saying…it's an elite way of saying the government knows better than the public what's going to happen in the future.
Lacker: So that was a big difference between the old models and the new?
Sargent: Yeah, big difference.
Lacker: So taking these models to the data and, you know, actually use them in policy, is kind of a difficult and sort of subtle task. Looking back at this Great Recession, a lot of people act as if, "Gosh, it should have been obvious in 2006." But you go into it and it looks like a shock. If you look at history it looks like, this is pretty unusual. This is sort of like a meteorite hitting or something. So when you build these models, you use historical data. How do you deal with the fact that you could get a shock come along that you just never saw in the data you're using to estimate your model? How do you think about that? How should economists approach that?
Christiano: Well, I can tell you how I think it was approached in the current situation, which is that economists, in fact, invest in lots and lots of different models and study lots and lots of different phenomena. As a consequence, we ended up with people running the Fed, in particular [Fed Chairman Ben] Bernanke, who had studied lots of banking models, and for various reasons he was interested in the Great Depression. Partly, he was into that literature, and he studied models where the banking system actually can become dysfunctional, and he had grown used to thinking about how worlds like that work. And in effect he had been sort of in the gym with those models lifting weights. So when the crisis struck, he did have a framework for thinking about all of this stuff. So I think this is an example of how it makes sense, the old phrase "don't put all your eggs in one basket" is right. We want economists studying lots and lots of different phenomena because we really don't know what we're going to be hit by. What we can hope for is, we'd really love to be able to forecast, but I'm sure things are going to happen that we don't forecast. What we can hope for, though, is that when things happen, that we have some kind of a framework for thinking about what to do.
Lacker: Tom, I know you've thought and worked about this topic.
Sargent: Yeah, the way I'd answer your question is, not just to back up what Larry said, but economists and historians have studied data on past financial crises. There's a great summary of it in the first chapter of a book called Understanding Financial Crises, which was written by Douglas Gayle and Franklin Allen. And their first chapter describes a myriad of data in which, not just the United States, partly the United States but mainly other countries, have experienced financial crises, and what the consequences of them were. And the consequences are—if you read what happened to countries after financial crises and adverse employment shocks, you'll see that the experience of the United States in the last two and a half years in basically right in line with the average; pessimistic as that may sound. From that data there's kind of a lot of just raw empiricism that, when things go wrong with financial intermediation, the whole process of getting money from savers to people that can productively use it is screwed up because banks aren't doing what they're supposed to do. That, in a dynamic economy, where banks are supposed to be doing that, that has a really adverse and long-term influence on unemployment, job creation, and that's why many economists think that the first order of business is to get the financial remediation business going again.
Lacker: So, if this was a time capsule and we were going to open up at the next Jekyll Island conference, what would you tell us?
Christiano: I think the Federal Reserve is a remarkable institution, in terms of—this is going to sound corny—but in terms of patriotic and civic-minded people that will serve on the board and on the staffs. You know, there are people who are serving on the staffs that could make four or five times as much on Wall Street. And they're making much less because these are big issues, and the decisions you and the board make affect us all now and in the future, so out of altruism and confidence.
Sargent: I think part of the strength of the Fed comes in its decentralized feature, so as we move forward, since there are so many difficult issues and so many difficult decisions and so on, I think debate is important, and the way the Fed is structured now promotes that sort of thing because there's a kind of a competitive atmosphere because of the sort of independence of individual things. And I think that's going to make it more likely that we get the right answers.
And it's interesting. One of Milton Friedman's favorite laws was the Law of Unintended Consequence, so as we learned again today when the Fed's initial creation with these twelve banks—why would you ever do that when every other country just has one? It was initially set up for political reasons to extend the support for the Fed. But then what's happened is the regional Feds have become intellectual centers that create diversity of views and first-rate research and a competition of ideas. So into Washington, there's a constant flow of new ideas, and you just have to believe that that helps make policy good.
Lacker: A positive unintended consequence. Let's hope we can preserve that.
Christiano: Well, there's also this partnership going on because…one place I see it is at the Chicago Fed. Well, the Richmond Fed is like this too. The Chicago Fed has top-notch research economists. So what's happened is there's a little bit of blending, and what you get out of that is leverage. So the Fed has an interest in getting some thinking going on, and by having their connections with the universities, they're getting a lot more people to think about stuff than just their own people. I think that's been very productive. I think all the developments in modeling, right from the very start, right from the Sargent and Wall stuff was going on…who was the president? You talked to some president a lot?
Sargent: Yeah. Mark Willis.
Christiano: Yeah, he thought you were terrific. So Mark Willis used to go and talk to Tom all the time. And so that makes Sargent think about, you know… Willis comes in and, he could have been thinking about something else. You know, God knows what.
Sargent: Medieval coins.
Christiano: [laughter] Yeah, medieval coins. Instead, this guy is coming in and asking, "Well I have to give this speech, and I have to clarify my thoughts about this problem." So that gets Tom thinking about it, and that gets Wallace thinking about it, and I think that's what happens with the Fed, and I think a lot of good things have happened because of it.
Lacker: Yeah. Pollination, yeah.
Christiano: You know, I'm going to give an example, and I'm not totally sure everyone would agree if this is a good or bad, but the whole thing—the whole analysis of the zero bound—reflects this partnership, because Eriksson and Woodford were intimate buddies and colleagues, etc., with people inside the Fed, for example, Bernanke. It was not twenty minutes that that paper was done before Bernanke was sold on it and was chatting with everybody else in the Fed about it. So the best thinking was available inside the Fed. The best thinking outside the Fed was available instantaneously because of these close connections. And I think this is a phenomenon of modern times. I think if you go back to the '70s, I think the board had a few fancy econometricians, so there was an element of this at the board. But I don't think if you went to Richmond, if you went to Minneapolis, I don't think you would find any economists, you know.
Sargent: It definitely shifted over the horizon. I'm told the political and public attacks we got made people realize—and some from academics—made people feel like, alright we've got to step up our game here. I think there was some of that going on in some of the other banks. But it's definitely been quite a journey.
Christiano: I think another recent example is Chari and Kehoe, I think, wrote a very nice paper on too big to fail. And those guys… Chari is basically a thinker. He thinks. And he'll think about whatever it is you have sitting in front of him. And you happen to have him sitting inside the Minneapolis Fed, constantly being harangued by people who worried about too big to fail or whatever, and that's how you get this interesting, intellectual stuff happening.
Lacker: It's been great talking to you, Larry Christiano and Tom Sargent.
Davidson: This concludes our podcast of the Jekyll Island Conference sponsored by the Federal Reserve Bank of Atlanta and Rutgers University. Thanks for listening, and please return for more podcasts. If you have comments, please e-mail us at firstname.lastname@example.org.