Roger Farmer on the Natural Rate of Unemployment Hypothesis and Prosperity for All

Read the full episode transcript

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: Roger, welcome to the show.

Roger Farmer: Great to be here. Thanks.

Beckworth: Yeah. It's a real treat to have you on here and I read your book really enjoyed it and we're going to get into it. But before we do, I like to find out how did you become a macroeconomist?

Farmer: Ah, good question. So when I was in the equivalent in the UK, I grew up in London when I was in the equivalent high school here, my sister dated a student at LSE who was studying economics. I was about 16 at the time and we would chat late in the evenings over the kitchen table. He got me interested, I took economics A level and then I went to Manchester University to study at the time.

Beckworth: Okay, so we can thank your sister then?

Farmer: Yes.

Beckworth: And her sweetheart. Okay. Great story. And you went into macroeconomics, Why macroeconomics over say microeconomics? What pulled you into macro?

Farmer: Well, and in Manchester I had a separate small department and econometrics department. And then I followed when I was thinking around about what to do next, there were two macro professors there, Michael Parkin and David Lagler to Western Ontario. I was using them as references and David suggested to me that and offered me money to come out to Western Ontario. And when I got to Western Ontario, the econometrics group was not as strong as it had been in Manchester, The macro group was very strong. So I followed what was strongly at the time, which was macro.

Beckworth: Fascinating. Now we're going to get into your book, but one thing I wanted to find out from you and that is, as I read your book, I was struck by how you were able to think outside the box to use that phrase. Most grad students go grad school and they follow their professors, they're trying to get tenure so they don't deviate too far from the mainstream. But you were able to go beyond the standard New Keynesian paradigm, you're able to think outside the box, and then in your book you go in, tell the story how you were fascinated by the early work on Sunspots. But what made you open to thinking differently? For many people may be hard to do that, what would you recommend or what was your secret to thinking outside the box and opening this other research agenda?

Farmer: I wish I could tell you. I mean, I just always wanted to explore my own different things. Interestingly enough, the thesis I first wrote was on contract theory. And my first job, I spent two years at Western Ontario as a student, and then I hung around Toronto visiting because my wife had moved there as a visiting student. And then end of the third year, I got a job as a lecturer. Now, the thesis I ended up writing was completely different from anything I was thinking about at the time and it was actually inspired by a lecture that Joe Stiglitz gave.

He was on the path at the time which is Repeal the Laws of Supply and Demand, and I put that together with some other things I was doing. So the idea at the time then was still the same thing as the one I've been pursuing in my current research, which is that we should really be thinking about notions of multiple steady state equilibria to think about macroeconomics as opposed to sticky prices. And I've had that view ever since I was in... it was in Ontario. So it's simply the way that I'd formulated that has changed. Now, Why I did that other than following convention? I suppose I've just always been pigheaded.

Beckworth: A good trait to have sometimes, huh? Okay, well, let's move into your book, and as I mentioned before, it's called Prosperity for All: How to Prevent Financial Crises. And in it, you outlined some problems with mainstream macroeconomics, which has dominated by New Keynesian, thinking the New Keynesian model and let's work our way through some of the problems. There's also, you outlined problems with the macroeconomy itself, and then we get to your solutions and your way of approaching macroeconomics. So the first problem with modern macroeconomics that you outlined is the natural rate hypothesis. Tell us about that and then what are the problems in mainstream macro with it?

Natural Rate Hypothesis

Farmer: Well, interestingly it's all centered around a simple diagram that I put together when I was writing an undergraduate textbook about 15 years ago. And I was looking for a way to explain the natural rate hypothesis to undergraduates. It struck me that a good way to do that would be to take very long averages of unemployment and inflation data. So I took tenure monthly averages at the unemployment rate, I took tenure monthly averages of the inflation rate. And I plotted them on a graph and I was expecting to see a vertical line, the natural rate.

And the idea would be that, over 10 year periods which have 120 months, you would expect to be averaging out as many months when inflation was above the expected inflation is when it was below. And the expectations of augmented Phillips curve which was dominant than in macroeconomics, which combined rational expectations with the natural rate hypothesis would have implied that if both the natural rate hypothesis and rational expectations were correct, you would expect to see this vertical line. Now, what I found in data was much close to horizontal than vertical, and that led me to start being skeptical of the natural rate hypothesis.

Beckworth: And so the national rate hypothesis says there is this underlying rate of unemployment in the economy. Also, you could argue an underlying potential GDP value and that the economy naturally gravitates back towards it. You gave it the analogy, it was real interesting of the rocking horse, could you describe that to us?

Farmer: Yes. So in the book and actually in a previous book I wrote, which was, How the Economy Works. I used the rocking horse metaphor, which was originally from Pixel and then formed the basis for Ragnar Frisch's important paper on Propagation and Impulse Mechanisms in Economics. And the rocking horse analogy is an analogy that really forms the basis for all of modern macroeconomics, both classical and New Keynesian. Well, Frisch argued or Pixel argued is that the economy is like a rocking horse in which the interaction of billions of human beings in markets represents the dynamics of the rocker. And he asked you to think of a child with a club randomly hitting the rocking horse and the economic analogy there would be any kind of shock to the system.

And if you look back at Pigou who had a lovely description of business cycles that he wrote in 1926 I think. He has a lovely verbal description of the economy that fits into this framework very nicely. So Pigou has agricultural shots, he has strikes, he has monetary disturbances, shots to competence are back there in Pigou. And all of these shots in a classical system, like a kid striking a rocking horse, if the child will then to refrain from hitting it ever again, the rocking horse would move back and forth that converge always to the same rest point. And I think that's a good analogy for both classical and New Keynesian economics in which the rest point is the natural rate of unemployment.

I can trust that in the book to what I call the windy boat model, which is a different metaphor and I think a metaphor for the Keynesian general theory. And in that metaphor, the for the economy is like a sailboat on the ocean with a broken rudder, and the shocks are gusts of wind that blow the boat in one direction or another. And the difference between the rocking horse metaphor and the windy boat metaphor, the economy is like a sailboat with a broken rudder. There's no tendency after the shocks are taken away for the boat to end up in any particular point. And again, if you draw the analogy with the economy, the implication is that an unemployment rate of 5% would be a steady state equilibrium, or an unemployment rate of 25% would be a steady state equilibrium.

Beckworth: Okay, so the standard view is there's a natural rate, we gravitate back toward it, there's deviations. Maybe that the path back takes a while, and what you're arguing is that's incorrect, empirically that's incorrect. And then theoretically it's also incorrect, you present a model for that, we'll go over in a bit. But your argument is that there can be multiple values where the unemployment rate can settle down for a sustained period of time, that is equilibrium unless it's shocked out of it or pushed out of it by some government policy. And yet that's the dominant view in macroeconomics right now. It's interesting, you mentioned in the book that big macroeconomic events help shape our understanding and bring the death of important idea.

So you mentioned the great depression brought the death of Say's Law, and that's where Keynes comes in and writes his book, General Theory. You also mentioned the great stagflation 1970s, where we had high inflation and high unemployment brought the death of involuntary unemployment that came into the modeling. And you argue that the great recession may be the death of the natural rate hypothesis going forward, people will reconsider. So a question I have for you is, is the profession looking more closely at this multiple equilibria idea? Are they beginning to discount more the natural rate hypothesis and taking the perspective you see, is there a movement toward that?

Farmer: Oh, yes. I think so. Very definitely. I think that much of the work that Larry Summers has been pushing on second stagnation is really much the same idea. He has a slightly different mechanism in line.

Beckworth: Okay. Well, let's move on to another problem with macroeconomics and the particular problem with New Keynesian economics. You call it in your book and this book was a fun to read, I really recommend it to everyone out there to go and read it, but you call it bastard Keynesianism, I got a chuckle out of that. And then also with the generative research agenda, I want to come back to what that means. But you list a number of problems with the New Keynesian approach to macro.

I'm going to just list one through five, you provided in the book. One, you say, prices are implausibly sticky in the model, two; inflation is persistent, three; there's no unemployment, four; welfare costs of business cycles are small and five; the New Keynesian model cannot explain bubbles and crashes. And this is like the baseline model, and I think what you're saying with the generative research agenda phrases, people have attempted to address these issues by adding bells and whistles and new things to the models. But at the end of the day, it's just becoming a big complicated mess. Is that the point you're making?

Problems with New Keynesianism

Farmer: That's exactly right. And it's based on that, if there are PhD economists out there who have time in between solving functional equations to actually think about what they're doing, I would urge you to read Lakatos, which is a wonderful philosopher of science, he was in and out of sea for a long time. And he talked about scientific research programs, and economists I think most of us meet Milton Friedman's positive economics at some point. And sometimes we go beyond that and we looked at Kuhn and revolutions, but there's a lot of interesting ideas in the philosophy of science that I think are worth delving into a little deeper. And what Lakatos argues is that most of science, not just social sciences, but most of science competing research agendas and suppose the best way to understand it, is that whenever we test something particularly in non-experimental sciences like economics everything we do is a test of a joint hypothesis.

And when the data don't do what we expected them to do, we need to modify something in the way that we were thinking about the world, and what Lakatos does is argue that we should separate the central propositions of an approach into what he calls a hard core and a protective belt. And the hard core is a set of propositions that you will never give up on, you'll simply modify the protective belt that is consistent with the evidence. If you accept this idea, then how do you decide whether a particular view of the world is correct, or whether it's not. And what Lakatos suggests is that some research programs are what he calls progressive and some of degenerative. And progressive... let me start with degenerative, degenerative research program is one in which time after time the research program is falsified in the sense that predictions it made about the world turn out not to be correct.

And it modifies its protective belt, it keeps adding what you call bells and whistles in order to accommodate the central ideas. Progressive research agenda is one that occasionally makes correct predictions, and there's no way that a committed adherence of one program is going to be convinced by evidence if the other program is correct. So in Lakatos' is view I think is that, the way we should spend our time trying to convince the next generation of graduate students, because those are the ones who are gonna actually have an impact on the future give up trying to persuade the old guys who've got so much invested in the particular program that they probably will never change their minds. So I believe it was a physicist who said, "Science progresses one funeral at a time."

Beckworth: It's of a thought. Okay. So the problems that you listed, I want to focus on one; prices are implausibly sticky in these models and both the monetarists and New Keynesian would be subject to that critique because they both either invoke sticky wages, sticky prices. And you instead turn to something else, you turn to incomplete market ideas, which I found very interesting. In two different markets, you argue the problem with the macroeconomy is that there's incomplete labor markets and incomplete financial markets. And those two things are the key... I don't call them frictions, but those are the key issues that you see in the macroeconomy that need to be addressed, as opposed to all the other things we talked about for the New Keynesian. Is that correct?

Farmer: Yes. Absolutely.

Beckworth: Okay. So let's talk about the first one; the incomplete labor market, and you bring up the idea that finding employment itself is a technology. It's a search that goes on and so you lay out a Keynesian search model. So search itself as a technology, the firm searches, laborers search. And you argue that this technology is at work but there's a missing market here, labor markets are incomplete. And in an ideal world, there would be a matchmaking firm, but they don't exist because of moral hazard problem. And I read this, you actually had a note in the back of your book, which was very fascinating, "What would a complete market look like in this setting?" So could you describe what this matchmaking firm is and why it's such a problem that it's absent?

Farmer: Yes. So I liked the idea that came out of search theory and particularly strong idea that I think was important in search theory, it's the one that you just pointed to, which is the notion of thinking about a technology for matching people with jobs. And empirically, I think one can think about the Beveridge curve, which is a downward sloping relationship between vacancies and unemployment as an isoquant. So if you think of the production function and you ask yourself, suppose the economy needed to fill 10,000 jobs a month, for example. Well, you could fill 10,000 jobs a month perhaps by having a 100,000 vacancies and 50,000 people searching for those vacancies, and that will be one point on the isoquant. Or you could have a 100,000 people searching for 50,000 jobs and that would be another point on the isoquant.

Now, if that truly were a technology, then one could then imagine that there would be a bunch headhunting firms that would operate the technology and they would purchase the inputs, which would be the searching time of the unemployed workers and the searching time of the recruiting department of the firms that had vacant jobs. And those activities would have prices, and they would sell the product which would be a match back to the worker firm pear. And the moral hazard issue is, I think you don't see those markets because they're clearly open to moral hazard. So if you thought the way that they would operate, the matchmaking firms would be going to unemployed workers and saying, "I'd like to buy the exclusive right to find you a job, and I'm going to pay you for the right to find you a job." And I'm going to say, "Fine, wonderful. Pay me a $1000 a week or an hour or whatever the going price is."

And then the matchmaking firm would come back to the unemployed worker and say, "Look, I found you a job." Now, we see attempts to operate those markets usually by governments. And they're subject to the problem that... First of all, people often sign on the dole under more than one name and take unemployment benefits that they're not entitled to. So if a private market were to do that, you could imagine a person signing up with more than one firm and selling the right twice. And then once one is offered the job, there's the incentive to say, "Oh, sorry, my mother is sick back on the East Coast, I need to go, I can't take that job." So paying people to be unemployed, doesn't seem to be a particularly effective way to allocate resources.

Now, as a consequence of those markets not being there, the price signals that would tell individuals whether 10,000 jobs should be filled by a 100,000 vacancies and 50,000 searching people, or a 100,000 searching people and 50,000 vacancies, that's missing. And as a consequence, you end up potentially with the wrong point on Beveridge curve, which again, I want to think of as on the isoquant. So if you were to give this problem to a social planner, you'd find that there was indeed an optimal point on the Beveridge Curve. You'd find that there was an optimal unemployment rate, you might call that the natural rate of unemployment. You would not want to push out unemployment below that rate because in to do so, you'd be using too many resources in the economy searching to fill jobs quickly as opposed to actually producing them.

So that issue that there are missing markets, I think was well understood by the theorists who were working in search theory. And once you start writing down simple models where the problem exists, and you find you don't have enough equations. So as a consequence of the missing price signals, if you simply assumed that firms and workers are crise takers. So firms paid wages is given and prices is given and workers paid wages and prices is given, and trying to close in a way. What I call classical search theory, which is the Diamond, Mortenson and Pissarides recognize that problem, and they solve it by introducing another equation, which is the Nash bargaining equation. So instead of being priced and wage takers, once a work that needs a firm, everything's renegotiated, even though the firm might have advertised the job at $20 an hour, once you get there you say, "Oh, no, I want $25."

And there's a bargaining process that goes on, and that bargaining process is what selects a point on the Beveridge curve. Now, it turns out that if you write down search models and you close them in that way, they don't do a very good job understanding the volatility of employment/unemployment over the business cycle. That was pointed out by China. It's known as the China Puzzle. If on the other hand, you take the view that I do, which is that, once we see a missing market like that, we should think of that as an opportunity to model the idea. This is, I think, one of the most important ideas in Keynes general theory, that there are potentially multiple steady state equilibrium unemployment rates.

Beckworth: Yeah. It's very interesting. And I really liked the stress you put on the price signal is missing. I mean, any good free market loving capitalists would appreciate that insight. We want markets to do their job, but the price signals are missing. It's not going to work, and so you have an incomplete market. Let's move on to your second concern. So the first being, incomplete labor markets, your second concern is incomplete financial markets. And you make the case that the participation is limited to those folks who were alive or the critique is that unborn generations cannot participate in the financial markets. And so is this the ideal world where there's the contract for every contingent outcome but that's not possible?

Farmer: Well, yes and no. Modern financial economists build off Gérard Debreu. So let's back up a little bit here. So much of modern macro, much of modern finance is built on Debreu chapter seven. So what I mean by that is in Debreu's book Theory of Value, he has a proof of the existence of an equilibrium and he makes some very general assumptions about the nature of the things that are being treated. And then chapter seven is a throwaway, he says, "Well, the space that I constructed, what I defined the commodity, the way I did it was sufficiently general that you could think of the commodity as being delivered at a particular date, at a particular geographical location and in a particular state of nature. So a loaf of bread in Paris in June of 2003 if it's raining, is different from a loaf of bread in Paris in June of 2003 if it's sunny. And you could potentially trade these things in advance.

And I think in Debreu that was just a throwaway. But then when Lucas comes along, Lucas really embraces that as a paradigm for all of macroeconomics. And when people talk about complete financial markets, they implicitly mean that every human being who will ever be born is present in some matter market at the beginning of time, they trade commodities contingent on every possible event that could possibly occur. And then the world just unfolds, the uncertainty is realized, the trades are realized and that's it. Now, if you were to think about the people who are doing the trading, there're really two macroeconomic paradigms that people have used. One is that the people doing the trading are infinitely lived families where the matriarch trades on behalf of her children and her grandchildren and her descendants, and the other is the overlapping generations framework where we're all selfish, we don't care about our children and they have to fend for themselves.

Now, in my view, the world is closer to the second than the first, but that presents a real problem for the complete financial markets' view. Because even if you grant that we can trade on every possible contingency that we could differentiate between costlessly, we still faced the problem that not everybody who will be affected by financial volatility is present to trade in the financial markets. So I draw an analogy in the book with fire insurance. So fire insurance was a wonderful invention, but in order to benefit from fire insurance, you have to buy the insurance before your house burns down. If you could trade in this preexistence market, it's a little like, "Rules caused us the veil of ignorance in his work on social choice."

So if you could be trading over events you would be born into, and I want to say to you, David, you could be born into the great depression. In which case you, you may well have a lifetime of misery, or you could be born into some parallel 1929 state of a world that looked a lot like the 1960s, where there was lots of growth, then you'd had a job and a car and you'd have been wonderfully happy and you could trade in the financial markets. You could buy an asset that pays off a lot of money if the great depression occurs and in order to pay for it, however, if when I flip the coin, I get the 1860s event, then you're going to have to pay for that by giving up a little bit of the resources that you might otherwise have had.

And if your risk averse, which is what we typically assume you would say, "Roger, yes, I'll take that insurance." And it's the fact that you and other people like you would be willing to pay for that insurance, which would eliminate the volatility that you would potentially see in the asset market. So in my view, the reason that we see volatility and in not just in asset prices, but in price earnings ratios, is the inability of our children and our grandchildren and our great-grandchildren to ensure against financial events that could potentially cause great misery.

Beckworth: Okay. So we have the two incomplete markets; incomplete labor market, incomplete financial markets. And as you just mentioned in your book, you provide evidence that the cyclically adjusted price earnings ratio, and also you give evidence that stock market … would cause unemployment. That speaks to the volatility in financial markets, these are real issues and so what is to be done about them? And you present your model of macroeconomics that helps us approach these issues.

We'll get to the solution in just a minute, but what you do and we'll talk about this and you can add to what I have to say here. But number one, you keep rational expectations framework. Number two, you allow for multiple equilibrium, which we've already talked about. And then the third thing that really I think shapes this, is this belief function that you bring into your model. So you operationalize this idea that beliefs are fundamental input to production by having a belief function, I believe it takes the place of the Phillips curve. Is that right?

Farmer: Yeah. That's right.

Beckworth: Okay. So tell us more about this belief function. Why is it so important?

Are Beliefs an Important Input for Production?

Farmer: Well, in order to answer that, cast your mind back to prerational expectations economics. So when Milton Friedman did his work on the Phillips curve, he needed to think about how people formed expectations. And at that time we thought of expectations as a different variable. So the price of a good tomorrow, call that P, was different from our belief of what the price of a good tomorrow would be, call that PE. Now, once you've added a new variable, which is an expectation or a belief, you needed an equation to determine how that variable was formed. And what we did and what Milton Friedman did in the theory of the consumption function was add something called adaptive expectations. And under that theory, your belief about what the price would be, would adjust in response to possible mistakes that you made in the past.

Now, cast your mind ahead and now we have Lucas coming in, and Lucas says, "Well, prices in the models we're writing down, end up being random variables. And if there're random variables, we should expect that the people living in our models would be treating them in the same way that an econometrician would treat them." And he said, "Well, if we treat them in that way, we don't really need to distinguish expectations from realizations, they're simply random mistakes." Now, in there, when Lucas did that work, he slipped in an assumption which went unnoticed by most people, which was that, there's a unique equilibrium in the world. Now, as long as there's a unique equilibrium and we decide to model things in the way that he suggested following move, "There's a random variable," then you don't any longer need to inquire as to how people form their expectations.

All that matters is that the probability distribution of the things they believe to happen is the same as the probability distribution of the things that do happen. And if there were no uncertainty, that would simply amount to perfect foresight to the assumption that people know what's going to happen. Now, in the models that I work with, and indeed in all monetary models I would argue, there are always multiple equilibria, there's always more than one thing that can happen in the future. And you're seeing that coming back to old and New Keynesian's now as they struggle with, how to think about periods of protracted deflation that we're seeing. When you live in the world where multiple things going to happen, you have to ask yourself, "How would a human being living in my artificial environment react to the fact that more than one thing can happen.

And here, what I'm suggesting is simply that we go back to Milton Friedman's notion of something like adaptive expectations, where people use what I call a belief function. And a belief function is just a generalization of that idea of adaptive expectations, it's the idea that in order to figure out what they think will happen, people look at what's happened in the past and they project forward to the future. So a belief function is a mapping from current and past observable variables to probability distributions of a future variables. And in a world of multiple equilibrium, it's the belief function which is the, "Lets, what will happen."

Beckworth: And the belief function is where the animal spirits come in, correct? It's the mechanism through which the psyche can determine what equilibrium you end upon, correct?

Farmer: That'd be correct. Yeah. So what I-

Beckworth: So that's what... Go ahead.

Farmer: So when I said the probability distribution of the future variables, I was giving myself room for people to react to shocks. And those shocks can be what I call a belief shock, that is, markets can lose confidence overnight and revalue what they thought paper assets will look.

Beckworth: Yeah. What was interesting to me as I read your book is this idea that beliefs themselves are fundamentals. I mean, just like productivity technology, levels of education, beliefs themselves are a key fundamental input to production. And your argument is they can determine a steady state, they can determine where an economy... I mean, there can be good equilibriums, there can be bad equilibriums. Now, I found myself asking this question, and I'm guessing you would argue I'm reverting to the old natural rate hypothesis thinking when I do this, but I asked, "Okay, do these belief functions... Is this a short run versus a long run phenomenon?" And I think that's my bias towards the natural rate hypothesis, because the natural rate hypothesis says, "There's a distinction between short run and long run." But I think your argument, I've read your book correctly is, the belief function can affect the long run too, not just the short run. Is that right?

Farmer: That's right. So let me give you a way of thinking about that. I wrote a paper a while back in the EJ, Confidence, Crashes and Animal Spirits. And there, the idea is the following, it's the simplest formalization of what I talk about in the book. So imagine some Lucas tree economy where people need to pick the fruit. So it's a Lucas tree model as a simple asset pricing model, and Lucas asked, "If a tree is a random flow of apples, how much would you be willing to pay for ownership of the tree?" So take that model and think about the fact that the firms who own the trees and you're going to now own shares in the firms, have to employ people to pick the apples. And imagine that the labor market is a search market subject to all the issues I talked about.

Well, if you were Diamond, Mortenson and Pissarides, any issues with that you'd say, "Fine." There's a bargaining weight, that bargaining weight determines the volume of employment in the labor market that determines the apples that will be produced into the infinite future, and that determines how much you would pay for the tree. Now, ask yourself what the world would say about the 2008 financial crisis. So what we observed in 2008 is a massive crash in the value of paper assets followed by a recession, that is the tree falls in value and there are a few apples in the future because people are picking them. Now, if you're coming from the fundamentalist view of the world, you would have to look for some... What you'd say is that standing in 2008, before the collapse of Lehman Brothers, people trading in the asset markets looked ahead and rationally anticipated that some fundamental event would mean that there would be less output in the future.

And so they downgraded their value of the market, and one thing might be say, an illegal decision that changed the bargaining weight in the favor of corporations instead of workers. Or the other way around, that would be an event that would change the value of assets. The other view of what happened is … Paine's beauty contest view. So Paine's had this view of the asset markets, he said, "The stock market is like a beauty contest in which the judges and not judging how beautiful the contestants are, they're judging how beautiful they think the other judges think the contestants are."

That's a lovely metaphor for thinking about a collapsing confidence, if you think that you won't be able to resell your shares, you don't value them now. And if that drop in paper wealth persists, people then are going to take decisions about how much they spend and if drops in the stock market last for more than a few months, and the stock market stays down, people stop spending. If they stop spending and if it's the bargaining weight which is endogenous as opposed to the value of the asset, you really see the whole process in reverse. That's the story that I tell.

Beckworth: I was wondering as I read your book, whether you think there's any similarity between what you're doing with this belief function, which is a way of modeling animal spirits and John Cochran's recent push for putting the discount rate front and central and understanding macroeconomic crisis. So he gave a 2010 presidential address to the American Financial Association was published in the Journal of Finance. He argues, when you look at asset pricing equations, it's really the discount rate that shapes the near to mid-term the business cycle frequency.

It's not the numerator, it's not the future earnings, it's the discount rate. And he's really stressed, we need to think more and longer and harder about what causes the discount rates to change. And it seems to me, at least that there's some similar story here between what you are arguing in terms of the animal spirits and confidence shocks and discount rates, to the extent that discount rates reflect changes in mood, changes in confidence. Do you see any similarities?

Farmer: Yeah, I understand. So when John says it's the discount rate, in his representivation framework, that's the only way that he can understand a change in price earnings ratios. Because the price earnings ratio is essentially a discount rate, there's an arbitrage condition which says that there's a connection between the ratio of the price and the dividend to the return you can get from borrowing. So what I'm giving John here actually is a theory of why he's perceiving that to be changes in discount rates, and it's because the representative agent is the wrong way to think about financial markets. And I go beyond that, it's because finite numbers of infinitely of families trading in asset markets, there's also the wrong way to think about financial markets. Incomplete participation in financial markets is critical to understanding how there can be changes in asset markets that rich people can't arbitrage away.

That's actually another thing I draw attention to in the book, which is that it was a remarkable couth for Farmer, that's J. Farmer to call his work, the Efficient Markets Hypothesis, because he slipped in a change in the definition of the way we use the word efficiency and economics. So in financial economics, when you say markets are efficient, you mean they're informationally efficient, meaning that nobody can make money essentially by trading in the markets unless they had inside information. That's very different from the other use of efficiency which is Pareto efficiency, which is the argument that there's no intervention that a government could make that would improve welfare. And once you have incomplete participation, you can have informational efficiency without having Pareto efficiency.

Beckworth: All right. Very fascinating. So we have the problems of incomplete labor markets, incomplete financial markets. We have the issue of animal spirits, belief, functions that can put us in any number of equilibrium or outcomes. And so you want to stabilize all that, and you propose doing that by targeting an index of asset prices. And you talk about doing that through ETFs, but could you share with our listeners how you would go about doing that?

Farmer: Yes. So the goal here with the me... So let me first of all just reiterate the mechanism that leads to this. I believe that animal spirits lead to variations in big fluctuations in the value of the stock market, when those are persistent, they lead to big drops in consumption spending and big drops or changes in economic activity and that Pareto inefficient. I think the reason that, that Pareto inefficient is because of market incompleteness, the inability of future people to ensure themselves. We have a mechanism that could substitute for that and that's government. I want to aside here because some of the people listening to this will probably jump in and say, "Well, why do we need government and why can the family do this." Barrow showed us for example that, if people love their children, they could step in and correct one of the inefficiencies from overlapping generations structures.

That argument does not work here, and it doesn't work because in order for the family to ensure their children or their grandchildren, you would need to leave assets to your children, which had positive value in some states of the world negative values in other states of the world. And although those trades would never be observed in the unique socially optimal equilibrium that it's the ability to make those trades is what enforces the equilibrium. You're not allowed to make those trades because we don't allow that bondage in Western European legal codes. So I wanted to throw that in because I'm now going to talk about why government has a unique ability to make a trade, and the reason is that the national treasuries have the ability to make the trades on the half of our children and our grandchildren that they would make if they were able.

Now, what would those trades look like? I think it's very dangerous to have governments buying and selling individual stocks. I think to a first approximation, capital markets probably work efficiently in a Pareto sense in the sense of moving capital between different industries. It's the intertemporal problem that I have difficulty with. And so to avoid the political issues involved in government buying and selling individual companies, I'm advocating the creation of an exchange traded fund define as broadly as possible on the whole market and valuated. And imagine, so the way this would work is the Fed or the treasury would go to private companies like BlackRock or Fidelity and they would say, "These are the characteristics of ETFs that we would be willing to trade," and Fidelity or BlackRock would set up these ETFs and they would start trading.

If the price of the ETF is currently a 100 as of next Tuesday, it will be trading at 110. And the reason that we think it should be trading at a higher value is because the economy is falling into recession, we think stocks are currently undervalued. That's our best guess as to where they should be trading, and the price of the ETF should be growing 3% annual rate over the next month. Now, you are a private investor and you recognize that the treasury or the central bank, whichever, Roger's government is operating this, is a very big actor that has the ability to enforce this price. So you go and you buy out the companies that exist in ETF and buy them up because you know that as of next Thursday, you'll be able to sell them at a higher price. If the announcement is credible, it's entirely possible that the treasury would never actually have to carry out any trades, they simply have to credibly announce the price in advance.

And the market will establish that price. Now, just as it's possible for the market to be undervalued, it's also possible for a market to be overvalued. And if the economy overheats, it would be necessary for whatever body was put in place to make these trades to essentially take away the punch bowl as Greenspan put. So I would envisage an institution, something like a modern independent central bank, something like a federal open market committee. We have an analogy in the UK right now, the UK has set up a financial policy, which is charged prudential regulation. And nobody seems to know what that means or what they're supposed to do, I had a particular plan in mind with what they should do. I think they're the obvious body or group, the scheme that I'm talking about.

Beckworth: So maybe a way of thinking about this is, traditionally we talk about a monetary policy that there needs to be a nominal anchor, the central bank needs to create a nominal anchor. And what you're arguing for here is, we need to have a belief anchor, we need something that can maintain and stabilize beliefs going forward. And knowing the book, actually call for both, you want to maintain an inflation target as well as have this asset price index targeted as well. I was wondering if you can draw any insights or any thoughts from what's happened in Japan, the Bank of Japan has bought up a large number of ETFs over there. I know it's a totally different setup, but are there any lessons to be learned from that experience that would help shed light for what you want to do?

Farmer: I think it was a tremendous experiment. I think looking to see what happens is going to teach us a lot. I wouldn't have gone about it in the same way that the Bank of Japan did, which is working on quantities, I instead be trying to implement price targets. The ideas I've talked about here are actually very close to Scott Sumner in many ways, and I've had conversations with him about this. He and I... When I first heard Scott talking about a nominal GDP targeting, I thought he was just talking about moving around interest rates to try and target normal GDP, he wasn't. Scott's talking about Fed interventions in the GDP futures markets, which is a really quite a good substitute for operating in ETFs that I'm talking about. So he was active trades in asset markets to maintain a nominal GDP target, where I differ from Scott is the second part here is, he thinks that's enough.

I think if Scott were here, he'd be arguing that the... he probably has the rocking horse model in the back of his mind, I'm not sure that's enough. So as you pointed out, I want to have these two targets. One is essentially targeting nominal rates and the other, you could think of it's targeting some real rate. Now, interestingly, this also turns me into what people have been calling Neo-Fisherian recently. I'm saying this because you asked me about Japan, and Japan has been trying everything they can to get inflation up by lowering interest rates and moving to negative rates.

Also, we've seen that now in Europe and Switzerland, probably not immediately in the UK, but I could see that happening potentially even in the US. I think that's exactly the wrong way to get inflation back up. I think if you want to restore a 2% inflation target, probably the way to do that is to raise interest rates to 2% and offset the inevitable recession that will occur with some version of the asset price targeting I'm talking about. If that was in feasible politically, I'd even go for some fiscal stimulus, but I think that's very much second best.

Beckworth: Do you foresee any time that the two goals could conflict with each other? So the inflation target, the asset price target, is there possibility of these would ever conflict with each other, or would they always be mutually consistent?

Farmer: In my worlds, they're mutually consistent only because I give up on the natural rate hypothesis, so okay. So sometimes people talk about the natural rate of interest and the natural rate of unemployment as if they were the same thing. In my world, you could be at the natural rate of interest in a real sense with any unemployment natural rate. And the reason because, if you get to the natural interest rate, what is that? It's the ratio of a price to a flow, right? It comes to the value of the stock market, for every real interest rate, what's the price earnings ratio.

The price earnings ratio could be correct with a price of a 100 and a dividend of a 100 or a price of 200 and a dividend of 200. And the dividend of 100 might be associated with unemployment rate of 10%, the dividend of 200 might be associated with an unemployment rate of 20%, for example. So it's the multiplicity of steady states that enables me to make the statements about there being potentially two prices that you may be able to target independently.

Beckworth: Very interesting. We are out of time. Our guest today has been Roger Farmer. Roger, thank you so much for being on the show.

Farmer: Well, thank you so much for talking to me. I've enjoyed every second of it.