Peter Ireland is a professor of economics at Boston College, a research associate at the National Bureau of Economic Research, and a member of the Shadow Open Market Committee. Peter joins David on the podcast to discuss the nuts and bolts of Federal Reserve policymaking, the role of monetary aggregates in monetary policy, and how more complex measures of money can teach us a lot about the stance of monetary policy.
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: Peter, welcome to the show.
Peter Ireland: Thanks, David. Thanks for having me on.
Beckworth: I'm glad to have you. As a person who's read your work over the years, it's a real treat to have you on the show. I want to begin, as I do with most of my guests, and ask, how did you get into macroeconomics?
Ireland: That's a great question. Actually, my answer is going to parallel, but then maybe go a little bit beyond, the story that Bob Hetzel told you just about a month ago when he was on the show. Bob and I were both undergraduates and graduate students at the University of Chicago. He was there in the early to mid-1970s. I think he finished up in '75. I arrived at Chicago in 1984 and stayed through the early '90s, so 10 to 15 years after Bob.
Ireland: By the time I got to Chicago, Friedman was gone. He had moved to Stanford at the end of his career, so I never really had personal contact with Milton Friedman, David. But of course Bob did. Bob was Friedman's student. But definitely, in Chicago, the spirit of Friedman is still very much alive in the 1980s, and don't forget, going back to that time, it was by no means clear how the Cold War was going to resolve itself. It was by no means clear which economic system, capitalism, socialism, communism, something in between, would really prove to be most successful in the long run. Especially, I think, as an undergraduate, because as a PhD student, really a lot of your effort goes into bringing yourself up to the frontiers in terms of the technical aspects of our field, but in terms of substance, as an undergraduate, for lack of a better phrase, I mean the Friedman spirit, that was something really exciting to me.
Ireland: You felt like these were important issues, issues that were very much still open for debate. That was a really exciting thing to be part of. Another point that Bob mentioned a few weeks ago with you that resonated with me was the Chicago approach to price theory. The idea that you could use the basics of the basic analytical framework given to us by consumer and producer theory to address a wide range of questions having to do with economic behavior, human behavior, and environments that went well beyond the sort of canonical textbook descriptions of utility maximizing, consumer choosing bananas versus apples, or the firm hiring capital and labor input. So, when I was a senior in college, and getting ready to apply to PhD programs, I was most excited about that kind of applied micro, maybe the Becker-Rosen style of labor, human resource economics.
Ireland: But before I finished my undergraduate days, it was actually the fall of my senior year, I took a couple of PhD courses at Chicago, including one taught by another one of your former guests, John Cochrane, and that class-
Ireland: ... was really the one that kindled my excitement in macro. One of the problems that we have even today teaching macroeconomics to undergraduates, but perhaps the problem was even more pressing 15 to 20 years ago, it just has to do with the fact that PhD-level macroeconomics requires a set of mathematical and statistical tools that really go well beyond what most college sophomores or juniors have. In a 35-person course, not everyone's obviously going to be a math or statistics major. So, I liked macro as an undergraduate, but to me, it seemed to lack the technical sophistication and rigor of price theory. But when I got to John's class and we did dynamic programming, and stochastic models, dynamic models, that really showed me that the same kind of analytic rigor, the same kind of technical sophistication that could be brought to issues in micro could also be brought to study issues in macro, and that's really what won me over.
Beckworth: Interesting. I didn't realize John Cochrane was your teacher.
Ireland: He was.
Beckworth: Fascinating. Now, you also studied under Bob Lucas, is that correct?
Ireland: I did. I mean, at the time, Chicago, it still is of course, but when I think about the faculty in the area of macro that they had at the time, my PhD thesis committee consisted... My principal advisor was Robert Townsend, but also on my committee was John, Bob Lucas and Michael Woodford, so-
Ireland: What a way to learn some monetary economics, being able to take something away from each of those guys. I mean, it was really incredible.
Beckworth: No kidding. Both intimidating and a great experience at the same time to learn from them.
Ireland: But Bob was great. I mean, even if I had not had him as a teacher, his scientific contributions are so great and so, just like Friedman, if you do work in macro, you cannot not be heavily influenced by Lucas. But in addition, Lucas was just an amazing, inspiring teacher. In his first-year PhD class, he really conveyed, like John did, the excitement of rational expectations, dynamic and stochastic macro to us as first year PhD students. Then, especially my memories of Bob come from later on in the more advanced topics classes that he taught to us as we started writing our PhD theses. What Bob used to do was just to pick the topic that he was interested in at the time, so one year optimal taxation, another year economic growth, and he would go through the hardest papers in those subjects just sort of before you on the board in class.
Ireland: And so, what I really took away from that is no one is endowed, or almost no one, not even Lucas is sort of endowed with the intuitive knowledge that in an instant can bring him or her up to the frontiers on any given topic. What you have to do is go to the library, or these days go online, find the hardest papers, work through them in great technical detail, so that you really understand what the assumptions are, how you go from the assumptions to the results. That's what Bob did for our eyes, and that was really inspiring. That was a learning experience, not just in terms of substance, but in terms of method. How do you do research on your own?
Ireland: I actually have one more story about that, too.
Beckworth: Sure, please share.
Ireland: He probably doesn't remember this, and John Cochrane doesn't remember it either, but I do. There was one summer when I was working on my thesis, and I was converging on the library that morning together with a few other econ PhD students, and we were standing on the steps talking about something, that I don't remember what, and just then Bob came up the walk, walked up the stairs and said his hellos, but kept walking quickly, as quickly as he could into the library, I guess to find a book or a journal. And just a couple minutes later, John Cochrane came riding by. He used to ride his bike to work, and he stopped, looked at us and said, "Hey guys, I think Bob is trying to tell us something."
Ireland: And you know, we all laughed, went into library and got to work as fast as we could. But that's another thing-
Beckworth: Hilarious. That's great.
Ireland: I mean, it just goes to show you that again, a lot of it is going to the library, reading the papers to keep learning. It's not like a field where you just sit there and wait for inspiration to come. You make it happen yourself.
Beckworth: It's hard work.
Ireland: That's right.
Beckworth: Yep. Well, that's a great story. It's a rich environment, to have other thoughtful people around you like that to inspire you, and to point you in the right direction. Well, let's move on into monetary policy, since you have yourself done a lot of work in this area in addition to having great teachers. And I want to begin for our listeners, because often on this show I just kind of assume everyone knows what monetary policy is, how it's done, and even I think among people who do know monetary policy, of journalists, observers, even some economists, I think sometimes we forget or gloss over kind of the institutional details, and I want to begin by asking you some very basic questions, and that is the following: What are the instruments of monetary policy? What are the intermediate indicators of monetary policy? And then the policy objectives? And then kind of tie those all together. Why is it important to know the distinction between those, among those, and then what is it useful in terms of understanding monetary policy?
Monetary Policy Basics
Ireland: Right. Well, let me preface my remarks by saying that when you use these terms, instruments, intermediate targets, targets and goals, the lines aren't always completely clear, and so people will debate or pick on you for saying something that they don't exactly agree with. But we could start from the beginning just by observing that the Fed Reserve and most other central banks around the world conduct monetary policy today by managing some short-term nominal interest rate, an overnight interbank loan rate like the federal funds rate, or perhaps if the central bank is running what's called a floor or corridor system, this is more akin to what the Fed is trying to do now, they might manipulate directly the interest rate that they pay on reserves and the interest rate that they charge on loans to banks, what the Fed Reserve calls its discount window facility, in order to manipulate the short-term interest rate and interbank rate.
Ireland: And so, we could argue about whether the federal funds rate of the United States is really an instrument, in the sense that it is not an idea created by the Federal Reserve, but most monetary economists I think would identify the federal funds rate or some short-term interbank interest rate as the instrument of monetary policy, meaning that it is the focus of the central bank's day-to-day actions. When you ask why, why most central banks around the world conduct monetary policy with reference to interest rates there, I think it's because of the famous results presented in a paper by William Poole, from the 1960s, that shows that if the demand curve for reserves, or the demand curve for liquid assets, monetary assets, exhibits a lot of high-frequency volatility, by setting or targeting a short-term interest rate and then just letting the supply of reserves adjust elastically in the short run, as opposed to pegging the quantity of reserves, the central bank can stabilize interest rates by automatically accommodating those shifts in demand, and thereby contribute to financial stability and prevent the volatility in interest rates that would otherwise occur from spinning over to the real economy, perhaps.
Ireland: Now again, different people have different views of that result. To me, it seems reasonable enough. I think it makes sense to make it as easy as possible for financial institutions, non-bank, public, to manage their holdings of liquid assets, and if that involves smoothing out fluctuations, very high-frequency fluctuations in the money markets, and if that can be done through targeting something like the federal funds rate, then that's worth doing. But-
Beckworth: Let me ask one question here. So, central banks do in practice use some kind of short-term interest rate to operationalize, to do monetary policy, the instrument you would you call it, but they could, as you're suggesting, also go the other route, and then directly control the growth of the monetary base.
Ireland: That's right, so I was... You're anticipating exactly what I was going to say next.
Ireland: So, even if you take the view that it's a sensible approach to target an interest rate to smooth out very high-frequency fluctuations, there's still two things to keep in mind. And the first one is that the federal funds rate is an interbank rate, which is determined in the market for very short-term interbank loans. It's not something that the Federal Reserve, as I said, calls out by fiat, or it's not a price that is fixed by law. How does the Federal Reserve bring about outcomes where the actual funds rate is close to target? It's by conducting open market operations and managing the quantity of bank reserves that it supplies.
Ireland: Ultimately, the Federal Reserve's ability to conduct monetary policy rests on its status as the monopoly provider of bank reserves and currency. So, this is a source of some debate or disagreement about what really can be called an instrument of monetary policy. If you want to ask what is it ultimately that the Federal Reserve actually controls, it's reserves, or more broadly the monetary base reserves plus currency. So, what would happen if the Federal Reserve abandoned federal funds rate targeting, and instead simply managed the quantity of reserves, or the monetary base? Interest rates, especially very short-term interest rates might exhibit, probably would exhibit more high-frequency volatility. I don't think that that would be a big deal as long as everyone understood that these were high-frequency movements that would average out over a period of days or weeks. The macroeconomic effect should be minimal, so in that sense nothing that the Fed could do with a funds rate target couldn’t also be done just by targeting the monetary base instead.
Ireland: And so, I see it as a kind of equivalence there. But another point that many people I think forget, and here perhaps more to their peril, is it's easy to go from a mentality that says, "We're going to target the funds rate to smooth out very high-frequency movements in overnight rates, to make it easier for banks and money market funds to manage their liquidity positions and so on." But the danger there, and I think we're seeing how dangerous this can be in the United States today, the danger is going even further and beginning to think of all fluctuations in all interest rates, even longer-term interest rates, and then all fluctuations across all prices, for all assets, as something to be avoided, and something to use monetary policy to clamp down on.
Ireland: It's important to remember as an economist that changes in prices convey information to the demanders of and suppliers of any good or service, in this case liquidity, and attempts by the Fed or any other government agency to smooth out, or clamp down on, or regulate movements in prices rarely end up achieving welfare-improving ends.
Beckworth: Do we have any evidence of a time and a place where there was no central bank, or a central bank targeted the monetary base, and allowed short-term interest rates to fluctuate? And what was the result?
Ireland: Right. Well, I'm not a monetary historian, so I can't speak intelligently about going too far back in time. I mean, under Volcker, during the early stages of the Volcker disinflation, it was certainly more emphasis on the monetary base, or measures, quantity measures of reserves, on more tolerance for fluctuations in the federal funds rate, although even today, I think there's still some debate about whether that was rhetoric, or whether that was something that Volcker really, and the FOMC really did do. I think here what you want to keep in mind is an important distinction in monetary macro more generally, and that's to think about the difference between a movement in the federal funds rate within a given regime, versus a movement in the federal funds rate across two different regimes.
Ireland: Today, if we suddenly saw the federal funds rate even jump up by 100 basis points, financial markets would be in a panic wondering what on earth was going on. But within a new monetary regime, where the Federal Reserve just said, "We're now conducting policy by managing reserves or the monetary base, and we're going to allow fluctuations in the federal funds rate." A fluctuation up and down by 100 basis points, again, would average out over a period of days, and through expectation channels wouldn't really have any effect the longer-term rates that govern spending decisions by households and firms.
Beckworth: Okay, so the Fed and most central banks actually use short-term interest rates, they adjust those based on some objective, and the monetary base, as you said, kind of responds based on what the target interest rate may be. If the Fed right now wants to keep the interest rate at a particular level, it will accommodate demand. Actually, now's not a great example because of the huge supply of reserves, but let's go back to a more normal period when there's a smaller stock of reserves. The Fed would set some particular interest rate for a period, and as the demand for those reserves would adjust, the Fed would simply adjust the supply. Is that the right way to think about it?
Ireland: That's right.
Beckworth: Okay, and so in that sense then, the monetary base or the bank reserves become endogenous, as economists would say, or would be kind of driven by the system that it's imposed upon it, and the alternate scenario we outlined is well, it could be a case where the Fed targets the monetary base and allows interest rates to be endogenous, but the potential danger you suggest would be is that there could be more fluctuations, and it depends on the regime, and the time and place. All right, let's move from the instruments then down to the intermediate indicators.
Beckworth: So, the instruments, just to summarize, are the tools, the operational approach the Fed takes to implement monetary policy. What role does the intermediate indicator play and what is an example of it?
Intermediate Indicators of Monetary Policy
Ireland: That's a great question, too. Well, it may be easier now to jump to the ultimate goal of monetary policy.
Beckworth: Okay, let's do that. Sure.
Ireland: Which would be inflation and probably some measure of real economic activity, the output gap, or the unemployment rate. In the United States, it's Congress that gives its instructions to the Federal Reserve, and the so-called dual mandate loosely speaking tells the Fed to pursue the goal of price stability and also maximum employment. And now the question is how do you successfully manage your instrument in order to achieve your goals for the ultimate target variables, goal variables, inflation or unemployment, given that there are in Milton Friedman's famous words long and variable lags between policy actions and the impact that those actions have on inflation and unemployment.
Ireland: And so, the role of an intermediate target is to give policymakers real-time information about whether their actions relating to the behavior of the instrument, or their setting for the instrument, are appropriate if they wish to achieve their long-run goals. So, examples of intermediate targets might be some broad measure of the money supply, which by according to quantity theoretic logic, sustained movements in the broad monetary aggregates would be an indicator that there's upward pressure on prices, and perhaps there are real effects in the short run on unemployment, as well. So, the Fed could use information in the monetary aggregates to guide its decisions for the funds rate or for reserves, or forecasts of the goal variables based on real-time information I suppose could also be thought of as an intermediate target.
Beckworth: All right, so where would the Taylor rule fit into this picture? Well, maybe you should explain the Taylor rule to our listeners. I'll let you do that, and then place it where you think it fits in kind of this map between instruments, intermediate, and then final goals.
Ireland: Right. Well, the Taylor rule in its simplest form is a simple mathematical equation that prescribes a setting for the federal funds rate as a function of the output gap and the inflation rate.
Ireland: So, it provides a guide for monetary policy makers. If the goals are to stabilize output and to stabilize inflation, and you see output being higher than potential and or inflation being above target, the coefficients from the Taylor rule on those two variables being positive, the Taylor rule will dictate that the central bank should raise the federal funds rate, tighten monetary policy, in order to bring output back to potential, inflation back to target, and vice versa. If inflation is too low, below target, or if output falls below potential, the Taylor rule would dictate to the Fed that it should lower the federal funds rate to rekindle inflation and bring output back up to potential.
Ireland: So, in a way, I mentioned the monetary aggregates as an intermediate target and we might want to talk about the role of monetary aggregates in more detail later. The Fed really does not use any more the monetary aggregates as intermediate targets. It instead has forecasting models, and if you want to ask how does the Fed operationalize the instrument intermediate target goal variables framework today, it's probably through using its own forecasts as intermediate targets. But the Taylor rule provides another way of finessing this issue. It too does not make reference to any measure of money, but instead it allows you to ask. Today, for example, we see inflation about 25 basis points below the Federal Reserve's 2% target. The output gap I believe is still slightly negative, so that tells us intuitively that monetary policy should remain accommodative, but quantitatively how accommodative, the Taylor rule provides an intertemporally consistent way of saying based on the Fed's historical behavior, given inflation, given the output gap today, what is the normal setting for the funds rate? And we can use that as the starting point for our discussion of the current environment.
Ireland: Do we want monetary policy to be easier than it would be normally under these circumstances? Tighter? Or about right, so to speak, in historical terms? The Taylor rule lets you do that.
Beckworth: All right, so if we think of that dual mandate prescribed by Congress, some measure of real activity and stable prices, and I'm going to just make this easy and call it nominal GDP as a rough approximation of that goal, nominal GDP being the total dollar amount of spending in the economy, and that would capture both of those elements to some degree. If that's our policy goal, we have our instrument, the interest rate, and we need something in between to help guide us, help us adjust our interest rate to hit that goal, the dual mandate goal, the Taylor rule is one way to do that, because it has components of that, has the output gap as you mentioned and inflation in it. But going back to money, money could also be in there, right?
Beckworth: I mean, if you took away the Taylor rule, you're suggesting that you could use money or monetary aggregates as a way to guide the path of interest rates in a manner that leads the Fed to hit the dual mandate.
Ireland: That's right. Well, go back first to the point we made about nominal income. One of the nice... I mean, when John Taylor originally proposed his rule in 1993, he had separate terms for the output gap and inflation, and higher weight on inflation relative to the output gap. If you just replace the output gap with an output growth, which is something that people like to do sometimes in economic models as well, and then on the right-hand side you've got both of the components of nominal GDP growth, and if you set the two weights equal to one another, you have a variant of the Taylor rule that adjusts the nominal interest rate, the federal funds rate, in order to stabilize growth in nominal GDP.
Ireland: So, you can think of nominal income targeting as something of a special case, perhaps, within a policy framework that uses the Taylor rule as a guide. Now let's dig a little bit deeper and ask, "Well, what role if any might be served by the monetary aggregates?" In the context of a Taylor-type rule, the role for the monetary aggregates would come in terms of their ability to forecast either future output or future inflation, so Milton Friedman's advocacy, for example, of policy strategies built around the money supply is based partly on the idea that the lags between policy actions and their ultimate effects on output, and especially on inflation, were so long and so variable that if the Federal Reserve tried actively to stabilize inflation, more often than not it would be moving in the wrong direction and destabilizing instead of stabilizing.
Ireland: So, if instead you think that there is a reliable link between money growth and growth in prices and or output by focusing instead on money growth, you can in a sense do more by trying to do less. By stabilizing money you can still provide a stable backdrop of price stability in the long run, but avoid costly mistakes that get made because you're constantly reacting to incoming data on output and inflation themselves, which tell us what prices are doing today, but don't necessarily tell us what effect monetary policy is having on future prices today.
Beckworth: Yeah, and it's interesting, as you mentioned earlier there is no money in monetary policy today. The Fed, using the Taylor rule or Taylor rule-like thinking looks at output gaps, inflation, but that wasn't always the case, right? In the '70s, and I think the '80s some extent, they looked at monetary aggregates. So, I guess my question as I'm sitting here listening to you, was the Taylor rule kind of a replacement or an attempt to get past the monetary aggregates, which we'll come to this discussion in a minute why they didn't seem to work at that period, and we'll get into discussions of Divisia and simple sums, but monetary aggregates were dropped. They seemed to-
Ireland: That's right. So, I think you can look at some of the popular writing, the speeches that John Taylor has given describing the intellectual origins of the Taylor rule. I think that for John, it wasn't so much driven by any kind of ideology so much as an attempt to be as helpful as possible to policymakers as they made their policy decisions in real time. So, if you go back to John Taylor's work in the late 1970s, he had a paper in Econometrica that used optimal control theory to characterize monetary policy and it was a linear rational expectations model, where wage contracting gave rise to monetary non-neutralities in the short run. But there his proposed rule, if you will, took the form of a rule for adjusting the money supply.
Ireland: I think the 1993 rule was prescribed, it was cast in terms of interest rates in large part because by then that is how the Federal Reserve was conducting monetary policy, and so John... If you look at the title of that paper in 1993 by Taylor, it's called something like “Discretion in Policy Rules in Practice.” So, if you underline in practice, I think that's what he was all about. Giving policymakers a useful guide for decision making at the time and today, and recognizing that probably because of the analysis, the decision had been made to switch to the funds rate.
Beckworth: Another way of saying that, my earlier question, and related to what you just said is did central banking give up on monetary aggregates? And then did the Taylor rule in practice replace them?
Ireland: Well, yeah, the simplest way to answer those questions is yes.
Ireland: I mean, let's not forget, though, that the ECB in the early days did have the two-pillar approach.
Beckworth: That's true. Right.
Ireland: And to be honest, I would prefer, I think it's a mistake for the ECB to jettison that second pillar, and I think it was a mistake by the Fed to give up completely on the monetary aggregates.
Beckworth: Okay. Well, let's move into that question, then, of whether money still matters or not, and as we just said, money fell from grace so to speak in terms of central banking practice, and that's because they were looking at monetary aggregates, M1, M2, based on simple sum measures. Can you explain what a simple sum measure is, the shortcomings with it, and what is the latest developments in terms of better ways to measure money?
Monetary Aggregates: Simple Sum vs. Divisia Measures
Ireland: Sure. Well, the simplest way of describing it is in terms of the same sort of aggregation theory it used for, say in computing real GDP, and to think about how we describe the process of completing real GDP to undergraduates in a macro principles class. A simple sum monetary aggregate is called simple sum because it basically takes a collection of assets, for M1 that would be currency and checking deposits, basically, and then for M2 it would include a variety of other bank deposits, as well. Money market, mutual funds, as well, and simply sums up their total dollar value. Ignoring the fact that in terms of the liquidity services they provide, currency is an extremely liquid asset. You can walk into just about any store and buy stuff with currency. You can usually pay for things by check. In fact, sometimes for certain types of transactions it's easier to pay by check, but then you think about say a savings account, where you've got to go to the ATM machine first, and you're limited in the withdrawals you can make. Or money market mutual fund where again, you're not going to be allowed unlimited check-writing privileges.
Ireland: The simple sum approach, although it's simple from a mechanical perspective, misses the fact that these different assets provide different amounts of liquidity services. So, in that sense, going to making the analogy to real GDP, it would be like as we say, we can't add apples and oranges. You can't add apples and luxury sedans. You wouldn't say that real GDP in an economy that produces two luxury sedans and one apple is the same as an economy that produces two apples and one luxury car. You can't just add them up and say in both cases the GDP is three. And yet, that in a nutshell is what simple sum monetary aggregation does.
Ireland: Alternative to simple sum aggregation, work proposed first by William Barnett in the 1980s based on aggregation theory, and so the question Barnett asked is if we wish to come up with more reliable metrics of monetary services, how on earth do we actually measure quantitatively how much extra in terms of liquidity is provided by currency, or a regular checking account, let's say, relative to a money market mutual fund? And one way of doing that might be just to guess. Say, "Well, maybe it's half of the liquidity services," and then to give it a weighted average, but what we do over on the GDP side is not to guess and say, "Well, we're going to say that a luxury automobile is the equivalent of 10,000 apples." Instead, we use information that's communicated by the price system about the relative value that consumers place on goods of different kinds, and so Barnett's insight is that we could use interest rate differentials to gauge how much liquidity services optimizing consumers are getting from different monetary aggregates, and from those interest rate differentials make inferences about how the different assets should be weighted.
Beckworth: Okay, so if we take this Divisia approach, we get a much better measured aggregate of money and money assets, and the issue is that back in the '70s, the '80s, I guess even the '90s, when there were a number of studies that found that money didn't do a good job predicting nominal income, or the relationship was breaking down, and one of the reasons the Fed and central banks abandoned money as an intermediate target, based on these studies, goes back to the question of how they measured money, right? Most of those studies, did they not use simple sum measures as opposed to Divisia measures?
Ireland: Yes, that's right, and in many cases subsequently, when Divisia measures replaced the simple sums in the exact same statistical test, somebody who's done great work along these lines is another one of your previous guests, Josh Hendrickson, and Josh had a paper a couple years ago, I think it's actually pretty recent, in the journal Macroeconomic Dynamics, in which he simply redid some of the statistical studies with one and only one change, and that's in the monetary series used. Showing that to the contrary, the information content in the monetary aggregates reappears when you use Divisia aggregates in place of the simple sums.
Beckworth: Okay, and you had a recent paper you co-authored. It was titled “A Working Solution to the Question of Nominal GDP Targeting,” I highly recommend to our listeners. But tell us how you use Divisia in terms of operationalizing a nominal GDP target.
Ireland: Right, that's a good question, too. Well, there were really two parts of that exercise. My co-author, Mike Belongia and I, we did take advantage of one of the problems in doing this kind of empirical work that Josh did and that we have done as well is that for a while it was difficult to obtain up-to-date data on the Divisia monetary aggregates. The St. Louis Fed provided them for a while, and then later on, Richard Anderson, Dick Anderson at the St. Louis Fed and Barry Jones were basically compiling the statistics on their own, and now Bill Barnett and his associates are compiling these data and making them available through the Center for Financial Stability website. So, part of it was just taking advantage of the new data on the Divisia aggregates that Anderson, and Jones, and Barnett had just made available.
Ireland: But there it was simply a matter of wanting to use the best measures of money, and based both on past empirical results and on the theory originally put together by Barnett, the Divisia aggregates were our clear choice. But in particular, what we showed in that paper, back in the 1980s and on into the early 1990s, at the board they actually were looking to some extent at the monetary aggregates, and they developed something called the P-Star model. The P-Star model was based on an assumption that in the long run, the velocity of money would return to its average or mean value, and so that if you looked at sustained growth in M2, you could compute the level of prices, P-Star, that the actual price level would gravitate towards as velocity returned to its long-run level.
Ireland: They were using simple sum M2, and maybe that was part of the problem. There were also issues with money demand instability, the instability of velocity for M2 shortly thereafter and the model was abandoned. What Mike and I showed in our paper was when you replace simple sum M2 with Divisia M2 or MZM, it didn't really matter much which aggregate we used. But in addition, and this was actually Mike's great idea, it's still not proved, even with the Divisia aggregates, that the velocity of money is a constant in the long run. It does seem to vary, and it varies for a couple of reasons. It varies because interest rates economy-wide change. That changes the opportunity cost to consumers to hold monetary assets, but it does seem like ongoing financial innovations have an effect on interest rates, as well, but Mike's idea was to kind of clean those low-frequency movements in velocity out of the series, or maybe a better way to put it would be to account for them and to track them just using a simple time series model that... It allowed for a time varying long-run average level of velocity, let's say.
Ireland: What we found was once we accounted for movements of that kind, the predictive power of the P-Star model for nominal income returned. If money growth for a sustained period of time is above the level that would be consistent with the velocity of its long-run goal, long-run average, then you begin to see upward pressure on nominal income growth and similarly, unusually slow growth in the Divisia monetary aggregates presage a slowdown in nominal income. So, the point of that paper was that if the Federal Reserve wanted to use nominal income as an intermediate target for achieving its dual mandate, it could do that in real time by conducting policy with reference to movements in the Divisia aggregates.
Beckworth: So, you show then in short money still matters. The Fed could still use monetary aggregate, as long as it's Divisia, to help guide its actions. Correct?
Ireland: That's correct.
Beckworth: Now, you also show in there, which is interesting when I first read that paper, something called a money gap. I think that's the term you used. But you show the distance between where the money supply is and where it would need to be for the economy to hit full employment. Is that a proper interpretation of that gap?
Ireland: Yes, basically. Well, just to elaborate ever so slightly, this goes back to the idea that there are... If you think about the equation of exchange, MV=Py, if you have a target for nominal income, Py, then you know that your ability to hit that target by trying to use a monetary aggregate as your intermediate target depends in large part on the stability of velocity. And so, what we did with the measure of m-STAR is to ask, accounting for the slow-moving trends in velocity brought about either by long-run movements, secular trends in interest rates, or perhaps financial innovations, where would M have to be to, accounting for movements in velocity, in order to hit the target for nominal income. And that's what we called m-STAR.
Beckworth: Okay, so maybe full employment's the wrong term. You're looking at where money would need to be in order to hit this level of nominal income compared to where it actually is, and that's what that money gap is.
Ireland: That's right.
Beckworth: Now, is there a mapping between that money gap and the interest rate gap that's in the New Keynesian model of thinking? So, the New Keynesian approach is there's some market-clearing interest rate, the natural interest rate, and the objective of policy should be to have the Fed adjust its interest rate to that magical natural interest rate which we don't observe. And the difference between those two kind of determines the stance of the monetary policy, so is there some kind of mapping between this money gap and that interest rate gap?
Ireland: The mapping would be there in the sense that if the New Keynesian model were to say that monetary policy is too loose, let's say, that would show up in a setting for the federal funds rate that was too low relative to the setting that would bring about a speedier return of inflation or nominal income to target. And similarly in our framework, instead of an interest rate that is too low, what you would see is money growth that was too fast.
Beckworth: Okay. Now, has this work gotten any recognition by central bankers? Are they taking another look at Divisia measures as a way to help them guide monetary policy?
Ireland: Well, I think sadly no. I think that's partly to their detriment. I should say that that's not entirely true. The Board of Governors, Ruth Judson and several of her co-authors had a very nice paper on M2 and what it tells us about the stance of monetary policy. That's a very nice paper at the Dallas Fed. John Duckett has done outstanding work on money demand, even through this period where the FOMC members themselves have certainly not been discussing the money supply, and I doubt discussing publicly the money supply, I doubt really thinking about the money supply either. So, it's not as though there's no one at the Fed that really cares about this, but-
Beckworth: But are these folks using the Divisia measure?
Ireland: Right, and that's another aspect of it.
Ireland: No. Now, I should say that if you take a look at the behavior of the Divisia aggregates and the simple sum aggregates over the last five to ten years, the differences are not that big, so I would guess that you could redo their statistics with Divisia or simple sum and reach the same conclusion. But no, I mean, I think that it is unfortunate that Barnett's work has not received the widespread recognition that it deserves.
Beckworth: All right.
Ireland: So, it is unfortunate that central banks have... It would be one thing to say that money demand instability or financial changes lead us to be skeptical of every movement we see in money growth. I think it's a mistake, however, to abandon all reference to money growth in a monetary policymaking framework.
Beckworth: Okay, so a question I've been dying to ask you is what monetary aggregate is the most appropriate in this day and age? Specifically you've mentioned M2, and let's just assume that we're talking about Divisias here, so we can put that issue to the side. But Barnett also has an M4 measure, and I've taken a liking to that because it includes institutional money assets. So, M2 would be more retail money assets. M4 would include all those retail money assets, but also have institutional money assets, and the reason that seems to be reasonable to me, and I would like to hear your views on this, is that the bank run that we had in 2007-2008 was on the institutional money markets. It was the institutional money supply that fell. So, M4, as you know if you look at M4, there's a much sharper decline. In fact, I think M2's probably relatively... That doesn't change much at all, am I correct?
Ireland: What is certainly true is that the growth rate of M4 has been lower than the growth rate of M2, yes. And a big part of that has been, has involved too what's gone on in the market for repurchase agreements, and the repo markets have been at the center of discussions of financial disruption, so for sure that's something that M4 will pick up and the narrower aggregates will not.
Beckworth: But from a quantity theory perspective, as a monetarist, you've got this spirit of Milton Friedman, Chicago School.
Beckworth: Is it appropriate to plug M4 into that equation of exchange, or am I going beyond what was originally intended?
Ireland: Right. Well, I don't think that... I think ultimately that's an empirical question, and the only way that you can really answer it is to say that for the particular study, the particular issue that I'd like to address, which level of aggregation gives me the measure of money that's most closely linked to the variables that I ultimately care about? I guess my own answer and my own reason for focusing on... You don't want to call them the narrow aggregates, because they're still broad aggregates, but more standard levels of aggregation like M1, M2, and MZM, is simply that those aggregates consist of liquid assets that are being held by the non-bank public, or the non-financial sectors that, and therefore my thinking would be they should be more tightly linked to actual spending decisions.
Ireland: So, if a firm, if revenues are increasing and they're hiring more workers, you would expect a firm to have higher balances on average in its checking account, and likewise if consumers have jobs, and are getting raises, and are spending money, their holdings of liquid assets will go up, too. Whereas once you get out to M4, you're capturing a lot of transactions that are going on between financial institutions, and that may be very useful for some purposes gauging levels of stress in financial markets, et cetera, but my hunch is that those measures would be less tightly linked to the actual spending decisions that determine say nominal GDP.
Beckworth: Okay. Well, let's move in the last few minutes we have to the Great Recession and the slow recovery that followed it, and you had a paper you had a paper you wrote recently for the Shadow Open Market Committee, where you addressed this question, at least maybe the post-recession period. The sluggish recovery. And you asked the question why has nominal income been so slow, the growth been so slow, and can you answer that question for us now?
Why Has Nominal Income Growth Been So Slow?
Ireland: Right. Well, I think the easiest way to... First, if you look at the data, it's definitely true that nominal income growth trended downward looking at the period. I mean, once you take out the period of the Great Recession itself, and just focus on the slow recovery versus what came before, there's clear downward movement in nominal income growth, so the question is why. If you decompose nominal income growth once again back into its two components, then just from an accounting perspective you're left with two possible explanations.
Ireland: One is that real economic growth decelerated, and the other is that inflation has been chronically slow. And in that Shadow Open Market Committee position paper, I think you have to concede that part of the slow growth in nominal income is reflective of slower real economic growth, which plausibly had something to do with things going on in the economy besides monetary policy. Slowdown in productivity growth, or maybe it's demographic changes, or maybe it has something to do with regulation, who knows? But if you take a look at inflation as well, inflation, like nominal income growth, has been below the Fed's target for quite some time, and if you take seriously the idea that inflation is in the long run a monetary phenomenon, I mean it's sort of getting to be the long run. And when you see inflation coming in again and again below target, you have to say that to some extent this is because monetary policy has been insufficiently accommodative.
Beckworth: Yep. Well, that is all the time we have for today. Our guest has been Peter Ireland. Peter, thank you for being on the show.
Ireland: Thanks for having me, David.