Patrick Horan and David Beckworth on *The Fate of FAIT* and the Future of the Fed’s Monetary Framework

In this special episode of Macro Musings, David Beckworth and Patrick Horan join guest host Carola Binder to discuss their newest paper, *The Fate of FAIT: Salvaging the Fed’s Framework.* Patrick Horan is a research fellow in the Mercatus Center’s Monetary Policy Program and Carola Binder is an associate professor of economics at Haverford College as well as a visiting scholar at the Mercatus Center. In addition to their paper, Pat and David also talk about the basics of flexible average inflation targeting, how it compares to temporary price level targeting, the differences between the Fed’s old and new frameworks, and a lot more.

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].

Carola Binder: Hi, Macro Musings listeners. This is Carola Binder. I'm a visiting scholar at the Mercatus Center and an associate professor of economics at Haverford College. Today, I'm very excited to be hosting this episode because your regular host, David Beckworth, is our guest, talking about a Mercatus Center working paper that he co-authored with Patrick Horan, who's also at the Mercatus Center. Their paper is called, *The Fate of FAIT.* That first fate is like destiny. The second is the acronym FAIT for Flexible Average Inflation Targeting. We will get into the details of what that is and what they think about it later in the episode, but, first, I wanted to give you a chance to meet Patrick in case you don't already know him.

Patrick Horan: Hey, Carola and David. Thanks for the very kind introduction. As Carola said, my name is Patrick Horan, and I am a research fellow at with the monetary policy team at the Mercatus Center. I joined Mercatus several years ago as a program associate, which is a more entry-level position at Mercatus. I was young, I was fresh out of college, relatively speaking, and I didn't know quite what I wanted to do with my career, but I wanted a job that was policy oriented.

Horan: I had studied economics and political science as an undergraduate, and I came across this job opportunity at Mercatus and, while working at Mercatus, I decided to get my master's in economics at George Mason as a part-time student. George Mason is one of the few places you can do graduate work in economics as a part-time student. While working at Mercatus, I got a master's, and then I decided to be a glutton for punishment and kept going, so now I'm getting my PhD. Coming to Mercatus, I knew a little bit about monetary policy, again, from studying economics as an undergraduate, but I really started to love monetary economics and policy as I started to learn a lot more here, and I've made that my research area. I study and write on monetary policy for work, and I'm also writing a dissertation that's focused on monetary economics.

Binder: Great. This paper that we're going to be discussing is one of the chapters of your dissertation, right?

Horan: Yes. That's the plan.

Binder: Can you tell us about maybe one of the other chapters?

Horan: The other chapters are still in earlier phases. One chapter compares different monetary aggregates, the simple sum measures that people are probably familiar with, M1 and M2, but then I also look at Divisia monetary aggregates. A few guests in the past, including Peter Ireland and Josh Hendrickson, have talked about Divisia. That's a different way of measuring monetary aggregates. It basically weights different forms of money based on how liquid they are. I use some vector autoregressions to try to predict the effects of money on different variables, including output and inflation.

Binder: Very cool. For today's show, let's first get into what exactly FAIT is, and then we'll talk after that about why and how you think FAIT needs salvaging. Yeah, just to start with, what is FAIT? When did the Fed start doing it? What is it?

The Basics of Flexible Average Inflation Targeting

David Beckworth: Well, FAIT, as you mentioned, is an acronym for the Fed's new framework which was introduced in August 2020, and it stands for Flexible Average Inflation Targeting framework. The Fed had a review under a process it was doing through 2019 to 2020, and the original motivation was how to deal with the experience we had in the decade leading up to the pandemic. As you know, we had a period of low inflation, low interest rates. The concern was we would be hitting the effective lower bound over and over again. There were several papers that came out that showed the likelihood was increasing. We would hit the lower bound, and if we do that, the Fed has less wiggle room, less policy space to cut interest rates, and so its hands are tied, so what can they do beyond going negative interest rates, which isn't very popular politically, and then there's a limit to QE. The Fed was wrestling with this, and what they came up with was this Flexible Average Inflation Targeting framework which allowed them to offset persistent periods of below 2% inflation to keep the average up at two and, by keeping the average up at two, it would keep the nominal interest rate up as well and provide more wiggle room for it. Pat, maybe I'll let you flesh that out a little bit in terms of the details of the framework.

Horan: Sure. As David said, the motivation was largely how to deal with below target inflation, below two. Prior to 2020, the Fed's goal was to try to hit 2% inflation each year, but they didn't take into account misses to that target, so bygones were bygones, if you missed one year, you didn't try to offset the following year, but because of the problems of undershooting persistently, the Fed was concerned it could be harmful to the Fed's credibility that it can't hit 2% inflation and then also lead to lower nominal interest rates and, as David said, less room to implement monetary policy in the event of a downturn, so then the idea was, okay, we'll make up for undershoots to bring the average back to two.

Horan: What David and I talk about in the paper, at least in the beginning, is that a lot of people were under the impression that the reverse would also be true, that while you say you're doing average inflation targeting now, so now that you've overshot, you're going to undershoot now to bring the average back down. The Fed has said over the past year, "No. No. No. Now our goal is just to go back down to two, so we undershoot from below, but not from above." David and I sift through a lot of FOMC statements, the long-term goals and strategy statement that they put out in 2020, as well as statements by top Fed officials, and they actually did clarify even then that the goal was to undershoot only from below, not from above, but we don't think it was very much at the forefront of discussions, so that's why a lot of people were confused, including a lot of prominent economists.

David and I sift through a lot of FOMC statements, the long-term goals and strategy statement that they put out in 2020, as well as statements by top Fed officials, and they actually did clarify even then that the goal was to undershoot only from below, not from above, but we don't think it was very much at the forefront of discussions, so that's why a lot of people were confused, including a lot of prominent economists.

Beckworth: I would add myself to that list. I, too, was confused. I, too, was thinking this would be more symmetric. I wrote a piece or two on this and, I think it's fair that many people were confused. Again, I'll throw myself out there. If you went to the previous framework the Fed had, so the flexible inflation target, just FIT versus FAIT, and you look at their statement on long-run goals and monetary policy strategy, so that's where they enumerated their new framework, in their statement from 2016 to 2019, they had that their framework would be a symmetric inflation goal, and they have those words symmetric inflation goal, and that they would be “concerned if inflation were running persistently above or below that objective.” Coming into this period of this new framework, if that's the thinking beforehand, it's easy to see why people would think, okay, that same symmetry would be applied to this makeup part of the new framework, but that wasn't the case.

Binder: Yeah. I remember, in the years before 2020, there was a lot of talk, including a lot of speeches by Fed officials, emphasizing that the inflation target was symmetric. I think it was because they had been undershooting the target, and some officials, including Charles Evans, were concerned that inflation expectations were actually becoming unanchored from below, that inflation expectations were below target, and they thought that part of the problem was that the public viewed the 2% target as a ceiling rather than as a symmetric target.

Binder: Part of that confusion, I think, could have come from the fact that what the ECB was targeting was inflation at or just below 2%. Theirs did have this kind of asymmetry in the way they worded it, and people might have just assumed either based on that or based on the fact that the Fed had been keeping inflation just below 2%, that 2% was actually more of a ceiling. I remember, in the 2010s, there was a lot of discussion about whether the target was a ceiling or not and then a lot of emphasis that it was not a ceiling.

Binder: In some ways, even though that, I think, has made it confusing for the public about this, with the AIT being asymmetric, it comes from the same idea because they used to be concerned that there was this asymmetry where inflation below 2% only was acceptable, and now they're emphasizing the other extreme, that inflation a little bit above 2% is acceptable, but if we go below, we're not going to accept that, we're going to make up for it. That is one important asymmetry in the FAIT framework that you talk about, but your paper actually talks about, I think, two different kinds of asymmetries, this one that we've been talking about on the inflation side and then a kind of asymmetry on the employment side, that there's an asymmetry in the way the Fed responds to negative versus positive supply shocks. Could you talk us through that asymmetry a little more? Because I think people are less likely to be familiar with that one.

The Important Asymmetries of FAIT

Beckworth: Let me read from the *Statement on Longer-Run Goals and Monetary Policy Strategy,* so the one that has the updated framework in it, and the part that we draw upon to illustrate this point, there's really two places, but the first key, big change, this is about two-thirds of the way down in the statement, it says, “The committee seeks over time to mitigate shortfalls of employment from the committee's assessment of its maximum level.” Previously, it would say, "deviations from the maximum level." This is shortfall. I think of Milton Friedman's plucking model. It's only from one side, from below. That was a huge change, and I do think a lot of people missed that.

Beckworth: Also, later in the statement, it says that should circumstances emerge where these two objectives, the inflation and the maximum employment, aren't in line, then they will reconsider how long it will take to normalize policy. That's referring to the case where you have a negative supply shock which doesn't cause output and inflation to move in the same direction. Demand shocks typically move in the same direction, supply shocks don't, and they say later, should we be in an environment like that, we will reconsider our path forward.

Horan: That's a good description. David mentioned flexible inflation targeting before, so even prior to adopting FAIT, the Fed had stated that in situations where they deem the employment and price stability goals to be in conflict, they would be flexible. Instances like that would namely be the negative supply shock example that David said where, suppose you had oil shock, for instance, that was temporarily raising prices and temporarily causing inflation to rise, but it would also be presumably hurting the employment goal and hurting the real economy, so you wouldn't want to tighten in that scenario.

Binder: Right. Is that really the definition of flexibility in some sense that the flexibility to look through supply shocks or does flexibility also have some other meaning?

Beckworth: I think, for the most part, that is a key part of it. The, I think, textbook understanding of inflation targeting up until this change was you look through or see through supply shocks, and you do that given you have credibility that you're going to keep inflation expectations anchored. I've discussed this with some previous guests on the show that, in some places, some emerging markets, they don't have that luxury. They can't see through supply shocks because it's going to further deteriorate already unanchored inflation expectations, but in advanced economies like the US and Europe, it's the case that inflation expectations were anchored and they had the luxury of seeing through it and just trying to respond to more trend changes in inflation driven by aggregate demand. I think that was a key part of that. I think one of the challenges we face today is that credibility being challenged, and that's why the Fed is responding.

Beckworth: Let me step back. The Fed is also responding, I think, to some demand-driven inflation, but maybe the ECB is a better example of this. In the ECB, it seems like that a lot of the inflation is supply-side driven, energy costs, and yet it is aggressively responding to it. I think the reason they're doing that is they're concerned about inflation fighting credibility. We're recording this and, today, the latest numbers can account for the inflation in the Eurozone, above 10%, so they're really worried about the expectations becoming unanchored there and are responding as a result.

Binder: Right. Just to make sure that I'm understanding this asymmetry, if you have an adverse aggregate supply shock, oil prices rise, that is going to increase inflation, also increase unemployment and, in the old FIT framework and in the new FAIT framework, if the Fed can recognize that that's a supply shock, they will try to look through it. Now, if you have a positive aggregate supply shock, so oil prices fall, it reduces inflation, reduces unemployment, then, in the old framework, they would look through that symmetrically the way they would look through an oil price increase, but, in the new framework, because that's not causing a shortfall of employment, they're not going to look through it. Is that right? They're going to say, okay, unemployment's down, that's not a problem in our objective function, and inflation is down, to the extent that we undershot our inflation target, we'll make up for that by allowing higher inflation down the road?

Beckworth: So, Carola, in the case of a positive supply-side shock, I think that may be correct on paper, but, in practice, even like in the previous framework, they often would take advantage of that. Now, I know that the point is though, if you're below target in this framework, they have to make it up regardless of what's driving it. As I'm thinking through this, I think the issue would be, would there be a scenario where you would have enough of a persistent positive supply shock to keep inflation low enough long enough to get that average down?

Beckworth: Now, of course, you could say we've seen that on the upside, right? We've seen persistent supply shocks on the upside. I guess, in my mind, it's a little more tricky to think through persistent dis-inflationary positive supply shocks other than maybe a permanent increase in the productivity growth rate that might lower inflation. In that case, they might have to tweak what they're doing and, yeah, bring it up. I guess that's a great question, so would they have to maybe even, in effect, change the inflation target in a situation like that? I think you're right, that is an implication of the current framework.

In my mind, it's a little more tricky to think through persistent dis-inflationary positive supply shocks other than maybe a permanent increase in the productivity growth rate that might lower inflation. In that case, they might have to tweak what they're doing.

Binder: David, I think that the first time that we met was at the 2019 Cato Annual Monetary Policy Conference. At that conference, Vice Chair Richard Clarida was talking about the Fed's framework review that was ongoing at that point, and he talked about the benefits of makeup strategies and indicated that that was something the Fed was really considering in its review, which it did indeed adopt. When I was listening and maybe when others were listening, they probably thought that he was referring to something like Ben Bernanke's temporary price level target proposal that Bernanke had been talking about and had written about it on the Brookings blog and other places. A temporary price level target I think has a lot in common with FAIT, so why do you think the Fed ended up choosing FAIT instead of the maybe even unwieldier acronym TPLT? What do you see as the main differences?

FAIT vs TPLT: Why the Former?

Beckworth: Well, I think they're very similar, and former Vice Chair Richard Clarida, as you mentioned, gave that talk. He had several speeches near the end of his time at the Fed where he said, "This is effectively a temporary price level target," but, yeah, name is different, and I think part of the reason might be just… you want to make an easy transition, you want to say you're still doing inflation targeting. Part of it is marketing. Maybe some of it is political. You want to make that easy transition to it, and I think there's probably some challenges in implementing, perfectly, a temporary price level target. Bernanke's motivation for doing a temporary price level target is very similar to what the Fed's motivation was, effective lower bound concerns. Something else that Bernanke mentions in his description of the temporary price level target is, you want it to be temporary because if you did a true price level target instead of a temporary price level target, you would have to actually reverse the price level increases that came from a negative supply-side shock.

Beckworth: One way to think about FAIT and temporary price level targeting as we've discussed is you make up for below-target inflation that's persistent, and the assumption there is, implicitly, most of that inflation will be coming from weak aggregate demand if you're going to be persistently below. The supply side issues we just talked about probably weren't wrestled with enough, but that's the idea. Once you go above it, the idea is you revert back to a traditional inflation target. The reason they stress that, I think, at least Bernanke stresses that, is that you don't want to be in a situation where the price level is up and you've got to tighten monetary policy and bring it down.

Beckworth: A good example I think would be 2008. 2008, the price level was taking off because of commodity prices, a similarly negative supply side shock. You wouldn't want, in that situation, to actually begin to raise interest rates and tighten monetary policy and bring it down, so Bernanke said, "Let's make a temporary price level target compromise." It has some of the benefits, the makeup side and, again, I think the concern was that we're going to be in a world of persistently hitting the lower bound and having low inflation and trying to make up for it. I do think, circling back to your question, FAIT is a version of the temporary price level target. I think maybe it was just an easier way to sell it because it still has IT in it, right? It's FAIT.

Binder: I think, with the temporary price level target, doesn't the temporary refer to, they'll make up for below-target inflation that occurs at the ELB or at the ZLB? The temporary is talking about when you do it. I think, with FAIT, it doesn't matter when the undershoot occurs. It's like, if we undershot inflation, even when we're not at the zero lower bound, we'll still make up for it. I think that might be the difference as far as I understand it. With TPLT, if you undershoot inflation, but you're not at the zero-lower bound, then you don't necessarily have to make up for it.

Horan: That could be true. It's not a hundred percent clear to me that it's that distinct. That sounds right, but I also get the impression from listening to lots of people at the Fed speak that they're likely to undershoot when they are at the ELB. When they're not at the ELB, then they should be hitting the target as they're supposed to be.

Binder: Although they undershot for a lot of the 2010s… Well, after 2015, there was a liftoff from the zero lower bound, but they continued undershooting. Right?

Beckworth: Well, Carola, I think that's a reasonable interpretation, what you just said about FAIT, but I think their intentions would be much more in the spirit of a temporary price level target, again, for the reason that, what would actually cause you to be persistently below your inflation target, and I think most of them, at least implicitly, are thinking it would be because of some negative demand shock, they're not thinking... which would put them at the lower bound. I think, implicitly, they're coming from, we're at the lower bound, but I think, yes, it could be definitely applied in that context because it wasn't very specified, and maybe that speaks to some of the issues with this. There's a lot of uncertainty, a lack of clarity in the details of this framework. There's room for improvement, and maybe we'll talk about later the next stage of the framework in 2024, 2025.

One way to think about FAIT and temporary price level targeting as we've discussed is you make up for below-target inflation that's persistent, and the assumption there is, implicitly, most of that inflation will be coming from weak aggregate demand if you're going to be persistently below. The supply side issues we just talked about probably weren't wrestled with enough, but that's the idea. Once you go above it, the idea is you revert back to a traditional inflation target.

Horan: Carola, you had asked, "So why call it flexible average inflation targeting and not temporary price level targeting if they're basically the same thing?" Maybe they're not. Maybe they're somewhat different like you just suggested, and David said, "Well, maybe it's good marketing to say we're still doing inflation targeting." That's one argument for it. A counter-argument would be that flexible average inflation targeting has a very loose definition of the word "average" though. If you're saying you're going to make up from below but not above, you're not truly doing average inflation targeting then, which creates a lot of confusion for a lot of people, and so it could be an argument for temporary price level targeting instead even though that might be more complicated. To me at least, it's clear what you're trying to do. You're making up from below and then you switch to 2% when you're not making up from below.

Binder: Yeah. There was a great headline shortly after the announcement that said, "Fed officials do their own math on average inflation." I think that gets to the point of, well, people didn't know how many years back are they going to look when they try to compute this average and how long into the future are they going to take to try to bring inflation back down to average.

Beckworth: I remember, when inflation was beginning to take off in 2021, I had numerous exchanges, debates on Twitter about what did that average actually mean, how do you measure that? Again, I think the takeaway is, that makeup, that average is when the average is below 2%, not above. It's very, very asymmetric, and that's the hardest part to wrap our minds around.

Binder: Yeah. I noticed in your paper that you cite a Financial Times article by Mohamed El-Erian as an example of another commentator who was dissatisfied with the FAIT framework. I looked up that article. It came out in May 2021, so when the framework was much newer and when this debate about whether inflation was transitory was just starting to rage, and he wrote, "The Fed has no operational choice, but to stick to its transitory narrative given its recent adoption of a new monetary framework. This framework has wired in a delayed outcome-based approach replacing the more preemptive inclination associated with the traditional forecast-based one. As such, the world's most powerful central bank is publicly committed to having to wait for several months of actual upward inflation deviations before responding."

Binder: I think what he's saying is that the Fed used to respond to expected future inflation. If you think of a Taylor rule, we usually write expected inflation there, but the new framework no longer includes a forward-looking reaction function. Instead, they actually wait to see what happens with inflation before responding. Do you agree with this interpretation of the difference between the old framework and the new framework? If so, where in the amended statement can we actually see that the new framework is less forward-looking?

Forward-looking Differences Between the Fed’s Old and New Framework

Beckworth: Well, two things, let me throw out the FOMC's interpretation of that, and then I'll give what I think. The FOMC definitely was saying that. The FOMC was saying, "We need to see the white of the eyes of the inflation. We need to actually see inflation," and then, once inflation got really high, they actually reverted back to expected inflation. The FOMC actually changed its tune on this. They were first saying what Mohamed El-Erian was claiming. They want to actually see inflation and then, when inflation began to take off, they began to invoke, “well, expectations are still relatively anchored.” I think you can definitely see that or make that interpretation based on this idea that you got to get the average up to a certain point. How do you know you're there? That's going to be a backward looking, "I have to actually see inflation above 2% for some time".

Beckworth: The FOMC statement from September 2020 to November 2021, let me read it here. This is their conditions for ending makeup policy. They said, "Labor market conditions have to reach levels consistent with the committee's assessment of maximum employment and inflation has risen to 2% and is on track to moderately exceed 2% for some time." That it has risen to 2% and is on track to exceed it, that implies they actually have to see it happening in real time. Now, they dropped that in November 2021, so I think, early on, the FOMC was stressing that and then, once it became the problem, they began to stress more expected inflation as the way to look at it.

Binder: Yeah, I think that's right. I remember back in maybe around 2015 or shortly before the Fed was about to start its liftoff from the zero-lower bound, there was a lot of pressure on the Fed to say, "Look, we know unemployment is really low, so you expect that inflation is going to start rising, but now is not the time to start raising rates now. Let's actually wait until we see inflation start to rise before you lift off from the zero-lower bound to give the labor market a better chance to recover," and this was a big debate at the time. Should the Fed respond to its expected future inflation or should it respond to the inflation that it actually sees? I think, back then, people were using "respond to the inflation you actually see" as a way to pressure the Fed to keep policy looser for longer.

Binder: We sort of see, over time, differences in whether people want the Fed to respond to expected inflation or respond to current inflation. That's, of course, going to be a big debate now as inflation likely starts falling. Well, when should the Fed stop its rate hikes or when should it ease up on the rate hikes? Should it wait until inflation actually gets back to target? I don't think it's too likely to do that. I think it just wants to see, well, we can forecast into the future that inflation is going to start to fall, so we can ease up on the hikes right now. This is a question with FAIT or with any other framework, really. Okay, so let's get a little more into your motivations for writing this paper, *The Fate of FAIT.* It's about trying to salvage the FAIT framework as you say. Why does FAIT actually need salvaging? You've alluded to some of the issues with it, but why don't you lay them out and say why you think we need to do something to change this framework?

Salvaging the FAIT Framework

Horan: That's a good question. Why do we say salvage it as opposed to scrape it and do something else or just keep it as it is? I think the argument against “keep it as is” is that it's created a lot of confusion and it's come under intense scrutiny, so you have very prominent economists like Larry Summers saying, "FAIT has not worked out well. We should scrap it and go back to the old system." That's the argument against keeping it as it is.

Horan: The argument for salvaging [FAIT], though, is that we think there are some good parts to this. We think the idea of makeup policy in general is a pretty good idea. Just even thinking back to economics in the case for why you want price stability, the public wants to have a good idea what prices are going to be going into the future so they can make plans today, so that's both households and firms, they make plans and they want to have a fairly good idea what the price level is going to be going to the future. You can think of a makeup policy as basically being a way to better keep the price level on track. However, that would imply that you need to be more symmetric in your approach. You can't just be from below. It should also be from above.

The argument for salvaging [FAIT], though, is that we think there are some good parts to this. We think the idea of makeup policy in general is a pretty good idea.

Horan: We also like the part that you don't want to tighten in response to negative supply shocks because, again, you'd be tightening in response to something that is presumably transitory and would just hurt the real economy more. However, again, you're asymmetric, so we want to try to make it symmetric, and that's where we bring our idea of the Fed should embrace what we've been talking about for a while, and that's nominal GDP level targeting, and you might say, "Oh, but isn't that something different? That's radical, so, no, you're not salvaging it. You're doing something new." We argue that you can accomplish both goals at once. You can have average inflation and you can also be doing nominal GDP level targeting. By that, I mean, for those who are not familiar, that's the idea that the Fed would try to stabilize the sum of spending over time. It would make up for deviations either way.

Beckworth: This is an idea actually that we saw Michael Woodford make in a 2013 paper. He argued, these are his words, “a commitment to a nominal GDP level path is completely consistent with a commitment to a medium-term inflation target.” If we take seriously this idea that central bankers should see through short-term inflation spikes due to supply shocks, then we shouldn't be worried about month-to-month, maybe even year-to-year limits in inflation. We're worried about where it's going over the medium run, so two, three, four years out.

Beckworth: If we are stabilizing inflation over that horizon, one way to do that would be to stabilize, basically, the growth path of aggregate demand. When we say nominal GDP, we mean aggregate spending, aggregate demand, and so Michael Woodford makes this argument, Bob Hetzel, who used to be with the Richmond Federal Reserve, makes a similar argument that, if you stabilize nominal GDP or aggregate demand over the medium term, you're going to get something similar for the inflation target. We think this is a way to preserve FAIT with a little tweak to it. In fact, what we do is we call it FAIT-N to represent the nominal GDP weight on it. In other words, the Fed can still come out and say, hey, we are doing inflation targeting, but we're doing it in a manner that preserves it over the medium term.

Binder: Okay. I think it's pretty easy to see how average inflation targeting, if you make it symmetric, is like price-level targeting because you have a path for the price level that the Fed wants to target. If you have inflation below, so the price level is growing too slowly, you fall below the path, and they make up for it and they bring the price level back to the path, and the same if you have inflation too high, you go above the price level path, you bring it back down, so as long as you keep average inflation constant, you're going to be moving along a price level path that has a constant growth rate. I think that part is pretty clear.

Binder: Then, if you go from price-level targeting versus nominal GDP level targeting, well, the growth rate of nominal GDP is the growth rate of real GDP plus inflation. I think, if the growth rate of real GDP doesn't change much over time, then it's pretty easy to see through the equivalence in the medium run between price-level targeting and nominal GDP level targeting. I think where people might think that they would be different is when you have the real growth rate change over time, say, because of demographic changes or the economy becomes less productive or more productive for some reason.

Binder: How would that work? Say, you expect that the real GDP growth rate in the long run is 3% and you want inflation in the long run to be 2%, so you pick an NGDP level target where NGDP is growing 5% a year, that's 2% plus 3%. If for some reason the real growth rate changed to something like 2%, would you have to change that NGDP target to reflect that, or would that mean that you were then going to have higher average inflation over time?

The Nuts and Bolts of an NGDP Level Target

Horan: The central bank would have to pick. It can either adjust its nominal GDP target to be consistent with the inflation target it really wants, 2%, or it can stick with the nominal GDP target, but then you have a higher trend inflation rate, and there's arguments for either way.

Beckworth: Yeah, and just to be clear, how would you actually implement a change in the nominal GDP target? It would have to be gradual over time because potential real GDP trend is something we don't observe in real time, and so you'd have to make adjustments. At Mercatus, we have this neutral level of nominal GDP, and this is one way to incorporate those changes. It looks at the consensus forecast of where potential real GDP is going, plus the expected inflation growth path over the long run. In some ways, it's capturing at least the markets or the consensus expectations. The Fed could do a gradual update like that, but there are trade-offs. You could adjust the target gradually, slowly. To be clear, historically, you don't see big swings in potential real GDP. They're usually relatively mild. It may not be a big deal in practice…

Binder: I mean, I would think that, for the public, it would be much easier for them to just understand and to find it credible for the Fed to say NGDP is going to be growing 5% per year than to have them make adjustments or even to say we're going to use this measure based on consensus forecasts. They might think that that's subject to being manipulated or it might just make it harder for them to plan. I would think it would be a lot easier to communicate if you just had a constant number, have it be like a round number like 4% or 5% rather than 4.8 or something. I mean, can we still call it FAIT-N? Does it still have this kind of equivalency to average-inflation targeting, or does it become something actually quite distinct?

Beckworth: Well, I think it does. You're asking, if you change the nominal GDP target, is it still a FAIT-N? Is that what you're asking?

Binder: No. I'm asking, if you keep it constant, if you keep the nominal GDP… NGDP level target, just a level that's growing at a constant rate, is it still equivalent to AIT even though you have this possibility of real GDP growth rate changing over time?

Beckworth: It would be an AIT whose average target is updated over time. That's what it would be. It would be a version of an flexible average inflation target, but you would have to update it.

Horan: Again, the real GDP growth or the potential real GDP growth rate doesn't change very much over short intervals, so this would be a slowly changing average rather than a rapidly changing average. One argument for keeping it the same, this is one argument you can make, is that if your real GDP growth rate trends down, now you have a higher new trend inflation rate. If you're keeping the NGDP target the same, arguably, Congress or other policymakers should do more to boost productivity growth to bring that inflation rate down. The Fed can just say, "Hey, we're just keeping spending on the stable path, which is our goal if you want lower inflation. Now, it's your return to help make the economy more productive."

Beckworth: Let me add to that also another reason to keep the target fairly stable, is that people make their financial plans based on some kind of implicit forecast of where their incomes, their nominal incomes, their dollar incomes are going. There's a financial stability argument for not changing the nominal GDP target in addition to the simplicity one, the credibility one. I would note that. Finally, let me just circle back, Carola, to a point you made earlier. If we do a nominal GDP target and if potential real GDP changes so that the price level permanently changes or the inflation rate permanently changes, the argument for allowing that is that the fundamentals of the economy have changed, scarcity has changed or abundance has changed.

Beckworth: If the real economy has shifted such that, say, we're poor because of negative supply shocks, one way to let that manifest is, things are now more expensive, we simply can't produce what we used to or vice versa if now we're more productive, things are cheaper, so we're allowing the price mechanism to reflect that. A price level target would not do that. A price level target would force all those changes back into contraction or expansion in real GDP. What we're saying is, let those relative price changes take place given changes in the real economy and just keep the Fed's eyes on total dollar spending or, equivalently, total nominal income.

Binder: Yeah. That actually gives the benefit that inflation becomes more the way people already understand inflation as inflation is the bad thing, and so when you have these bad times when real growth is low, inflation is high, and people already take high inflation as a sign that the economy is bad and low inflation is a sign that the economy is good. As it currently stands, they're not always right when they do that, but with NGDP level targeting, they would be right when they use that rule of thumb, which I think is a nice feature of NGDP targeting that's not always emphasized.

What we're saying is, let those relative price changes take place given changes in the real economy and just keep the Fed's eyes on total dollar spending or, equivalently, total nominal income.

Beckworth: That's really neat. I learned that from you. We did our previous podcast together, and you have this in your research. You show that people do confuse bad times with inflation, and it really resonated because, if we've been following the news, you know that a lot of people have said we're in a recession already even though it's hard to define that with the GDP numbers, labor markets are still growing robustly, but your research really helped me understand this, that people associate the two, recession and inflation, so very interesting.

Binder: One weakness of FAIT that you point out is that it relies too heavily on discretion. I think discretion and flexibility are not exactly the same thing. We talked about flexibility being more a reference to how they respond to supply shocks, but do you think that FAIT gives policymakers more discretion than FIT, and how would that compare to the amount of discretion policymakers would or could have under NGDP level targeting?

The Discretion of FAIT vs NGDPLT

Horan: I don't think flexibility and discretion are inherently the same thing. You could distinguish between the two. I think, frequently, when you have these flexible regimes, they do end up entailing a lot of discretion. When it comes to seeing through the supply shocks, it's also hard to see in real time, is this increase in prices due to a temporary adverse supply shock or no? I think, to some degree, policymakers then need to require some discretion to figure out what they're responding to.

Horan: I would say that FAIT entails more discretion than the previous framework. I don't think it necessarily has to be that way, but I think it does and that the FOMC retains the discretion to determine how much makeup policy is enough. There's no strict time horizon for what constitutes an average. There's a good argument for not wanting to have a strict average, and that is that, let's say, if you had a strict average timeframe and you made up and now you said, okay, now it's time to then go below to bring the average back to where we want it to be to satisfy the timeframe. If you run into an adverse supply shock like an oil shock, just when it's time to do that, then you're in a bind. That's an argument against having a strict timeframe for your average, but there's also a good argument for nominal GDP targeting because, by construction, you're not trying to respond to supply shocks.

Beckworth: Yeah. I would add to that also the definition of what a shortfall for maximum employment means. It's not defined, even in their statement. Let me read again from their Statement on Longer-Run Goals and Monetary Policy Strategy. This is what they say about the maximum level of employment. They say, "The maximum level of employment is a broad-based and an inclusive goal that is not directly measurable and changes over time owing largely to nonmonetary factors that affect the structure and dynamics of the labor market." They acknowledge that it's hard to put a finger on exactly what it is in real time, and then you add on top of that that you're only going to look at shortfalls from it. I just think that it adds a lot of confusion.

Beckworth: Going back to your question on flexibility versus discretion, the way I think about it would be, take a Taylor rule that prescribes... That Taylor rule gives a central bank flexibility, but it also makes it systematically respond in certain prescribed ways to changes in inflation and changes in output, and discretion would be having a big error term on that Taylor rule where the central bank is maybe going beyond the rule or outside the rule, and I think there's a lot of uncertainty. Circling back to nominal GDP, again, we keep plugging this, there is a lot of things we don't know in real time. I mean, Josh Hendrickson and I have a paper in JMCB in 2020 about even a Taylor rule. A Taylor rule has embedded in it this idea of an output gap, but we don't know in real time what the output gap is, and that can create more noise, more uncertainty, so keep it simple, keep it to some stable growth path and you avoid some of those problems. I think a simple nominal GDP target does avoid some of the temptation and problems created by trying to divine in real time with a Taylor rule where we should be.

I think a simple nominal GDP target does avoid some of the temptation and problems created by trying to divine in real time with a Taylor rule where we should be.

Binder: Do you imagine NGDP level targeting or FAIT-N implemented with some kind of rule where it was pretty clear to the public how quickly the Fed would get NGDP back to its target and what exactly its response would be, or do you think it would still be implemented with a similar amount of discretion as the Fed has had in recent years? Would it help if the Fed published the kind of rule it was going to use?

Horan: I think it would help if they published an explicit rule. In the paper, we show several potential rules about how you could go about doing nominal GDP level targeting. So, we provide several Taylor rules where the central bank sets its target interest rate in response to the estimated neutral rate plus our NGDP gap term. You could do just that or you could do a forecasted NGDP gap at a certain number of periods out or you could do a hybrid rule where you add in the current gap and you put some weight on it plus some weight times the forecasted gap, and then we also show some inertial rules. Inertial rules are rules that are meant to adjust the target interest rate more slowly if the central bank does not want to adjust its target very rapidly for whatever reason. I'm pro published rules-

Beckworth: Carola, let me go back to your question about discretion. I think what you're getting at is, whenever you do makeup policy, there is this potential for a lot of discretion. To be fair, this is a critique of any kind of level target, any kind of makeup policy. Whether it's a price level target, a nominal GDP level target, they both require some catch-up period. There's debate. Even people who love nominal GDP targeting, some of them want to do nominal GDP growth rate targeting, and then many of us want to do level targeting. One of the reasons people prefer the growth rate version is because there is discretion involved.

Beckworth: Bennett McCallum was a famous proponent of nominal GDP targeting going back to the 1980s, and he didn't like the idea of level targeting for this reason, and so I think it's a fair critique. Any kind of makeup policy is going to empower the central bank to make up, and how they do it could be conditional on what's happening. It's not clear that there's an easy answer. However, what I would come back with is my hope is that, if this rule were credible, if it were believed by the public, then the public would do the heavy lifting in terms of the makeup policy. If you believe the central bank was committed to maintaining stable aggregate spending, then the public would adjust, you'd see velocity adjust when you have shocks, and so hopefully the credibility by itself would do the heavy lifting, would do the makeup policy, and you wouldn't have to rely on the discretion of the Fed to do it for you.

Binder: What about something like what Scott Sumner has proposed where the Fed would use an NGDP futures market? I think that would basically remove the need for discretion. Did you think about something like that in your paper, or do you think it's too unlikely, too far from the status quo?

Horan: We briefly mention it. However, it is a radical proposal in the sense that it's very different from how central banks operate currently. It's an interesting idea. For the purposes of this paper, we didn't bring it up. Alternatively, there are other ways to potentially do nominal GDP targeting. The McCallum rule, David mentioned Bennett McCallum, the McCallum rule is a feedback rule for the monetary base in response to changes in nominal GDP, but then we'd have to switch to base rules, and central banks have not gone in that direction. They're more interested in interest rate rules.

Beckworth: It'd be very difficult with the interest on reserves, but, yeah, Scott's rule is fascinating, a nominal GDP futures target, and this might be viewed as a baby step in that direction.

Binder: Right. If the Fed were to decide they want to implement what you're calling FAIT-N, what would they actually have to do to make that change? Do you think they would have to go through another framework review? Would they have to amend the statement on longer-run goals? Would they have to make sure Congress is on board, or what would that whole process look like?

The Process of Adopting FAIT-N

Beckworth: Well, I think they could do it themselves without going to Congress. Part of the reason we want to call it FAIT-N is to say, look, we're still doing inflation targeting, so make everyone happy in Congress and the body politic. They're not radically changing things. Look, in your own work, Carola, you've noted, for example, from the post-great financial crisis up until this period, it looked like they effectively were doing a version of nominal GDP targeting. If you look at nominal GDP, it was relatively stable, so we could say, look, we're just formalizing what we have already been doing.

Beckworth: Yes, you would have to make changes to the statement on longer-run goals, the monetary policy strategy, you'd have to adjust the SEP and maybe add some additional metrics to it and be explicit about what you're doing. It would be a framework review that would be required, and we have one coming up in 2024, 2025. In fact, one of the motivations for the paper, in addition to everything we've discussed, is we want nominal GDP targeting to be a part of that conversation. Our hope is that this paper will generate some increased conversation and possibly lend itself to taking it more seriously this time than last time. Last time, it was briefly considered. Lars Svensson had a paper as part of the review. I know there were some other internal papers at the FOMC, the Board of Governors, that was done on the different approaches, but we're hoping the nominal GDP approach will get more time on the floor to be considered on the next review.

Horan: I want to make a point. Carola, you asked about how would the Fed go about switching to this new system. Just a practical point, one argument for negative interest rates would be that, to implement this kind of framework, if you're doing a Taylor rule, the interest rate that it prescribes does get negative at times, so in 2008, 2009, and again in 2020. This is also true of the traditional Taylor rule. That's an argument for the negative interest rates. If you truly want to follow the rule, you have to go negative, unless you have some other means of setting policy like a base rule or a futures market rule, but, again, we're not doing that right now.

Horan: In the paper, we show what this NGDP gap rule to do FAIT-N would look like in the great recession period and then more recently in 2021, 2022, and we see that, had a forward-looking NGDP gap rule been used in 2008, the FOMC should have cut the federal funds rate more quickly than it did in reality, and then, more recently, we show that such a rule would've prescribed raising the rate before the Fed actually did so in reality. It would've called for a more expansionary policy earlier on in 2008, potentially mitigating the great recession and then, more recently, would've called for a more contractionary policy which plausibly could have mitigated the inflation surge we've had.

Binder: Wouldn't a standard Taylor rule have done both of those things also, implied looser policy sooner in 2008 and-

Beckworth: Definitely in 2021 Taylor rules… in fact, David Papell and some others have some neat papers showing this, but 2008 was a little trickier because inflation… if you look at its headline CPI, it was relatively high. I mean, it's not too far below where it is today. If you look at the FOMC speeches and statements, they were very hawkish all the way up till mid-2008 because of high inflation. I think that's both they're worried about the headline number, but also that it's being fed into these Taylor rules. I mean, how you define it... You mentioned it earlier. The Taylor rule, where you put expected inflation in, I think that, yeah, maybe if you had a medium long-run inflation expectation term in your Taylor rule, maybe it would be very similar to a nominal GDP rule, but if you use a simple Taylor rule that looks at current inflation, then you might make a mistake in a large negative supply shock environment.

Binder: I guess you told me that the next framework review is in 2024, 2025. Hopefully and probably inflation will be back down to target by then, which I think would help with any kind of change in the framework. I think changing the framework right now with inflation so far above target would probably not be the best idea. It might come across as opportunistic or they're just changing strategies in the middle of this high-inflation episode would not give off the best appearance. Do you think that 2024, 2025 is going to be too soon? Does it matter what conditions are like at that time for how likely it is that they actually go about changing the framework?

Horan: Well, again, we're arguing that this would salvage FAIT by modifying it, not scrapping it because we agree that the Fed tends to be small-C conservative. They don't like to make dramatic changes unless really absolutely necessary. Again, that's an argument for salvaging it rather than scrapping it and just calling it nominal GDP level targeting. I don't want to make predictions about what 2024 and 2025 will look like because I think people have been pretty bad at forecasting the past few years. Hopefully, things are a little bit calmer and they can make a decision without panicking about high inflation, or hopefully not recession, too. Yeah, I think it's reasonable how a lot of people may have different perspectives on how soon is too soon when it comes to major changes.

The Fed tends to be small-C conservative. They don't like to make dramatic changes unless really absolutely necessary.... [and] that's an argument for salvaging [FAIT] rather than scrapping it and just calling it nominal GDP level targeting.

Beckworth: Well, my hope is that the Fed will look back at this experience and see two big things. One, how do you deal with environments where you have inflation, high inflation that's in part driven by negative supply shocks. Look to Europe, even more so of a problem, and nominal GDP targeting is a meaningful way to deal with that, and then, secondly, to rely on forecasts. Our approach is you have to look at forecasts of nominal GDP, not look in the rear view mirror. I think this period has been incredibly useful in telling that a traditional inflation-target approach has some severe challenges when you have these big supply shocks, so I think it will be a great opportunity to discuss this.

Beckworth: Another option that has been discussed before the pandemic, and now I'm seeing more people talk about it, is why not settle on a higher inflation target, why not just set it at 3%? I think, the point you raised earlier, it would undermine the credibility. If the Fed in the next framework review would say, hey, let's just go for 3 to 4% inflation, I think it would look awfully convenient and would definitely have an impact on its credibility. It opens the door for something like nominal GDP targeting, so, I mean, in some ways, if there's a silver lining to this whole mess we've been through, I think it's setting the stage for a robust discussion of a nominal GDP targeting-like framework.

Binder: Yeah. I agree. I think something like increasing the inflation target to 3 or 4%, I think that actually could provoke Congress to say, no, listen, we've given you a price stability mandate. 3% inflation or 4% inflation, that implies the price level is going to double every 25 or 20 years or whatever it is, and that's not price stability. I think they would put themselves at a lot more political risk raising the inflation target. I think the easiest thing would be for them to say, "Okay, we've seen the cost of high inflation. We've seen how it hurts families. We are changing the symmetry here. We will make up for past overshoots as well as past undershoots."

Binder: That would be the smallest change they could do. They justify it easily. I don't think it would upset Congress because it's actually more committed to their price stability mandate, but maybe without changing how they describe deviations versus shortfalls from employment. I imagine you would think that that would still be a favorable change even though it wouldn't go all the way to NGDP level targeting if they were to just make the framework symmetrical with regards to undershoots versus overshoots of inflation. To me, that seems more likely and maybe a good thing.

Beckworth: That would be an easy next step for sure, but I would want the Fed to be clear that these would be corrections when the inflation is driven by demand, because you can imagine a scenario... If we're in Europe and you applied that kind of rule in Europe, you would be tightening when, in fact, it may be the case that the inflation has been driven by the supply side. The danger is, if you're too narrowly focused just on inflation, you might again get the wrong prescription. I think, again, going back to what we've talked about earlier, that's implicit in this that if the average over time is deviating, all of that typically happens from aggregate demand pressures, not from aggregate supply shocks. I think you can get there, but, yeah, you're probably right, that's probably the easiest next step.

Binder: Or something like make up for persistent overshoots or persistent-

Beckworth: Persistent, yeah, that would be great.

Binder: Alright. Thank you so much for talking about this paper. I hope all of the listeners will go look it up, read it. I think they can find it on the Mercatus website. Great chatting with you then, and I will talk to you soon.

Beckworth: Thank you, Carola.

Horan: Thanks so much, Carola.

Check out our new Macro Musings merch here, and use the promo code NGDP for 10% off!

Photo by Win McNamee via Getty Images