Carola Binder on Political Pressure and the Twin Deficits of Central Banking

Carola Binder is an assistant professor of economics at Haverford College and is an associate editor of the Journal of Money, Credit, and Central Banking. Carola is also a member of the CEPR Research and Policy Network on Central Bank Communication, and joins the show today to discuss her work on central banking and populism. David and Carola also discuss the link between central bank credibility and popularity, the twin deficits of central banking, and why NGDP targeting could be an easy transition point from current inflation targeting regimes.

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: Our guest today is Carola Binder. Carola is an assistant professor of economics at Haverford College and is an associate editor of the Journal of Money, Credit and Banking. She's also a member of the CEPR Research and Policy Network on Central Bank Communication. Carola joins us today to discuss her work on central banking and populism. Carola, welcome to the show.

Carola Binder: Thanks for having me. It's really great to be here.

Beckworth: Great to have you on. Now we crossed paths recently at a conference and you had a great paper you presented and I'm excited to get you on the show to talk about it. Before we jump into it, I want to learn, how did you get into economics? What motivated you to pick this career path?

Binder: Sure. So, I did my undergraduate studies at Georgia Tech. I was a math major there. And sort of by chance I took an internship the summer after my sophomore year at the Georgia Senate Budget and Evaluations office, which is like the state budget office there. So this was roughly in 2008 or 2009 I think. So we were in the midst of the recession and the budget crisis. I found the work really interesting. I actually ended up staying on for a full year rather than just the summer and just became more interested in policy that way.

I honestly applied to economics PhD programs on something of a whim, I don't think I really knew what I was getting myself into, just that I liked and was good at math and was kind of interested in policy. So I applied for some econ PhD programs and ended up attending the program at Berkeley.

Beckworth: Yeah. So how did you make the decision to go into macroeconomics versus some kind of applied micro field?

Binder: That decision was largely inspired by the Romers’ class at Berkeley. So David Romer and Christina Romer-

Beckworth: Oh, interesting.

Binder: They teach a really great class on US macroeconomic history. And it was probably a combination of just how interesting that class was and their approach to macroeconomics, as well as the timing that I was in graduate school. I started at Berkeley in 2010. And, of course, the macro economy was just the most interesting topic at the time. There was so much going on in central banking and monetary policy that it just was what interested me the most, even though I originally probably thought I'd be a theorist since I had the math background.

Beckworth: Yeah. So I guess Christina Romer must have just come back from serving in the White House during that time. Is that right?

Binder: Yes, she had just come back.

Beckworth: Okay. So you were seeing kind of policymakers firsthand teaching your class, you had witnessed the Great Recession, you'd seen it back in Georgia, at the state level and then at the macro level. So very fertile time to be thinking about these issues, and so you went into macro. And you've done some really interesting work since then, since you've been out. I want to talk about a paper, several papers you've done on central banking and populism. You have one paper that's titled *Political Pressure on Central Banks*, and we'll get to it, but an interesting collection of data you've done, 118 countries over the past decade, looked at political pressure on central banks.

It's a very timely topic for those of us who follow this topic closely. As we all know, in many advanced economies, populism has been rising here in the US. Of course, we've seen this interaction between President Trump and Fed chair Jay Powell that often happens on Twitter. But a lot of interesting debates going on about this, and I happened to hear you present a paper that also touched on this, another paper of yours. This paper was titled *Nominal GDP Targeting and the Public.*

Now of course, with nominal GDP targeting and the title, all listeners will know I got excited just off the bat. But you had a really interesting story to tell and I want to work our way through this. And this is really a story about how central banks can deal with populism, kind of how to nip it in the bud and be proactive as opposed to waiting for populism to kind of force changes upon them.

And so, you have this great quote you begin with, and it deals with nominal GDP targeting. I'm going to read it, it's from Ben Bernanke. So in 2011, the Fed was considering different approaches to monetary policy. And as we know, they ended up with inflation targeting which they adopted in 2012. But you had this great quote that comes from his book that surrounds the discussion of these different approaches, and nominal GDP targeting was one of them, and in particular, this quote surrounds that discussion. So I'm going to read it from your paper here.

So, this is Ben Bernanke talking, recalling his discussion and he says, "We consider the theoretical benefits of the approach but also whether it was desirable or even feasible to switch to a new framework at the time of great economic uncertainty. After lengthy discussion, the committee firmly rejected the idea. I had been intrigued by the approach at first, but came to share my colleagues' reservations about introducing at that time. Nominal GDP targeting is complicated and will be very difficult to communicate to the public as well as to Congress which would have to be consulted."

And you go through your introduction, and later you say, "In 2011, the FOMC thought that the time was not right to switch to a new monetary policy framework. In this paper, I suggest the time is better now. I argue that the status quo is so unpopular and precarious that a new target would do more good than harm for central bank credibility. In the current context, the case for nominal GDP level targeting is especially strong."

Okay, so that's kind of a setup for this paper and the discussion of populism and central banking. But let's work our way through the paper. And one of the first points you make in the paper is that central banks are unpopular by design. What do you mean by that?

Central Bank Unpopularity

Binder: Right. So this is based on Kenneth Rogoff's 1985 paper that kind of talks about how we might deal with the inflationary bias in monetary policy or deal with the time inconsistency problem in monetary policy. And that's by appointing a central banker who is both independent and conservative. But conservative here means that the central banker is more inflation averse than the government or than the general public.

So, what that implies is that, the central bank is going to be pursuing policies that are politically unpopular in the short run, but help keep inflation down in the long run. So even if the policies are beneficial in the long run, they're always going to be unpopular in the short run. So, the central bank is always really going to be subject to pressure for easier, more inflationary monetary policy.

Beckworth:  Yeah, so central banks are unpopular by design. And I believe in that discussion somewhere you talk about survey Alan Blinder did and he found that for most central bankers, credibility was the same or mounted to a low inflation dedication or a commitment to low inflation. And that credibility is important to them but it's also very unpopular.

Binder: Right. And central bankers and people discussing central banks tend to talk a lot about central bank credibility, but not talk so much about central bank popularity, I think partly because we think of central bankers as these technocrats who don't care about, don't need to care about their popularity. But the point I try to make in the paper is that to at least some extent, they do need to care about their popularity because politicians can always ultimately change central banks even if central banks have some degree of independence.

We think of central bankers as these technocrats who don't care about, don't need to care about their popularity. But the point I try to make in the paper is that to at least some extent, they do need to care about their popularity because politicians can always ultimately change central banks even if central banks have some degree of independence.

Beckworth: I completely identify with that because it's easy to get caught up on a research project, you just kind of assume or take for granted that central bank is independent. But here in the US, they're a creation of Congress and Congress can change what it creates. And again, as we mentioned earlier, President Trump has been very vocal about his unease with Fed policy, kind of another symptom of this link that you talked about between credibility and popularity that we need to take seriously.

Along those lines, you also discuss an idea called the twin deficit. So Andy Haldane’s twin deficits. Can you explain that to our listeners?

The Twin Deficits of Central Banking

Binder: Yeah. So Andy Haldane came up with the phrase twin deficits to refer to deficits of understanding and deficits of trust in central banks. The idea is that the public lacks an understanding or a very thorough understanding of what it is that central banks do. And then they also don't trust the central bank. And he knows that these two deficits are, of course, very closely related. And a lot of my own work has documented evidence of these deficits.

So I've looked a lot at consumer survey data and even conducted some of my own surveys that show that consumers don't seem to have a great idea of who works at the Fed, what the Fed's goals are, what inflation has been recently. And they also, therefore, lack much trust in the Fed. Others have shown this for other central banks as well.

Beckworth: Right. So there's a deficit of trust, a deficit of understanding. I want to just cite some of the numbers you mentioned in the paper which references other work and some of your own work. But you mentioned in terms of deficit of understanding, only one fourth of those surveyed knew of the Fed's two percent inflation target, is that right?

Binder: Yeah, this was in a survey that I conducted online with one of my former students, Alex Rodrigue. We did a survey on Amazon Mechanical Turk, and one of the questions asked people if they knew the Fed's inflation target. Half of them said that they did but then only half of those actually could give us the correct number.

Beckworth: All right. So what does that tell us about the deficit of understanding that most people don't know or are aware or maybe care about the Fed's inflation target?

Binder: Yeah, that they're not aware of it. I don't know that that's automatically a bad thing. So, I think a lot of times the research on central bank communication is about how can central banks get more people to know more about central banks. I think most people don't like really spending their limited time and energy reading about and thinking about central banks. So, I think the ideal monetary policy just keeps the economy stable enough that most people don't even necessarily have to know about what the Fed's target is. They can just rely on the economy being stable.

I think most people don't like really spending their limited time and energy reading about and thinking about central banks. So, I think the ideal monetary policy just keeps the economy stable enough that most people don't even necessarily have to know about what the Fed's target is.

But at least the theory of inflation targeting is that it works through expectations, and by giving this explicit target, it should anchor expectations around that target. And for most people, it doesn't seem to have done that.

Beckworth: Yeah, that's a fair point. This may be a sign of an accomplishment, a good thing.

Binder: Right, right. Yeah. That's the point I want to make, is that we don't necessarily need to say that that's a bad thing. But if people didn't know the inflation target but they all felt confident that inflation was just low and stable, but they just didn't know that the number was two, that would be good. But a lot of people will give inflation expectations that are like 10 percent or something like that. So, in that case, I don't know that that's a good sign.

Beckworth: Yeah. So they don't know because they don't need to know. You also mentioned in there that not many people were familiar with Jay Powell or the head of central banks. Is that right?

Binder: Yeah. At the time of our survey, it was Janet Yellen that was the Fed chair. I think we gave them a multiple choice question with three possible options, and two thirds correctly picked Janet Yellen as the Fed chair.

Beckworth: Alright. And again, this could be a good sign, maybe it's a great thing that they don't know. You also highlight that in the UK, New Zealand similar findings in terms of the deficit of understanding. But this can also be a challenge if inflation is low, maybe it's too low, maybe they sense something's wrong, they can't put their finger on it. President Trump, for example, he will, of course, he knows a lot more, he's following, he's a little more informed maybe than some of the people in the survey, but maybe the populism that's behind him senses something's wrong and they can't put their finger on it. And this leads to some challenges.

Now, you also talked about the deficit of trust and you highlight that polls show little trust in the Fed. Flesh that out for us a little bit.

Binder: Sure. So, this is mainly based on survey data from the Gallup poll and other similar polls that ask people about different institutions and politicians, and the questions are always phrased a little bit differently, but they'll ask them something like how much do you trust the Fed or the Fed chair to do the right thing for the economy? Or how confident are you in the Fed? It's a little hard to know how to interpret these results, but generally, I think the levels of reported trust in the Fed are probably lower than the Fed would like to see.

And this past year, I've actually gotten to visit a lot of central banks all over the world. And it seems to be a common concern is how do we increase public trust in the central banks? I think this is especially the case because of the political pressures that the central banks are facing and the more that politicians are trying to erode public trust in the central banks, the more the central banks need to worry about this.

A common concern is how do we increase public trust in the central banks? I think this is especially the case because of the political pressures that the central banks are facing and the more that politicians are trying to erode public trust in the central banks, the more the central banks need to worry about this.

Beckworth: Okay. So those are the twin deficits, the deficit of understanding, deficit of trust. And your point is this matters because it harms the legitimacy and the independence of the central bank?

Binder: Right. Because it might ultimately lead to kind of externally imposed policy changes, changes in the central bank's framework or mandate or target or governance structure. Or it might just lead to kind of informal political pressure that at least affects perceptions of independence.

Beckworth:  Okay, so there are these challenges to central bank independence, legitimacy, and ultimately, credibility. And you've done some research, and this is where your other paper comes in. You cite this in your work, but your paper you've gone out, and you've looked at 118 countries over the past decade. And tell us about the findings from that study and how you use it in this paper. What does it highlight for us?

Political Pressure on Central Banks

Binder: Sure. So if I can just explain a bit about the methodology of the data collection. What I was interested in is getting some quantitative data on how frequent is political pressure on central banks. We have pretty good existing data on legal central bank independence. But I think there's kind of a difference between de jure and de facto central bank independence, and I was interested in de facto central bank independence.

So, what I did was to take a narrative approach where I read through, with the help of some of my students,

I read through country level reports from the Economist Intelligence Unit and from Business Monitor International for every country that they had reports available for. And some of these countries are in monetary unions, but in total, there was reports on countries and 118 different central banks.

And in each report, I recorded whether there's any discussion of a politician putting pressure on or interfering with the central bank, and then some details about the nature of the pressure, whether it was pressure for easier policy or tighter policy or something else. Whether it was election related, whether there was any mention of replacing the central bank governor. And then also whether the central bank was reportedly trying to resist the pressure or was succumbing to the pressure.

So, what I ended up with was this large panel data set that lets me look at the different characteristics of central banks or of countries that are associated with pressure on the central bank, as well as what happens to the macro economy after these episodes of pressure.

Beckworth: So I'm looking at your chart, figure one in the paper, and it shows this time series. Well, I guess it's one indicator from your panel set it shows the percent of banks facing pressure. And you see a spike real recently here. So, what did you find, what relationship did you find with this data? So if there's an increase in political pressure, any interesting developments that happened afterwards?

Binder: Yeah. Afterwards, there's more inflation. And it makes sense because more than 90 percent of the time, the pressure is for easier monetary policy. So, I find, in the cases where the central bank is succumbing to the pressure, there's immediately much higher inflation. And for the central banks that resist the pressure, kind of gradually over time, their inflation starts to rise.

So, that could be for two different reasons. It could be that they are trying to resist the politician, but eventually, they give in. It could also be that they continue to resist the politician but their credibility gets eroded. So, inflation expectations rise and that in turn pulls up inflation.

Beckworth: Yeah. And you have another chart in there where you use this data as well as data from another study. And you show that those countries that can be categorized as populist tend to have more political pressure on them. Is that right?

Binder: Right, that either have a nationalist executive or a populist executives. There's better data on nationalism. So that's what I mainly report in the paper.

Beckworth: Yeah, so very interesting findings here. Now, let's take all this, take these concerns, and what we want to do ultimately is tell the story that you do in the paper, what should the central bank do? How can it be proactive because these political pressures are real. And you start with a case study that was really fascinating and surprising when I heard you talk about it. That is the case in New Zealand. So tell us the really fascinating story from that country.

The New Zealand Case Study and its Implications for Central Banks Moving Forward

Binder: Sure. So the Reserve Bank of New Zealand is really famous for being the first central bank to adopt inflation targeting. That was because of the Reserve Bank of New Zealand Act in 1989 that made price stability the primary objective of the Reserve Bank. And they at the time defined price stability as having CPI inflation in the range of zero to two percent. What I explained in the paper is how over time that inflation targeting became more and more flexible in New Zealand, meaning more willing to allow either higher or more variable inflation in the hopes of getting maybe more stable real activity.

So, the way that came about, one important part was that there is an electoral system reform and 1996. The electoral system changed from first-past-the-post system to a mixed member proportional system. That meant that the two main political parties there now needed to form coalition governments with smaller parties in order to get a majority to govern. So, that gave more important role to the New Zealand First party, which I don't know if they describe themselves as populist but they certainly have some some populist tendencies. And they were really part of driving more and more flexible inflation targeting.

So the target range got wider. They kind of changed the horizon at which the bank was supposed to hit the inflation target to be further in the future. And eventually, in the 2017 election, the winning party included a coalition with New Zealand First, and part of that coalition agreement included a promise to reform the Reserve Bank New Zealand. So these reforms which came about in 2018 included a change from inflation target to a dual mandate so that price stability is no longer, it doesn't have a special role as the primary objective. It's a lot like the US where there's a price stability objective and an employment objective.

At the same time, the Reserve Bank also switched from having a single governor making monetary policy decisions to having a monetary policy committee. And the reason I presented this case study in my paper is because I think it makes the point that the political will for how the central banks should behave does matter because politicians can change the central bank. And kind of after the fact of these reforms, the Reserve Bank of New Zealand governor, Adrian Orr has praised the reforms noting that central banks need to adapt to societal needs to maintain their legitimacy, which I totally agree with.

What other central banks might learn from it is that if these kinds of changes are coming, maybe they should drive the change more proactively.

But I think that in this case, the reforms were sort of externally driven. So, what other central banks might learn from it is that if these kinds of changes are coming, maybe they should drive the change more proactively. So maybe other central banks, well, kind of going back to that quote from Bernanke about whether the time was right for a change in 2011. If they still are intrigued by NGDP targeting, they might want to do it now. And that could actually fend off populist impulses to change the central bank. It would actually show that the central bank is willing to depart from the status quo that they know is unpopular, and that could actually help their credibility.

If they still are intrigued by NGDP targeting, they might want to do it now. And that could actually fend off populist impulses to change the central bank. It would actually show that the central bank is willing to depart from the status quo that they know is unpopular, and that could actually help their credibility.

Beckworth: Yeah, it's an amazing story. And you have this great line where you say, "The first central bank to do inflation targeting is the first one to abandon it." So it went from kind of a pure inflation target to a dual mandate. And just like New Zealand was the first inflation targeter, it was the canary in the coal mine, it's also maybe the first one to show what can happen as you highlight, that political pressure can lead to changes in the laws and the governing of the central bank.

So it's kind of a canary in the coal mine, maybe all the central banks around the world should be watching. It was a leader in the past, maybe it is also now. Here's the question I have, how vague is the dual mandate? Because our dual mandate is a little vague in the US. Is it a little more precise over there or is it fairly vague too?

Binder: I'm not entirely sure. I think it's quite similar to in the US.

Beckworth: So you could take it, I guess, and run with it and apply a nominal GDP target over there. That's one of the arguments I've made is that a nominal GDP target, and you make this as well, is consistent with the dual mandate law in the United States. And it'd be interesting to see if New Zealand would kind of be a first mover and be someone to experiment and show the world that nominal GDP targeting can work.

Binder: Yeah, that would be interesting.

Beckworth: I often think that. I think like maybe a Bank of Canada or Reserve Bank of New Zealand might need to do it before the Fed does. But who knows, maybe the Fed will take your advice and become the first mover and jump on it.

One other question about the Reserve Bank of New Zealand before we move on from that particular example. So I know in the past, they had written into their laws, the regulations, that the finance minister, I believe, could fire the governor if he didn't hit his inflation target over the business cycle over some horizon. Is that still the case with this dual mandate and this committee now or have they loosened those screws a bit?

Binder: That I'm not sure about.

Beckworth: Okay. I always thought that was an interesting incentive. My understanding is though, it never was really used or tried. It was a potential way to incentivize good central banking but never was really effective. I mean, there's always ways to wiggle out to say, hey, we had this shock hit us and we had to do this. So, not a very credible threat I guess is the impression I've gotten. Even if that was an interesting idea at the time.

Well, let's talk about the implications of all this for the Federal Reserve here in the US, coming back home. And again, this is a great episode for all the central bankers out there who are listening, particularly those who might work in some part of the Federal Reserve. And you go through again, as you mentioned earlier, the case of New Zealand is very telling and maybe important for us to think about here in the United States. And you give some interesting points that would suggest to avoid the populism forcing on the Fed changes, maybe the Fed takes it upon themselves. Presumably the Fed would have to consult Congress as well when it made these changes.

But you highlight some important points, I think, in thinking about this. The first one is a smooth transition. So talk about that. You provide some really powerful graphs that compare nominal GDP to the unemployment rate relative to what the Fed thought was its natural rate as well as to the inflation rate relative to its target. So talk about all that and how it paints a picture of a smooth transition.

Transitioning from Inflation Targeting to Nominal GDP Targeting

Binder: Sure. So one of the charts that I show in the paper is the log of US NGDP over time. And what's really striking from the figure is that if you start in around 2011 or 2012, the log of NGDP just grows linearly, meaning that I plot a four percent growth path since 2012 on the same figure and log NGDP just stays right on that growth path. So it's as if the Fed were unofficially using an NGDP level target where they're targeting four percent growth per year.

So that could make the transition more smooth because they could argue that they were kind of merely making official what they had been doing for the last decade or so. They might not want to choose four percent, they might want to choose five percent. But I think it could help them show that this might not be that big of a difference. We've been doing this for a while and the economy has been pretty strong.

Beckworth: Yeah, it's a powerful argument. If you look at your inflation chart, the Fed has to explain why does it persistently miss its target. Well, it wouldn't have to say that with nominal GDP. It would say look, we've been doing this, as you mentioned, whether you draw a trend from 2012 or from 2015, it's fairly stable, almost a straight line.

The other thing that's interesting about this is if the Fed made that transition, it would eliminate all of this confusion and discussion over the Phillips Curve. You would just focus on nominal GDP and not worry about, well, is the Phillips Curve broken down or is there some nonlinear relationship? Is it hiding somewhere? Because you would just focus on nominal GDP, not get hung up over what's causing inflation to be below target. This is I think a powerful point not just based on simplicity, but also based on the fact you highlight that your forecasters are using r* and u* when they think about forecasting the economy, is that right?

Binder: Right. When the Fed is kind of pursuing its dual mandate the way it currently does it, one important parameter they looking at is u*. So, the natural rate of unemployment. It's an unobservable variable, we can't actually ever know what it is. So the Fed has to estimate it and they also have to communicate that to Congress, which is pretty difficult to do. It came up in Chairman Powell's testimony to the House Financial Services Committee in July of this year when Congresswoman Alexandria Ocasio-Cortez was questioning the Fed about why it had continued cutting its estimates of u* yet we continue to see unemployment falling below those estimates without any rise in inflation.

I show in another paper that private sector forecasters have really just started to give up on using u* at all for forecasting. So, the first quote from Bernanke that you read also talked about the difficulty of communicating NGDP targeting to Congress. But I think that the status quo is quite difficult to communicate to Congress because they have to talk about this Phillips curve, which might be exhibiting structural changes over time, and which involves unobservable parameters that might be changing over time. I think that's actually harder to communicate than an NGDP target might be.

But I think that the status quo is quite difficult to communicate to Congress because they have to talk about this Phillips curve, which might be exhibiting structural changes over time, and which involves unobservable parameters that might be changing over time. I think that's actually harder to communicate than an NGDP target might be.

Beckworth: Yeah, this would be a great approach for those who want to let the economy continue to grow and bring people into the labor markets. So the Fed wouldn't be worried about the economy overheating with inflation. It would just say, look, let's continue to grow nominal GDP on this path we've been on. Let labor markets continue to bring people off the margins into the labor force. And I think it'd be a great remedy, a great fix to some of these tough questions and get away from, “oh, we've gone too far”.

It's striking how much the FOMC's estimate of the national rate of unemployment is changed since they started raising interest rates. Adam Ozimek has this paper you've probably seen, where he estimates how many more jobs could have been created had they had the estimate of u* back then that they have now. They would have maybe even raised rates at a much later date. So, very interesting. And again, it gets to this dicey question of when do we raise rates, when do we lower them? What does the Phillips curve tell us?

All right, so that's a smooth transition point. It would be kind of easy to keep going along the path that we're on. You also make this point, it would improve understanding and trust. So explain that.

Binder: I think for households and consumers, it would improve understanding because of the way many consumers think about inflation, which is that I don't think they think about inflation the way a typical macroeconomist would. They kind of think of inflation as a bad thing. So when they're responding to surveys of expectations, a lot of consumers just give a high number for their inflation forecast if they think the economy is doing poorly and give a low number if they think the economy is doing well.

So it's kind of, for one, it's sort of hard for the Fed to then use that to gauge their own credibility. And also, with GDP targeting, by construction, inflation is going to be countercyclical. So when real GDP growth is low, inflation is going to be higher, and when real GDP growth is high, inflation is going to be lower. And that's actually the way a lot of consumers kind of like to think that the economy works. So, it would make their models, which are intuitive to them, a better match with reality.

Beckworth: So what you're saying is that the public, they are nominal GDP targeters at heart, they just don't realize it yet.

Binder: I don't know if they're nominal GDP targeters, but I think they have kind of memories of the 1970s or 80s when the inflation was high and the economy was bad, and they still hear about developing economies where you hear inflation is high and that's a bad thing.

Beckworth: I'm being silly. I'm just trying to highlight the fact, which was interesting and I didn't realize until I'd read your work, that most people tend to see inflation going up during bad times and vice versa during good times.

So, interestingly enough, nominal GDP would confirm those priors as you mentioned. So, another reason to chalk up and support nominal GDP targeting. So, on many levels, it would be an easy transition, it would be continue doing what we're doing, it would fit your priors. And you bring up another interesting finding that you've got in the literature as well as your own work. And that is when you ask people about their forecast of inflation, that initially they're way out there, but once they become aware that they're being asked in this survey and they get asked again, they be more cognizant and start thinking about it. And there's these big revisions. Is that right?

Binder: Right. This especially shows up in the New York Fed's relatively new survey of consumer expectations, because in that survey, the same respondents will take the survey 12 times in a row at one month intervals. So it seems like after the first or second or third time that they're asked what do you think inflation is going to be, they start giving numbers that are much more reasonable. So what seems to happen is that they go online and look up what inflation has been or they somehow or another inform themselves, and they revise their forecasts down by something like four percentage points after the first time they take the survey.

Beckworth: So this could present problems for the Fed when it's thinking about what does the public think about inflation moving forward. Is that right?

Binder: Right, because then your survey respondents are no longer a representative sample of the US population because they're more informed than the average person in the US population. You could still look at just the first time respondents and that would be kind of more representative of what the average person would report.

Beckworth: But there's a problem there, and then kind of the whole, as you mentioned earlier, one of the premises of inflation targeting is you're signaling to the public, and you got to know what they're thinking through these surveys and maybe these surveys aren't very informative, or as informative as they should be.

Now, I bring that up not just to criticize inflation targeting, but interestingly enough, this tendency to revise their forecasts isn't true for these surveys when they look at households outlooks of their income, their nominal incomes.

Binder: Of their own nominal income.

Beckworth: Their own nominal income, yes.

Binder: As far as I know, I don't know of any surveys that ask about aggregate nominal income and nominal GDP. I think it's sort of as you would expect, people kind of have a good idea of their own nominal income or how much it's been growing, and they're also not easily able to do a quick Google search that would give them a better idea.

Beckworth: Yeah, so this would be more robust. Even if the Fed began doing nominal GDP targeting, because households have their own feel for what they're facing, they probably wouldn't change as much as they would with inflation targeting. Is that fair?

Binder: Yeah, I think these panel conditioning effects would be smaller. But they might need to introduce a new survey or a new survey question that includes a question that's specifically about NGDP. I would love to see that. I don't know exactly what the responses would look like. But if I had to guess, I would think they would be a little more what we would consider reasonable than the way the inflation expectation is like.

Beckworth: Yeah. And one of the critiques I often get is how are you going to measure nominal GDP in real time. But you touched on some of this. We already have some of these surveys that ask at the monthly frequency, what do households think their nominal incomes will be going forward. And the Fed could draw upon that. And to some extent, I think the data is endogenous to the target the Fed picks.

If the Fed picks a nominal GDP target, there'd be a whole lot more resources dedicated to determining real time estimates of nominal income, nominal GDP. I think of some of the big data projects. Claudia Sahm is coming on the show in a short while. She's worked on some big data projects at the Fed. I think JP Morgan has this real time data set based on millions of credit card transactions and debit card transactions. So, I suspect that if the Fed went down this path, it'd be an even richer data environment.

Binder: Right. Yeah, I don't think that data concerns should be really kind of the big limiting factor in the kinds of policies that central banks choose to pursue because it's just getting, I think, easier and easier to get better and higher frequency data about either actual aggregate statistics or about expectations.

Beckworth: The real concern here was populism, not data availability. So in the grand scheme of things, getting your data is going to be a whole lot easier of an obstacle or a challenge to deal with in dealing with populism and the effect it will have on your independence and legitimacy.

So you make this great case for nominal GDP level targeting. And again, it's a great kind of proactive way to deal with the pressures for central banks to do more. But in a disciplined way, you still provide a medium to long run nominal anchor. And there's better maybe trade-offs and short run between employment and inflation. Any other thoughts on this proposal or your paper?

Binder: Just to tell you, I hadn't actually really thought much about NGDP targeting until I was asked to write a paper for this Cato Monetary Policy Conference about targets and mandates. So what I started thinking about was, well, what kinds of targets or mandates do we want given what we know about the politics of central banking and about expectations formation and central bank communication.

And so I only really, very recently kind of started talking myself into supporting NGDP targeting. So I'm pretty new to it. I think I maybe come across as more convinced in the paper than I am in real life. It was a really fun experience participating in that conference and kind of writing a paper that is a little bit more of a policy recommendation compared to my usual sort of more just academic research.

Beckworth: Now, that's interesting. So, just to clarify, you were wrestling with how do we deal with populism and central banking, and you kind of came upon nominal GDP targeting as an answer?

Binder: Yeah. Even compared to just a few years ago before I had seen so much of this political pressure on central banks and done my own research on it, I was kind of more in support of the benefits of central bankers having discretion because I thought that they're technocrats, they know really well, they understand what's going on in the economy. We should give them room to use their discretion, which I think it would be kind of true in a world where their central bank independence was very strongly enshrined.

But when it's not, I think the benefit of having a single explicit quantitative target is just so much greater because, one, it gives them that accountability and it leaves a lot less room for politicians to argue that “well, you need to cut rates for this reason or that reason.” There's just a really clear measure of kind of the stance of policy and “do we need to ease or to tighten?” It's a lot simpler to decide when you have a single explicit target. I really didn't believe that even a few years ago.

I think the benefit of having a single explicit quantitative target is just so much greater because, one, it gives them that accountability and it leaves a lot less room for politicians to argue that “well, you need to cut rates for this reason or that reason.” There's just a really clear measure of kind of the stance of policy and “do we need to ease or to tighten?” It's a lot simpler to decide when you have a single explicit target.

Beckworth: Fascinating. You've had your kind of conversion on the road to Damascus experience here. Well, you said sort of. You said your paper's stronger than your own person.

Binder: Yeah, I reserve the right to change my mind. I’m still a young economist.

Beckworth: Okay. Hey, I'm convinced. You've convinced me obviously, and of course, I was convinced already. But one thing I want to highlight in the last few minutes we have remaining today, something that was just really striking to me is the challenge that comes with trying to measure inflation expectations from households. And also, who is shaping inflation expectations? So you've touched on this and you mentioned kind of the revisions and the problems and are these surveys representative?

You've done a lot of other research along these lines. I want to touch on just a few of them. You've had several papers, I'll mention their names. One from 2015, *Whose Expectations Augment the Phillips Curve?* And in that one, you find that only certain types of people really get the expectations that are consistent with what's actually happening. And again, this speaks to what we've touched on earlier on who's actually paying attention to these issues. Speak to what you found in that paper and how it ties into our earlier conversation.

Inflation Expectations and its Impact on Inflation Dynamics

Binder: Sure. So in that paper, what I do is to estimate Phillips Curve relationships using average expectations of different demographic groups. So, I might regress inflation on the inflation expectations of high income, medium income and low income consumers and some measure of real activity. And in that case, the coefficient on expectations that's going to be statistically significant is the one on the high income consumers' expectations. And similarly, I find that college educated respondents or male respondents seem to have the inflation expectations that are driving inflation dynamics.

Now, that doesn't necessarily mean they have more accurate expectations because in fact, even college educated households inflation expectations are less accurate than professional forecasters' expectations. The means grade forecast errors are larger. But if you include both the consumers' expectations and the professional forecasters' expectations in the same Phillips Curve regression, you find that it's those consumers' expectations that are driving inflation dynamics, especially the high income or highly educated consumers' expectations.

I think that that has to do with who are the price of wage setters in the economy. And it's not necessarily professional forecasters. It's like those college educated households’ expectations are a pretty good proxy for price setters’ expectations.

But if you include both the consumers' expectations and the professional forecasters' expectations in the same Phillips Curve regression, you find that it's those consumers' expectations that are driving inflation dynamics, especially the high income or highly educated consumers' expectations. I think that that has to do with who are the price of wage setters in the economy.

Beckworth: Okay, so, they're the ones who demand a higher wage because they expect prices to go up. And it actually happens, it actually affects the real economy and what goes on. Interesting, I found that very interesting. You probably saw there was a recent study, I believe it was in Europe, where they looked at all this data, they matched I believe like 20 Scandinavian countries with really great data sets where they matched household balance sheets and they matched a number of surveys.

And one of them was IQ test. And they found that people of higher IQs tend to understand what's going on. And that seems to kind of at least somewhat consistent with what you found there. Or am I off, I'm not sure how to link that up to your work.

Binder: Yeah. I'm not really comfortable with knowing how IQ is measured, or how IQ, for example, I don't think that, well, I find that male expectations drive inflation a lot more than female expectations. But I don't think that there's that big of a difference in male and female IQ. But yeah, there is something to just that different groups are more informed about inflation than others. And it's often correlated with things like, did you go to college and do you have like high income or do you have investments in the stock market, things like that.

Beckworth: Yeah. So that IQ might just be a proxy for the better educated folks. And just to be clear, when you go to places with really, really high inflation, you do see more people becoming engaged, lots of people become engaged. It's just like we said earlier, they don't have to be engaged now because things are relatively stable.

Binder: That's like a rational and attention type argument, where when inflation is low and stable, if it's costly for you to go spend your time looking at inflation statistics, then it's actually rational for you to not look at them. But if you're somewhere where inflation is really high and/or really volatile, then it might be more worth your time to pay attention to inflation.

Beckworth: You have a paper or two on this, right? And you draw some conclusions about the relevance of these sticky information models. Can you speak to that?

Binder: Yeah. So sticky information models are based on the idea that it's not totally free to update your expectations even if there is publicly available information on inflation, you still might not look at it at every moment of your life. So you pick some, you allow some time to pass between when you update your expectations.

 In these models, it's really important to know, what's the probability in a given period that somebody is going to update their expectations? So a lot of studies have tried to estimate that stickiness parameter.

So, the point I make in a short paper is that the frequency of our survey data actually matters a lot for our estimates of the stickiness parameter. So, in the Michigan survey, when people take that survey, if they take it twice, there's a six month gap in between when they take it. But the New York Fed survey, they take it with a one month gap. And in the Michigan survey, it actually looks like people don't revise their expectations very often. So a lot of times there'll be no change from when they take it one month to when they take it six months later.

But in the New York Fed survey, you see very frequent revisions to their expectations. But if you kind of take that monthly data and pretend that it's six monthly data, so you only look every six months, you also see that it doesn't look like people revise very often. Because what they'll do is they might say two one month, and then they say three, and then they go back to two, and then they go back to one. But they end up on two again. So you miss all those changes in the middle by using too low frequency data.

So using the monthly data rather than the six monthly data, I actually find that people revise their expectations more frequently than we would have otherwise thought.

Beckworth: Which raises questions about the relevance of a sticky information model, ultimately, right?

Binder: Right. It might be that they are updating their information frequently but they're getting really noisy signals. So every time they drive to a gas station, maybe they change their expectations of inflation, but that signal is a pretty noisy one.

Beckworth: Okay. Well, we have a few minutes left and I want to end on one other related area. You have a paper titled *Federal Reserve Communication and the Media* in 2016. I know it refers to a huge undertaking you took looking at a bunch of news stories few years back. But based on what you've learned there and your understanding of the literature on communication and all the survey data you've looked at, you've done a lot of work, very labor intensive work looking at surveys, creating this populist data set.

I'm curious what you think about the Fed's communication approach. In particular, the Fed's press conferences, all the speeches that are given, Jay Powell has gone to having a press conference after every FOMC meeting. Lots speeches are given. I'm wondering if you have any sense on whether there's more signal or noise that comes with this increased access to these central bankers and the talks that they have. Any thoughts?

Signal vs. Noise in Federal Reserve Communication

Binder: Yeah. So actually, I recently got to be a discussant on a really interesting paper by Michael Lamla and Dmitri Vinogradov and it's now in the Journal of Monetary Economics. What they did was to run their own surveys a couple of days before and a couple of days after about a dozen different FOMC meetings that were associated with press conferences.

And they asked people not only their inflation and interest rate expectations, but also whether they had heard any news about the Fed. And they did find that going from a couple days before to a couple days after a press conference, there was a big increase in the number of people who said they'd heard some news about the Fed. But they were not in able to detect any effect on either inflation or interest rate perceptions or expectations.

So, it's kind of an interesting result because it says that maybe, there's some kind of signal people are hearing, “oh, the Fed had a meeting, the Fed chair did a press conference,” but they're not necessarily able to extract any information from that, at least not about inflation or interest rates. I think it's still possible that maybe it has some effects on consumer confidence. But it's interesting that they didn't detect any effect on especially interest rate expectations.

Beckworth: Yeah, that is interesting. There's maybe more noise, I don’t want to say more noise, less signal, but there's definitely maybe more noise and the signal hasn't changed. They're not getting any better at least at determining the path of interest rates.

One last question along these same lines. Do you think the Summary of Economic Projections are useful for the public at large or just for us Fed watchers?

The Importance of the Summary of Economic Projections

Binder: Well, I think that it's mainly the Fed watchers and people working in finance who are looking at the summary of economic projections. The only time I really saw them make the kind of the popular press was what I referred to earlier when Congresswoman Ocasio-Cortez mentioned the Fed's estimates of u* at the congressional hearing. And what she was actually referring to where FOMC’s longer run projections of unemployment. So that actually made mainstream news, but I think normally the SEP probably doesn't make the really mainstream news.

There is an interesting survey that David Wessel and some people at Brookings did, a survey of Fed watchers, asking them about their opinions of different forms of Fed communication. And with the SEP, they really wanted at least some link between the projections of different variables. So with the SEP, we see all the dots, all the projections of unemployment, all the projections of the fed funds rate or of inflation. We don't know which unemployment dot goes with which inflation dot or which interest rate dot.

So that means that we can't really infer anything about different FOMC participants’ reaction functions or the models they're using of the economy from those dots because we don't know which one goes with which.

So it might even be more useful to just provide a consensus projection if they're not willing to link those dots across variables.

So with the SEP, we see all the dots, all the projections of unemployment, all the projections of the fed funds rate or of inflation. We don't know which unemployment dot goes with which inflation dot or which interest rate dot. So that means that we can't really infer anything about different FOMC participants’ reaction functions or the models they're using of the economy from those dots because we don't know which one goes with which. So it might even be more useful to just provide a consensus projection if they're not willing to link those dots across variables.

Beckworth: That's a great point. I like to think going back to kind of the main topic of our conversation, nominal GDP level targeting and how it might simplify and make life easier for the Fed, the Summary of Economic Projections itself might be simplified for that as well. And a lot of these considerations, how well does the public read the Fed's language? How easy is it for the Fed to communicate to the public? All these things might be simplified if they adopted a nominal GDP level targeting approach.

And that's the argument you make in your paper and of course, I'm sold on that and I hope the central bankers listening today will also take it to heart.

Well, with that, our time is up. Our guest today has been Carola Binder. Carola, thank you so much for coming on the show.

Binder: Thanks for having me. It was a lot of fun.

Photo by Drew Angerer/Getty Images