Scott Sumner's Unconventional Monetary Theory

A Macro Musings Transcript

The following transcript is from Macro Musings, a podcast series hosted by Mercatus scholar David Beckworth that explores the past, present, and future of macroeconomic policy. Subscribe on Apple Podcasts or your favorite podcast app.

David: Today's a special Macro Musings episode. We're live at the University of Texas at Austin as part of the financial crisis symposium titled Ten Years Later: What Does the Data Say? This event is hosted by the Center for Enterprise and Policy Analytics at the McCombs School of Business, and our guest today is Scott Sumner. He's a colleague and a former guest, so our listeners will be familiar with him, and in the spirit of this conference, Scott has agreed to come on and discuss what he thinks the data tells us about what happened 10 years ago during the crisis with regards to monetary policy. Scott, welcome back to the show.

Scott Sumner: Thank you for inviting me, David.

David: So kind of the standard narrative surrounding the crisis is there was a housing boom, a bust. It affected financial markets, eventually turned into severe financial panic and we had the great recession. Now you take a slightly different view on what actually happened and transpired. You acknowledged all those things happen, but you actually say there's a piece of the puzzle that most observers miss and that's what the Federal Reserve did during a pivotal period during that time. Can you tell us what that is?

Scott Sumner: Right, and since my view is a little unconventional, let me start off first with an analogy and then tie it into my view. So think about somebody who has a mild cold and it gets worse. It turns into a really bad cold. They go to the doctor and they find out they've actually gotten pneumonia and the cold medication they've been taking is not effective. They really need to take antibiotics. It's a different kind of illness. So I would use that analogy to sort of explain my interpretation of the events leading up to the Great Recession.

Scott Sumner: The cold, the mild cold was the early stages of the financial crisis in 2006 and 7. We know the housing market boomed and then went into a bust. We know that some banks got into trouble. But during 2006 and 7, the overall economy was still holding up fairly well, even as housing was declining. Then in 2008, this sort of cold morphed into pneumonia. It became actually a different kind of illness, not just that the financial crisis was getting worse and worse and the economy was getting worse and worse. That's how it looked at the time to most observers.

Scott Sumner: But behind the scenes we're starting to see a decline in nominal spending. Dave and I focus a lot on nominal GDP. And as that started declining, you started to see a big increase in unemployment, and importantly, the financial crisis got a lot worse. And I think that wasn't recognized really until it was too late. It was really sort of in 2009 that policy makers recognized they had a major problem with aggregate demand, not enough spending. And they had to move from trying to rescue the banking system to try to have more effective monetary policy to keep the economy on an even keel.

Scott Sumner: But that recognition came too late because policymakers were inappropriately interpreting two key macro variables, interest rates and inflation, and the misinterpretation of those variables caused monetary policy to go off course in 2008. Do you want me to talk about—

David: Yeah, tell us about what happened in 2008, because in a part of your story is that the Fed kind of sat on its hands with regards to the monetary policy. It's very active responding to the financial panic, but between April 2008 and October 2008, it kept its target rates steady and it was signaling it actually may even do interest rate hikes because of the inflation. Why was it confused by the inflation of 2008?

Scott Sumner: Right. So, as you may know, the Fed does inflation targeting at about two percent, but it's really more complicated than that because there's different kinds of inflation. There's supply side inflation, which is created by shocks like sudden increases in oil prices, and then demand side inflation caused by overspending in the economy. It's really demand side inflation that the Fed is concerned about. There's not much they can do about supply side inflation.

Scott Sumner: Well, in the middle of 2008, inflation got pretty high even as we're going into this sort of banking, financial, housing crisis. And so the Fed was looking at that high inflation. I mean they sort of understood that oil prices were high, but nonetheless, they became increasingly worried about inflation being too high. As a result, even as late as September of 2008, after Lehman Brothers failed, the Fed refused to cut interest rates, which were two percent at the time, and they cited a fear of high inflation as the reason they refuse to try to stimulate the economy after Lehman failed.

Scott Sumner: A lot of people kind of forget this. So this misinterpretation of the inflation numbers led them to be not expansionary enough in monetary policy, and that's what allowed nominal spending to really plunge in the second half of 2008. The other variable they were misinterpreting was interest rates. Although at a certain level, most economists understand that interest rates are not a perfect indicator of the stance of monetary policy. Nonetheless, most economists overemphasize interest rates when thinking about monetary policy and the mistake people make is to assume that relatively low interest rates represent easy money. Through all the middle of 2008 that you referred to David, the interest rates were being pegged at two percent.

Scott Sumner: Now that sounds like easy money, but what's called the natural rate of interest was plunging sharply below zero because of the problems in the housing market and banking and so on. And as the natural rate of interest plunged into negative territory, pegging interest rates at two percent actually represented an increasingly contractionary policy, and that's because interest rates alone are not a good indicator of monetary policy. You have to look at the path of nominal spending in the economy to see if money is too easy or too tight.

David: Okay, going back to the confusion though, the Fed did this because they saw inflation spiking in 2008, and a true flexible inflation targeter in theory should see through temporary spikes in inflation, right? So the difficulty in real time, I think your point you're making is it's hard to know in real time if that inflation's going to be permanently higher, or kind of a one-time shock it was in 2008. So you know Monday morning quarterbacking, you're looking back. We say, "Well the Fed should have not been worried," but they were. And something you mentioned before is that if you look at Fed fund futures contracts and look at the one year ahead market forecast where the Fed funds rate would be from January through June of 2008, it actually was going up.

David: So by June 2008, Fed fund future contracts were at three percent. So the market was expecting the Fed to actually raise rates. The Fed was signaling through its talks about inflation fears. There were many people in the FOMC were really worried about inflation and that is sending a signal that if anything, they're going to tighten policy. So not only is the natural rate going below, but what you argued is that they were signaling a rate hike takes place. So your contention then is that this was a pivotal point, a catalyst that made things worse. Do you have any evidence from market prices that support your view?

Scott Sumner: Right. So sometimes David and I are called market monetarists because we emphasize market forecast more than traditional monetarists. And one point I would like to emphasize is that this is not just Monday morning quarterbacking. Even at the time, I was very concerned about a Fed policy, and actually, very concerned about Fed policy for the first time in decades. That is, I'd gone several decades thinking, 'Yeah, they're doing a reasonably good job and I don't have any major complaints. Maybe I would tweak it this way or that." But later in 2008, especially around September, it was just very obvious policy was off course. It was too contractionary.

Scott Sumner: Let me give you one example. In the meeting after Lehman failed, where they refused to try to stimulate the economy, they cited a risk of high inflation. And it's true that over the previous 12 months, inflation had run well above two percent because of the spike in oil prices. But if you looked at what's called the tip spread, which is the bond market forecast of inflation going out over the next five years, that was only 1.two percent on the day of the Fed meeting. So when the Fed met and said they're not going to ease policy because they're concerned about inflation, the markets were actually telling the Fed, "No, inflation's going to probably run well below two percent over the next five years." And guess what? That's roughly what happened.

Scott Sumner: Inflation did run well below two percent over the next five years. Now the markets aren't always right, but the market forecast needs to be taken more seriously, especially when you're in a fast moving crisis situation where backward looking at like last year's data doesn't tell you what's really going on in the economy at that moment. It's like driving by looking in the rear view mirror instead of looking down the road ahead as to where the economy is going. And we feel that market indicators are some of the best evidence we have of where the economy is going and the Fed should have paid more attention to those indicators, which were apparent, even at the time.

David: All right, so your contention is the Fed didn't do as good of a job as it should have done during that period. In fact, it was a pivotal turning point. If we look across the Atlantic at the European Central Bank, how did they do in 2008?

Scott Sumner: Well, they made similar mistakes and even worse mistakes, to be honest. The ECB is set up with a little more of a contractionary monetary bias than the Fed for various institutional reasons. But most central banks operate sort of along the same sort of policy regime. And so it's not surprising that when faced with a certain set of events, multiple central banks will make the same kind of mistakes. What's interesting to me about the European Central Bank is that the fact that they made even a more contractionary monetary policy in 2008 is itself very revealing about the crisis.

Scott Sumner: Like let's go back to the standard view that I'm wrong and what really went on is we had a banking and housing crisis which created a great recession, that it wasn't tight money. That standard view would suggest the recession should have been much worse in the US than in Europe because the US had much greater exposure to things like the subprime mortgage crisis. European banks also had exposure, but the housing bubble had not been as large in Europe as in the United States. So if that standard story is correct, Europe should have gotten sort of collateral damage. And that's the perception, by the way, in 2008.

Scott Sumner: At the time, the Europeans were kind of gloating that, "Oh, look at the US. They've been doing so well for years talking big about the success of their economy, but now they've been brought low by this reckless lending and so on." So everybody in Europe felt it's going to be a lot worse in the US. In fact, it turned out the Great Recession was much worse in Europe and the reason is very clear. European Central Bank monetary policy was clearly more contractionary all the way from 2008 for the next five years or so than the US, and they actually had a deeper double dip recession than the US had.

Scott Sumner: And that double dip recession correlates very strongly with monetary policy decisions. They raised rates in July of 2008, right in the middle of that crisis period. And then during the recovery when our Fed was trying to stimulate the economy with QE, they raised rates twice in 2011 and promptly went into a double dip recession in late 2011 that lasted all the way to 2013. So the crisis in Europe was very closely correlated with clear errors by the European Central Bank.

David: Okay. So we can count our blessings we didn't have the ECB running monetary policy.

Scott Sumner: It was even worse, yes.

David: Are there any central banks that did a good job during this period? This is a very depressing conversation so far. Can you point to anyone who did a decent job at least?

Scott Sumner: Yeah. I think several. Israel I think did a fairly good job. The one I point to usually is Australia, which is an economy that is sort of like the US economy, if you were to look around the world and say, "Which economies most closely resemble the United States?" Australia has not had a recession in 27 years. As an aside, keep that in mind when people tell you we can't go more than 10 years without a recession because we've never done that. It can be done. So how has Australia been able to avoid recession?

Scott Sumner: I think they've had a monetary policy that's done a better job of keeping nominal GDP growing along a relatively steady trend line. Now they do have fluctuations and they did have a slow down during the global recession of 2008 and 9. But if you sort of look at the trend line in Australia, they tend to come back to the trendline fairly closely. And so they were able to keep nominal GDP growing with a more stimulative monetary policy. I will concede there's a little bit of luck involved there because the way central banks operate is by adjusting interest rates, and when interest rates hit zero, it's harder for central banks using that tool. I have trouble with that tool, but as long as they're using interest rates as a tool, it's easier when interest rates are above zero. And in Australia for various reasons, interest rates are traditionally higher than in other developed countries. So they were a situation where it never felt a zero and they continued to use their ordinary interest rate adjustments as a tool to keep nominal GDP growing at an adequate rate for Australia.

David: So Australia had a similar housing boom, credit boom, very similar in that regard. And they didn't have a great recession is what you're saying?

Scott Sumner: Yes. And by the way, I also think the housing boom and bust has been misinterpreted. Not only Australia, but New Zealand, Canada, Britain, and the United States, all sort of the major English speaking economies of the world had enormous housing booms from 2000 to 2006. Now the standard view is prices in America got way too high in 2006 and had to crash for that reason. However, if that were true, wouldn't housing prices have crashed in these other four English speaking countries? In fact, they didn't. They tended to move sideways or even drift higher in the other four. So I don't believe that a crash was inevitable. Perhaps some pullback was, but the severity of the house price crash in the United States I think was clearly related to the severity of the Great Recession.

Scott Sumner: So let's assume there were some excesses in certain markets around the US and there would have been somewhat of a drop as we saw in 2006 and 2007. What wasn't necessary is the big drop in nominal GDP, which put a sort of second round of decline into the housing market. And if I'm not mistaken, in parts of the country like Texas, there really wasn't much decline at all during the first wave because I don't think you guys had had the bubble. You have a more stable price here because of zoning rules and so on, but you did have some decline when we got into the deep recession because Texas was also affected by the national economy. So it's also interesting to compare regional economies like Texas to the more bubble areas like Nevada, Arizona, California, and so on. And when you do that, you can sort of see with a national crisis of a fall in nominal spending hitting the housing market as a second wave and it's spreading from localized problems to a national housing problem.

David: Okay. So again, your contention is that there may have been an ordinary, garden variety recession in 2008, but it did not need to be a great recession?

Scott Sumner: Exactly.

David: A deep and sharp. Now you keep throwing this term around, nominal GDP. And maybe for our listeners, what is nominal GDP? Why don't you keep talking about that as opposed to say inflation is a measure of monetary policy?

Scott Sumner: Right, So nominal GDP is not the GDP number you might have heard on the radio this morning when they talked about GDP going up I think 3.6 percent. It's both real GDP plus the inflation component. It's actually the total increase in dollars spending in the economy. And I think it's a very unfortunate thing that as a convention, the term GDP implicitly refers to real GDP, because in terms of government policy, nominal GDP is actually a better indicator of what's going on in the economy in terms of stabilization.

Scott Sumner: Now, real GDP is more important for our living standards and that's why people pay attention to it. But it's fluctuations in nominal GDP that tend to destabilize the economy. Let me explain why that is. So nominal GDP you could think of is like the total value of all gross income earned by everybody in the economy. And that, before the Great Recession, trended upward at about five percent a year, three percent of that was real growth, another two percent was inflation. People make all sorts of decisions based on an expectation that that will keep growing over time. Companies sign labor contracts. People borrow money. Well, nominal GDP is the total resources that companies have to pay their workers. Nominal GDP is the total income people have to repay their debts, or companies, or even governments like the Greek government has to repay its debts.

Scott Sumner: So when nominal GDP falls sharply, as it did in 2009, below trend, below what was expected, what happens is companies ended up laying off lots of workers and people end up defaulting on lots of debts. Now is that just too in the Great Recession? No. All throughout history, you see this over and over again. Argentina, around 2000, the US in the Great Depression. Whenever there's a fall in nominal GDP, you get high unemployment and you get a debt crisis, or almost always. The Great Recession was no different. Instead of growing at five percent as had been expected in 2009, it suddenly fell by about three percent. That's eight percent below trend. That's a big shock to an economy. That's a lot of income that people and companies expected to have that suddenly wasn't there. And again, it shows up in the labor market and in the financial market because of contracts.

David: So nominal GDP is dollar income, or another way you framed it in other settings is it's also the total amount of money spent. You mentioned that earlier. So total money spending on the economy and your argument is the Federal Reserve, over the meeting the long term should have quite a bit of control over it.

Scott Sumner: Yeah. Let me make one other point. Why is it not real GDP? It's because our contracts are almost always nominal. Your mortgage is a nominal amount of money. You repay your labor contracts a nominal amount. So if we signed real contracts indexed to inflation, I'd be focusing on real GDP as stabilizing, but we tend to sign nominal contracts and so it's the fluctuations in nominal GDP that's destabilized the economy.

David: Okay, so you're critical of what the Fed did in 2008, but you know since then they had three rounds of QE, interest rates went zero percent for a long time. Many, if not most observers would say monetary policy was really accommodative, really easy, really supportive. What is your take on the recovery period?

Scott Sumner: Right, so I think that they're, they're looking at the wrong indicators and in a way they kind of should know better. The two indicators often cited for easy money are either low interest rates or QE. The term QE refers to a big increase in the monetary base, sort of printing money, if you will. Now, what I find interesting is that if you look at academic work by economists, they'll acknowledge that both of these are unreliable. Like low interest rates could reflect easy money, but it could also reflect a weak economy, just as high interest rates might reflect either tight money or high inflation, pushing up interest rates.

Scott Sumner: It's not always tight money when rates are high, it could be hyperinflation or something. And we also know the monetary base is unreliable. It could be an easy money policy if you inject money into the economy, but sometimes you're just accommodating a crisis where there's a demand for liquidity. I mean, they even did QE in 1932 in the Great Depression, but in retrospect we feel that monetary policy was too contractionary during the early '30s. They didn't really do enough relative to the demand for liquidity that was so high during the crisis period.

Scott Sumner: So when I talk about easy and tight money, I'm doing it in relative to what's needed to hit the Fed's targets. Like is monetary policy expansionary relative to what would be needed to keep nominal GDP growing at an adequate rate? Is it too much, too expansionary, creating high inflation? Or too contractionary creating recession? I don't look at interest rates or QE. I look at growth in nominal GDP relative to what's desirable, and judge the stance of monetary policy on that basis. Too much or too little relative to what's needed is I think the only useful measure of easy and tight money.

David: So was monetary policy highly accommodative 2009 to 2015 when they started [inaudible 00:22:29]?

Scott Sumner: Not according to my definition because nominal GDP growth was very sluggish, subpar during the recovery. We had a subpar recovery in both nominal GDP, and not surprisingly, in real GDP. And by the way, those two don't always go together. They're very different types of variables. Like in Zimbabwe, you had hyperinflation with nominal GDP going up and real GDP going down. But in America, over the business cycle for various reasons, nominal and real tend to be correlated in the short run. So-

Scott Sumner: Nominal and real tend to be correlated in the short run. So if you can have a smooth path for nominal GDP, fluctuations in real GDP will be smaller than if nominal GDP is going up and down because of inadequate monetary policy.

David: So let's take your argument as given. Why didn't the Fed act more aggressively in 2008 and say, 2009? Why not a ease more aggressively? What held them back?

Scott Sumner: Yeah, so quite a few things. As I said early on, they first misdiagnosed the crisis in 2008. They thought fixing banking would solve the problem, whereas they needed to actually change monetary policy. Later, they did recognize they needed to change monetary policy. They cut interest rates to zero in December of 08. They started QE in 09. But, they didn't do what I would call the whatever-it-takes approach. In my view, what the Fed should do is take whatever instruments they're using for policy and set them in a position where they expect growth to be where they want it to be. This is called targeting the forecast.

Scott Sumner: If I could use an analogy, let's say you're on an ocean liner going from Europe to New York, and halfway across the ocean you're talking to the captain on the deck and say, "When should we reach New York?" The captain says, "Well, we expect to land in Boston in three days." And you go, "Boston, I thought we were going to New York." And the captain says, "Well, yes, but because of wind and waves, we've been pushed off course so now we forecast that we'll end up in Boston." You might ask the captain, "Well, have you thought of adjusting the steering so that the city you expect to land in is the one that you actually had targeted at the beginning of the voyage?"

Scott Sumner: Well, that's just common sense with ships, but with monetary policy we don't always do things that way. At the end of 2008, the growth of nominal GDP that the Fed expected over the next year or so was less than what they desired. They were hoping to be surprised on the upside. That's the situation I think monetary policymakers should always avoid. They should do as much as necessary, whatever it takes so that their forecast of growth and nominal spending is equal to what they want to occur to nominal spending. That concept is called targeting the forecast. Because the Fed is a rather conservative institution in a small [inaudible 00:25:35] of cautious, they're often reluctant to do unconventional or aggressive steps. But I think that kind of caution is actually short sighted because down the road it leads to even bigger problems which force it to do even more extreme steps.

Scott Sumner: And again, here I can give you an example. Early on, the ECB, which is more cautious than the Fed, was very sort of contemptuous almost of Bernanke's aggressive moves towards QE at the Fed. And the ECB was more cautious, they weren't pursuing monetary stimulus. They thought he was being a little bit reckless and they were doing it the right way. But as a result, Europe went into a much worse recession, a double dip downturn. And then in the long run, the ECB was forced to do even more radical things than the Federal Reserve. The ECB got so deep into recession that they did negative interest on bank reserves, negative interest rates, something that was too radical even for the Fed to do. And they did QE, a lot of QE in Europe. But that was because there earlier caution had allowed the economy to slide much too deeply into recession.

Scott Sumner: These problems are much easier to address if you do them quickly and strongly with whatever it takes to keep the economy on an even keel than if you wait to where the economy slips so much that you have to take much more controversial and aggressive steps.

David: Okay. So your argument is the monetary policy was effectively too tight during the crisis and to some degree as well afterwards. And one of the points that people would maybe agree with you on is that the fact that the Fed's inflation targets two percent and it has persistently undershot that. So if you look the core inflation measure that it likes, the core PCE deflator, it's averaged about one and a half percent since the crisis. And some observers would agree with you and say, yeah, the Fed has persistently undershot over the medium, the long term. The Fed should guide that the two percent. So its shortfall relative to its own target in inflation terms. Yes, you would love nominal GDP targeting but relative to inflation.

David: What should be done? Should the Fed overshoot? I mean, should there be a symmetric inflation target? The Fed even says it has a symmetric inflation target but it persistently has undershot it. We're finally back at two percent, but almost a decade of below two percent inflation. Should there be an overshoot to compensate for the undershooting?

Scott Sumner: Okay, so let me first explain level targeting, which is what you're talking about here. And interestingly, Ben Bernanke advocated level targeting before he became Fed chair, not while he was at the Fed, but again, after being Fed chair is again advocating it. And I think the reason is that the Fed itself was resistant to the idea. He discovered once he got there.

Scott Sumner: But level targeting is you imagine on a graph paper drawing a line for the price level rising at two percent a year over time, two percent inflation. Now imagine that if you miss that line up or down, you try to come back to the line. So if you have one percent inflation, one year, three percent the next to get back to a two percent average. The idea of level targeting is to always promise to come back to that pre-specified trend line. And there are properties of level targeting that help stabilize the economy if it's done right, and that's a very important qualifier.

Scott Sumner: I think for level targeting to work, it has to be done in a very timely way. So, in 2009 when inflation fell to zero, it would have made sense to aim for four percent inflation the next year, to catch up. In other words, in that case, inflation would have averaged two percent over those two years. What you want to do, if you want to catch up, is do this during a period where it will help to stabilize the economy. And what do I mean by help to stabilize the economy? Well, if inflation is not going to be two percent every single year, you'd like it to be above two percent when it will help solve the problem of low unemployment and you'd like inflation to be below two percent during a boom period. So you'd want inflation to actually be counter cyclical, go against the business cycle. Low inflation during boom, high inflation during recession.

Scott Sumner: In fact, inflation in the US and most countries tends to be a little more pro-cyclical. It fell during the Great Recession and now that we're coming back to a boom, now inflation's finally getting back to two percent and might go above two percent. But yes, I do think there should be sort of a catch up period after low inflation. But you'd like that catch up period to occur when unemployment is high, like 2010, 11, 12, not 2018, 2019, 2020 when the economy is perhaps booming. That is actually destabilizing because when you have high inflation during a boom, it can lead to financial excesses, and when you have low inflation during a recession, it can lead to more unemployment.

Scott Sumner: So, by all means have some flexibility in the inflation targeting regime but do so in a way that is stabilizing. Low inflation during booms, high inflation during recessions.

David: And to be clear, you're still advocating price stability, it's just on average, you hit two percent. So you would have gone above two percent after the crisis and then below during an expansion. So, you're still getting on average two percent price level growth interest ...

Scott Sumner:            Yes. In fact, if you're worried about this, if it sounds kind of fishy, like you're cheating on the inflation target, it's actually more solid, more reliable than the current inflation targeting regime. Because under the current regime, when they miss, they can miss three or four years in a row and then nothing is done about it. So it doesn't have that much credibility then. When you do level targeting, you basically guarantee that in the very long run, inflation will average say two percent a year, because you'll always try to come back to that trend line. So with level targeting, you have a much stronger idea of what the price level will be 10 or 20 years out in the future.

Scott Sumner: I'll give you an example. Japan currently has a two percent inflation target which they adopted a few years ago. I think four years ago. They've consistently missed on the low side. A lot of people think they're going to keep missing for the next few years. So people in Japan can't really have any confidence about where the price level will be 10 years from now because will it keep going along at one percent inflation like the last few years or will they have their two percent target achieved? Well, that would be a very different situation 10 years down the road. With level targeting, the Japanese investors and citizens would have a much better idea of where prices will be 10 or 20 years out in the future even though a year to year there would be fluctuations.

David: Okay. So Minneapolis Fed president, Neel Kashkari, he's pretty public, he's in the media a lot, he had an op-ed in the Wall Street Journal yesterday where he argued that the Fed should pause its rate hikes and allow this inflation overshoot to occur to be symmetric. What I'm hearing from you is it's too late. The horse is out of the barn. If you're going to have that symmetry, you do it when the events actually occur. You can't make up for it many years later. Is that fair?

Scott Sumner: Yeah. I find it ironic that I'm hearing these arguments now. I was making the argument for higher inflation back in 2009 and 10 and the Fed was sort of resistant to, and you were making similar arguments. They were resistant to these pleas from people like us and now when there's talk of higher inflation, it really isn't, it's not clear to me how it helps the economy. The labor market seems reasonably strong. You can debate about where exactly the natural rate of unemployment is. But 3.7 is certainly historically a pretty good figure. We really needed the stimulus back in 2009, 10 and 11. I just don't see any justification for pushing inflation above two percent at this point in time.

Scott Sumner: If that's what they're thinking of doing, they really need to rethink their whole approach, the so-called monetary policy regime, how they think about doing policy over the business cycle and come up with an approach that's more stabilizing.

David: Okay. So others are having this conversation now as well regarding what should be the monetary policy regime moving forward. Lessons learnt and the number of prominent individuals have promoted different possibilities. Ben Bernanke has talked about a version of a price level target, which you've just described a few minutes ago. Larry Summers talked about a nominal GDP level target, which both of us are sympathetic to. And Olivier Blanchard has talked about having a higher inflation target, going from two percent to four percent.

David: I know you prefer the nominal GDP level target version. Why do that? So if any of these proposals out there are being discussed, why pick nominal GDP level targeting over the other ones?

Scott Sumner: Well, first, let me make a point about raising the inflation target. I want to be very clear on this. Raising the inflation target would definitely work in the sense that the proponents advocate. This is why they're proposing this. It would basically solve the problem of zero interest rates limiting monetary policy, what's sometimes called the liquidity trap. So, during the period of zero interest rates, central banks did have trouble hitting their targets. If you had a four percent inflation target, inflation would be so high that interest rates would almost certainly stay above zero. So that would solve one problem.

Scott Sumner: The reason I favor nominal GDP targeting is I think we could solve the problem with a lower cost. After all, higher inflation also does impose some costs on the economy. Because of our tax system for instance, it tends to discourage savings and investment. So I would like to solve this problem without just raising inflation to four percent from two when I don't think that's really necessary. What nominal GDP level targeting does is it gives the Fed an ability to deal with a crisis like 2008 far more effectively. I don't think interest rates would've fallen to zero in 2008. If we'd had that policy regime in place, I think it would have played out much like it actually did in Australia. So if you look to Australia, that's the sort of vision I have for what the US would look like under nominal GDP targeting.

Scott Sumner: There would be an occasional mild recession, but they would be less frequent and less severe than under the current policy regime. That's partly because we don't get thrown off by inflation signals with a supply shock of 2008. And it's also because of the level aspect, it gives monetary policy more attraction during a depressed period. Because as you promise to come back to the trend line, it creates more positive expectation.

Scott Sumner: I would use the analogy of putting sand under a tire on an icy road. I grew up in Wisconsin where we did this a lot. If you put sand under the tire, the tire can grip the road better on a slippery surface. With level targeting, the same tools, interest rate cuts, QE, whatever, are much more effective than under the current regime where if they miss their target, they just forget about it and go on from there. They don't promise to come back to the trend line. So, promising to come back to that trend line creates stabilizing expectations in a way that makes the economy more stable than otherwise.

Scott Sumner: And that's why a number, not just Larry Summers, but quite a few economists on both the left and the right I would add have been recently advocating nominal GDP targeting. And so it's grown quite a bit in popularity since you and I started talking about it back in 2008 and I think that's a good sign. That's a sign that a lot of economists are now beginning to see the appeal of this kind of policy regime as compared to what went wrong during 2008.

David: Okay. Now you have a very unique version of nominal GDP targeting that you would like to see done. You want to have the market effectively do monetary policy through what's called a nominal GDP's futures contract. Could you explain that to us?

Scott Sumner: Right. So this is sort of the market part of market monetarism and that's the label that is often given to us. So, I talked about targeting the forecast, like setting policy where the growth you expect a nominal GDP is equal to what you want to occur. Setting the forecast equal to the target. How do you do that? Which forecasts do you use?

Scott Sumner: Traditionally, you have a bunch of experts at the central bank try to forecast where the economy is going. But as we saw after Lehman failed, their inflation forecasts were wildly off course. The market forecast at least in that instance were much better. While market forecasts are not perfect, I think on average they're the best. So how can we harness market forecast? One way is to create a nominal GDP futures market and set monetary policy so that the market thinks that growth and nominal GDP will be equal to what is desired to occur. So if your target is four percent growth and nominal GDP, you want to set your monetary policy to a position where the market thinks nominal GDP will be growing at four percent.

Scott Sumner: So, that's the basic approach. There's many technical ways you can do this, some of which are sort of more extreme, more farfetched science fiction type proposals. I've kind of over time settle on a more pragmatic, less controversial approach that I think would be more acceptable and it would be almost as good as the pure system. And so, the proposal I'm talking about now is something I call the Guard Rails Approach. You might notice that when a truck backs up, sometimes there's a beeper that goes off when it's about to bump into something. And actually cars are like this too now, right?

Scott Sumner: So what would be an analogy for monetary policy that would warn the Fed when they're off course? I propose if the target were four percent nominal GDP growth, you have the Fed promise to take a long or short position on nominal GDP futures contracts at three percent and five percent. Those become sort of the guard rails along the highway. You're aiming for four percent but you don't want growth to slip below three or above five. So if investors think it's going to go above five and take a long position on the contract, the Fed is forced to take a short position. If investors think it's going to fall below three and take short, the Fed will take a long position on the three percent contracts.

Scott Sumner: The Fed is essentially betting that nominal GDP growth will be within that range. Now, if you think that's unfair to the Fed to make them gamble in these markets, here's my response. Well, if they think that the market is smarter than them, then they should just set monetary policy where the risks are balanced. People are taking both short and long positions in roughly equal amounts so the Fed can't really lose in that situation. If the Fed thinks it's smarter than the market and the market's betting that nominal GDP will fall sharply as the market would have been late 2008, and the Fed is thinking, well, we're much smarter than the markets, we know what we're doing, fine, you can bet against the market and then if you're wrong and lose a few billion dollars, you can go to the congressional committee later and explain why you thought you were smarter than the market and what went wrong with your forecasts.

Scott Sumner: But I think what the Fed would do is try to set policies such that they're not exposed to too much market risk and they would take these market signal seriously if all the investors were piling up on one side or the other, like clearly showing that nominal GDP is likely to be way too high or way too low. That's why I call this the Guard Rails Approach. It still gives the monetary authority discretion within the guard rails. They can sort of adjust policy a little bit this way or that freely. But if they go too far off course, there's a real price to pay and that would sort of tend to keep monetary policy on course I think.

David: Okay. So we're a long ways away from nominal GDP future contracts. That's fair, right? It's a moon shot, it's a great idea, great project. So in the meantime, how do we avoid the mistakes of 2008, repeating them. One of the proposals you have is an accountability project for the Fed. Makes them more cognizant of mistakes they've made without changing regimes. Tell us about that.

Scott Sumner: This is a proposal that I think is just common sense really, to have the Fed periodically say once a year revisit past decisions and provide a report to Congress as to whether previous policy settings were too expansionary or too contractionary in retrospect. And importantly, this report would include specific metrics that the Fed looks at. Like in the 1960s and 70s, the Fed might've said, well, in retrospect we were to expansionary because look, inflation's real high right now. Or in 2009, the Fed might've said, you know, in retrospect, a year ago we were too contractionary because we've gone into this deep downturn and inflation's fallen to zero below our target.

Scott Sumner: Some people wonder, well, would the Fed really be willing to admit mistakes. That's why the metrics are so important. If they just ask the Fed, did you do a good job? Of course the Fed is going to try to put a happy face on things. But they'll have to tell congress, like basically here's the metrics we look at, here's your dual mandate you gave us, here's how we interpret it, two percent inflation, high employment, which we interpret as an unemployment rate of x, say a four and a half percent or whatever. And here's the numbers and here's what we should have done in retrospect given all of those numbers. If it's not plausible, Congress is going to ask some questions clearly.

Scott Sumner: But I have actually a hidden motive for this, if I could just spend a minute on this. I'm actually not trying to do this in sort of punitive way. I actually think this would help the Fed, and especially the chairman of the Fed or chairwoman. The reason is that heading the Federal Reserve is sort of like hurting cats. You've got all these fed officials that are not on the same page in terms of what they want to do. You've got hawks and doves and so on. It's hard to get all the committee to agree on a common goal.

Scott Sumner: What this kind of process would do is over time, force the Fed to really think clearly about what metrics constitute success. And so, we would start to make the dual mandate, which is currently very vague, have more specificity associated with it. Here's an analogy that might help you understand this point. Ben Bernanke as an academic favored inflation targeting. When he first got to the Fed in 2003 and even ...

Scott Sumner: When he first got to The Fed in 2003, and even when he became Fed chair in 2006, the Fed wasn't really onboard with the idea of an explicit inflation target. They like to keep more flexibility. But Bernanke wanted an explicit two percent inflation target, so what did he do? He said, "Okay, here's what we're going to do. I'm going to have all of you people give me forecasts for key economic variables, every three months let's say. You're going to write down your forecasts, and importantly, all of these forecasts are going to be under the assumption of 'appropriate monetary policy' and these forecasts will extend out over many years. Not just a forecast for next year, but five years from now."

Scott Sumner: Well obviously if you have an inflation forecast for five years in the future under "appropriate monetary policy," that forecast really represents that person's view of what's an appropriate inflation rate, right? If they think inflation would be two percent five years from now, under appropriate monetary policy, they're telling you they think two percent is the right rate of inflation. Over time, as The Fed did these forecasts and the press observed them, the press noticed that most of the forecasts were clumped around two percent.

Scott Sumner: The press started to treat these numbers as an implicit Fed inflation target. The press were saying, "Yeah, we know you're not officially admitting your inflation target is two percent, but your long run forecasts are out around two percent, it's pretty obvious that's what you're doing," especially since Ben Bernanke, as an academic, had advocated that, right?

Scott Sumner: So once that settled in as pretty much common knowledge, by 2012 he was able to get The Fed to finally commit to an explicit two percent inflation forecast and partly he was able to do this in 2012 because he pointed out that unemployment was still high at that time, we were struggling to recover from the Great Recession, and an explicit inflation forecast would help the recovery by giving investors confidence that the Fed would try to boost inflation up to that target.

Scott Sumner: It was still running below two percent at the time, so he picked a good time to have this instituted, but he had laid the groundwork for this by getting them to provide these forecasts and commit. I think this sort of accountability, looking back at mistakes, identifying them and having explicit criteria metrics for what constitutes success would gradually nudge the Fed to think more about nominal spending as a single indicator of whether monetary policy has been too expansionary or too contractionary.

Scott Sumner: My hope is by making policy more and more clear, explicit, transparent, it would also become more effective over time.

David: Okay, we have some time left for questions. I know Scott has discussed a lot of interesting and maybe new ideas for some of you. So, we're opening up the floor to questions and there is a microphone here. Please-

Speaker 1: Please try to say your question a little loud, because perhaps this is not going to be as loud as the microphone they have right here, and maybe you can repeat the question for [crosstalk 00:49:21]. It might be a good idea.

Scott Sumner: Yeah, I'll repeat the question, sure.

Speaker 2: Hey Scott and David. I'm a big follower of your blog, Scott, The Money Illusion. One thing that wasn't talked about was over this time period, there was also this reassertion of Keynesianism during the Great Recession. So, my question for you is strategic. You've written a lot and I would agree, that basically fiscal policy doesn't work. How does this Keynesian overlay and consensus about liquidity trap and efficacy of monetary policy and, "Oh, the Fed is out of bullets when we hit zero," how does that impact change in the monetary regime? Larry Summers is now saying, "Hey, nominal GDP targeting's great" but he was also one of the guys saying, "We're out of bullets."

Speaker 2: When push comes to shove, are we going to be back in the Keynesian foxhole where monetary policy's "pushing on a string?" How do you think about that more from a communications and from an almost political economy perspective than from a technical perspective?

Scott Sumner: Right. One of my goals as a market monetarist is to basically avoid the need for fiscal policy. I personally don't think fiscal policy's very effective, and I think it's also wasteful. But at the same time, I want to just as an aside, give Keynesians a little credit on one point. I do think Keynesians were correct that there was a big aggregate demand problem in 2009 and '10.

Scott Sumner: I personally view this issue as the Achilles' heel of the right, if you will. Conservatives were too quick to dismiss the need for more aggregate demand, more stimulus in general. Where I think Keynesians go off course is putting too much weight on fiscal policy, which first of all I don't think is even politically feasible in our system of gridlock, but putting that aside, I don't think it's very effective.

Scott Sumner: And one reason I favor something like nominal GDP level targeting, is I think monetary policy would continue to be effective even in a difficult period like 2008. Let me use this analogy. If you look back through history and let's say you're looking at it from a right of center perspective like mine. The period where I think the left did best is where nominal GDP growth tends to be stable.

Scott Sumner: Say, like the '90s, right? If nominal GDP is growing at five percent a year, every year, then it looks like the capitalist system works pretty well. It looks like you don't need government bailouts and fiscal stimulus and all this stuff, so you often get even left of center political parties that are fairly market friendly in that environment. But when nominal GDP falls, due to bad monetary policy, like the Great Depression, like Argentina 2000, like 2009, monetary policy doesn't get blamed. It gets misdiagnosed as a failure of capitalism.

Scott Sumner: Now the left turns against free market economics and they favor fiscal stimulus, they favor bailouts, regulations, all sorts of other things, because it doesn't look like the free market economy is working well. Sometimes I believe conservatives are their own worst enemy. Sometimes conservatives will stick to in a stubborn way, to monetary policies that are leading to inadequate nominal spending, that will cause a recession, it'll make capitalism look bad, and then it'll open the door for more left wing ideas.

Scott Sumner: That's a long answer, but does that get at your question? I'm advocating this partly because I think nominal GDP level targeting would make it clear we don't need fiscal stimulus. It wouldn't be in that environment. If we go to zero interest rates, you really h very to think about calibrating your growth rate such that it's high enough to avoid zero rates.

Scott Sumner: I have this phase I use, inflation or socialism. I like to offer this choice to conservatives. If you really want super low growth and prices or tight money, interest rates will fall to zero and central bank balance sheets will get huge with QE. In countries with even lower inflation than the US, like Japan and Switzerland, the central bank balance sheet's like 100 percent of GDP, roughly.

Scott Sumner: Well, that's not desirable from a conservative perspective either, right? The government buying up this much in assets essentially. There's an interesting trade off here. Conservatives traditionally oppose socialism, government buying a lot of stuff, but they also tend to oppose inflation. The zero interest rate environment forces conservatives to confront a painful choice.

Scott Sumner: Sometimes you might have to have a little higher trend rate of inflation to keep interest rates above zero, so the central bank isn't forced to go in and buy up lots of assets, as they almost are forced to do at zero interest rates, to prevent a depression from occurring.

David: All right, next question.

Speaker 3: Yes hello. Sorry, first a comment then a question. First I would like to say that I think the idea of holding our Fed chairman or woman accountable is good, but I think the exercise you propose is pretty much impossible. We have no idea how things would have turned out had they done anything different. It's impossible to test that, so I think it really is impossible thing to try to do seriously at least. As micro economists, there's quite a bit of disagreement, so I don't know how you could take that to practice.

Speaker 3: Second, what I don't understand of your proposal for nominal GDP targeting is what you do in situations where you have something like stagflation. So, you have depressed growth, and then you just want to have depressed growth and inflation to meet your target. Is that your… How do you guarantee that that won't happen?

Scott Sumner: First of all, my focus is on monetary policy, but I would be the first to acknowledge that the biggest problem society faces are not monetary in nature. They're real problems, not nominal shocks. So, there are lots of things that monetary policy cannot solve and one of those is stagflation.

Scott Sumner: Stagflation is basically caused by a real shock, a decrease in aggregate demand. When that occurs, what you really want to do is just make it so that the stagflation problem, say like an oil embargo in the Middle East, or a bad regulation that hurts the economy, doesn't spillover and cause even more problems than necessary.

Scott Sumner: When you have a real shock to the economy, whatever it is, it's going to depress real GDP. How do you avoid that causing unneccesary collateral damage? The best way of doing so is to keep nominal GDP fairly stable and then what happens is as real GDP growth goes down, inflation goes up a little, keeping the total sum of the two about the same, that's nominal growth.

Scott Sumner: You do that to stabilize labor markets and financial markets, so that the real shock to living standards, that's inevitable from any kind of real shock, doesn't cause unnecessary unemployment and so on. Let's take the housing crisis, and think of it as an individual level. If you had accumulated a lot of debts and was having trouble repaying your debts and had a financial crisis within your family, you'd borrowed too much, would you think that a sensible solution would be, "Let's take a long vacation and stop working?"

Scott Sumner: No, what you'd actually want to do is maybe buckle down and work harder so you could pay off your debts. It's unfortunate you'd borrowed too much, but let's not make the problem even worse. What we did during 2008 is we had a situation where people had borrowed too much, and we made it worse by unemploying lots and lots of people. Whereas they actually should have been working harder to repay the debts they'd accumulated.

Scott Sumner: How do you do that? You do that by keeping nominal GDP growing at a pace where the labor market will stay in equilibrium, and the real shock doesn't have collateral damage on other sectors of the economy. Yes, it's unfortunate you'll have stagflation sometimes under nominal GDP targeting, but what I would say is any alternative, either lower inflation or higher inflation, would be even worse than keeping nominal GDP growing at a steady rate. It would be compounding the problem.

David: Got a question over here on the left.

Speaker 4: My question is framed with an eye towards the next recession. I'd like to know if there's an upper limit on how high the monetary base can get before we hit some sort of nonlinearity or singularity?

Scott Sumner: Yeah. Once I think the Fed has bought the entire world, then ... I mean, even I think Ben Bernanke and others have joked about this. You know, people used to talk about there's nothing Japan can do to create inflation, for instance. They would make this argument. Really, when you think about it, it's a silly argument, because if it were true, it would actually be wonderful news for Japan.

Scott Sumner: If there were nothing Japan could do to create inflation, that means the Japanese Central Bank could buy all of the assets of the world, Japan could become fabulously rich, they could all stop working in Japan and just live off their investments, and they'd have no inflation, right?

Scott Sumner: Just from that thought experiment, we know that obviously if they started down that road, they'd get a lot of inflation, printing money and buying up stuff. My first comment is as a practical matter, there's really no limit to what you can buy. Now, as a legal matter, governments might say, "Central banks cannot buy equities or various things" then you've got a problem.

Scott Sumner: So, what you really have to do there is you have to go to the policymakers and say, "Okay, here's some of the choices we face. We could have a higher inflation target, so we don't hit the zero bound and have to do QE, or we can hit the zero bound and go into a great depression. If that's what you want us to do, we can do that, or we can hit the zero bound and buy massive amounts of stuff, like in Switzerland and Japan, but we want you as Congress to think about which option you think is the least bad. Let's have an intelligent conversation about that."

Scott Sumner: Now, is that going to happen? No, in the current political environment, it's difficult to get Congress to agree on anything, but I mean conceptually, those are the sort of choices you have to think about. A higher trend inflation rate, a do whatever it takes, or just passive, "Yeah, we'll go into a depression, because we'll hit a liquidity trap and not do much." Those are I think the three options.

Speaker 1: I'll ask the last question then. You suggested there's a number of economists in both left and right that have been coming to an agreement that something like nominal GDP targeting is a good idea, the Fed should be trying to do more of that to correct from the mistakes of the past. Is there a path to get there that doesn't require legislation? In other words, under the current mandate of the Fed, could the chairman somehow write a set of rules or try to organize their interpretation of the mandate, such that they're doing something akin to nominal GDP targeting, as opposed to us trying to go through Congress and getting that done?

Scott Sumner: Right. I think there is. I think the easiest way to do it would be to basically first of all, let me remind you that nominal GDP is real growth plus inflation. You're putting equal weight on these two variables. That's very consistent with the dual mandate, because the real growth correlates with employment and the inflation obviously is the other half of the dual mandate. So, nominal GDP is actually a wonderful single variable to address the dual mandate, because it's already got both components embedded in it.

Scott Sumner: Now, in terms of the two percent inflation target, what the Fed would need to do is come up with a forecast for trend rate of growth and real GDP, let's say they forecast that trend at two percent a year, they'd add two percent inflation and say, "We're going to target nominal GDP at four percent." Now, what The Fed could promise is that they would periodically revisit their trend forecast and they would say every couple of years, based on new information, we will re-evaluate our estimate of trend growth and real GDP, and if it's revised to 2.3 percent, we will then raise our nominal GDP forecast or target to 4.3 percent.

Scott Sumner: It'll be the trend rate of 2.3 plus the two percent inflation. If they were to do that with periodic adjustments, that would keep inflation close to two percent. Now, David and I actually don't think it's essential to keep inflation at two percent. We think of nominal GDP as the real thing, like in the Coke commercial, and it's the thing that really is important for the economy. If you stabilize nominal GDP growth at a given number, it's okay for inflation to move up and down a little as real GDP adjusts over time.

Scott Sumner: But if you want to stick with that dual mandate, you would have a nominal GDP target, but gradually adjust the path as trend rates of growth in GDP evolve. But they don't evolve very much. The whole 20th century, basically we had a trend rate of three percent roughly. Didn't vary very much. It's probably a little lower in the 21st century, but we're not certain.

David: Okay. With that, our time is up. Our guest today has been Scott Sumner. Scott, thanks for coming on the show.

Scott Sumner: Thank you very much.

Audio recording provided by the LAITS Audio Development Studio at the University of Texas at Austin.

Photo credit: Spencer Platt/Getty Images

About Macro Musings

Hosted by Senior Research Fellow David Beckworth, the Macro Musings podcast pulls back the curtain on the important macroeconomic issues of the past, present, and future.