JW Mason is an assistant professor of economics at John Jay College and a fellow at the Roosevelt Institute. He joins Macro Musings to discuss his recent paper, *What Recovery? The Case for Continued Expansionary Policy at the Fed?*. JW argues that the recovery from the Great Recession was unusually weak, and as a result, the American economy is still not at full employment, leading to continually underestimated potential GDP.
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: JW, welcome to the show.
JW Mason: Well, thanks for having me.
Beckworth: So tell me, how did you get into macroeconomics?
Mason: It was a kind of roundabout route. I only took as an undergraduate at the University of Chicago, I only took one economics class and I said this is definitely not for me, I have no desire to ever encounter any of this stuff again.
Mason: And I spent quite a number of years doing policy work for the AFL–CIO and other labor groups, for the New York Working Families Party and a bit for city government in New York, and what I realized is if you want to have original ideas about the world and really think about things in a sort of deep way but you also want to be in the room where the decisions are being made, economics is maybe the only field or one of the most obvious fields that lets you bounce that, let you have that.
Mason: So I decided to go back to graduate school, and honestly from my point of view macroeconomics is economics, it's not a hard question. If you're interested, again, in sort of thinking about what's going on in the world and being part of policy debates and part of the political world, macroeconomics is where the action is.
Beckworth: Well, it's nice to hear you say that. When I was a grad student I remember once a fellow grad student ran into me in the elevator and I was first year in the program and we exchanged what we're doing and he was some kind of applied microeconomist and I said, "Oh I'm going to do macro." He looked at me and said, "I don't believe in macro." You still get that sometimes, but it's great to have someone so enthusiastic about macro on the show.
Beckworth: So let's talk about your article. As I mentioned it got a lot of play in the press lately, and its title again is *What Recovery? The Case for Continued Expansionary Policy at the Fed.* Let's begin by motivating your article, by taking about the recovery. As the name implies. Tell us, has the recovery been unusual in any means?
The Unusual Economy Recovery
Mason: Well yeah, it's been unusually weak on almost every dimension. Exceptionally slow recovery by the conventional metrics in terms of how employment and output, but also extremely slow investment growth, which a lot of people don't realize. A large fall in employment population ratio, which really hasn't recovered at all since 2008.
Mason: And output growth, although it has been positive, has not closed the gap at all, either with the sort of linear trend that was a pretty good guide to GDP up until 2008, or with forecasts made before the recession.
Mason: In other words, if you compare GDP today with what was almost universally predicted, both by the CBO and by private forecasters prior to 2008, we're as far from those earlier predictions as we were at the bottom of the recession. There hasn't been any closing of that gap at all and there certainly is no case since World War II where you could say that, where there's been this total failure to close the gap with the pre-recession trend or pre-recession forecasts.
Beckworth: That was fascinating to think about that, that the current deviation of real GDP from its trend path is unparalleled for length and persistence. And typically there has been some kind of reversion to the trend path. So it's fascinating to see that.
Beckworth: You mention in your article, there is a whole academic debate over the stochastic shocks and stuff like that, but it's a good starting point and we have to be able to explain why has there been a persistent deviation from that.
Beckworth: In going back to the bigger point, this slow recovery. Now, I've done some work, I've tinkered around and one exercise I just did some autoregressive models, purely nice statistical models and if you plug in, for example, things like output gaps, even as flawed as they are, we'll talk about in a minute, any number of measures, they all show a much more robust recovery if you just, you plug in the data, let it run say from June 2009 when the recession ends, all of the statistical models would have expected forecast at a much quicker recovery, kind of you have a severe, sharp decline, you expect to bounce back to kind of compensate for it, takes you back on path. But none of that happened.
Mason: Yeah, I think some people sort of want to dismiss this and say, "Well, deviations are always persistent, there's no reason to expect a return to trend." I think that's actually a very problematic view. First of all just as a historical fact, in the US, in fact there has been a pretty stable trend and there have been, in the past, periods of faster growth following downturns. That's just a fact as far as the US is concerned, not in every other country in the world.
Mason: But secondly we have a notion that we have a long run growth path that's dictated by the supply side fundamentals, the technology and population and so on, and then we have short run phenomenon that are accommodated by demand and monetary phenomena.
Mason: If you just sort of say well of course deviations are persistent, you can't tell that story. Either you go the sort of real business cycle where everything that happens in the economy is about supply side factors and recessions are due to some technological regress or people's sudden desire not the work, which I think is outside of a handful of academic departments, is very hard to take seriously.
Mason: Or you can argue that, well, monetary disturbances, demand disturbances, or normally permanent in their effects, which is maybe not quite as crazy but is not something that most people want to argue. But if you're not going to go one of those two routes, then you should expect some reversion to the trend after a downturn or after an inflationary boom. So the fact that we haven't seen that I think is not something you can just dismiss.
Beckworth: Yeah. And you point out in your paper, I want to spend some time talking about this, that forecasters... So not only do we have this deviation that seems unusual and you have to put an explanation out for why it's there, but you also note that forecasters did not anticipate this deviation as well.
Beckworth: And you kind of use the CBO as a benchmark but I think any set of forecasters would have come up with this, and I've seen other research that has highlighted this as well. But forecasters did not anticipate the persistent below trend growth that we've seen. You mention the CBO did anticipate some of the decline in the labor force participation, but did not anticipate the entirety of that.
Mason: I should stop there because it's a little bit tricky. The standard CBO, they publish these long run economic forecasts on a regular schedule. They don't actually explicitly forecast labor force participation. What they forecast is employment growth. They did forecast the slowdown in employment growth, you can't determine from the material they publish whether that is actually a prediction of slower labor force participation or slower population growth.
Mason: And it's not actually entirely clear which of those is behind that forecast but it is the case that they forecasted a deceleration of employment growth comparable to what we've seen.
Beckworth: Yeah, so maybe the bigger point here is this. Something that's structural like changes in demographics should be anticipated, right? I think that's the point you were making in your paper, is that if there are these changes occurring they should have been seen going into 2007/2008.
Beckworth: So the sudden collapse is explained by something other than these anticipated events. And you have some numbers that you throw out. Well it's the CBO's numbers, and correct me if I'm wrong, the CBO anticipated 7% of the 15% below trend path of real GDP by 2016. Let me make that more clear. So today, real GDP is 15% below its trend path going into the crisis.
Beckworth: The CBO forecasted or anticipated 7% of that 15% decline, and you attribute the difference, and I think some of the 7% as well, to the fallout from the Great Recession, is that fair?
Mason: Yes, that's right. It's the lingering... I suggest that we can, and obviously a lot of other people have made this suggestion as well, that this is the lingering effects of the Great Recession, that this is the damage from the Great Recession that we have not recovered from.
Beckworth: Yeah, and you've mentioned there's a series of papers that have made a similar argument. So Larry Ball has a 2014 paper, then he has a paper with [inaudible], Larry Summers in 2014, and then [inaudible] in 2016. And they all kind of use this, I want to say identification strategy, this technique where they kind of control for the anticipated change in potential GDP from demographic factors and then try to look at the rest of it being attributed to demand side. And they find that a lot of the shortfall is due to demand forces.
Beckworth: Let me ask this question here. So, for the part the CBO did not anticipate, where does most of that come from? The decline since the crisis.
Underestimating the Trend Path of Potential GDP
Mason: Well, first of all if we compare it to the pre-recession trend, the linear trend, about half of the shortfall is productivity, about half of the shortfall is slower employment growth, labor productivity here. We say the CBO anticipated the slowdown in employment, that's true, but I think we have to be a little careful about that because many people say well this is the natural demographic result of population aging.
Mason: But that's probably, doesn't seem to be what's going on. In fact, the employment, the population ratios in the older age groups have actually risen over this period. So we don't see the sort of... and the big falloff in employment is at the lowest age groups, the youngest groups. So, it's hard to say that this is the sort, even though the numbers more or less match up to what the CBO predicted, that doesn't mean that it actually is the phenomenon that they predicted, and it doesn't mean that it was a phenomenon on the base of demographics. Numbers might happen to match but coincidentally, and I think when you actually disaggregate a little bit that's more of what we see.
Mason: There's some other interesting results that actually aren't in the report but we're going to include in a follow-up, which is if you break down employment by education or by race, what you see is that for more educated workers and for white workers, the behavior of employment looks more demographically predictable. It looks more driven by the changing age structure. You can explain more of it simply by assuming that within each age group people's employment rates are fairly stable. For the highly educated white workers, that works pretty well.
Mason: If you look at less educated groups and you look at non-white workers, particularly African-American, the falloff in employment is much greater than you would predict on the basis of demographics. And to me, the story that sort of fits there is a kind of queuing story, there's a line for employment.
Mason: Some people are at the front of the line, some people are at the back of the line, and when labor demand is weak it's the people at the back of the line who don't end up employed. Now, the people at the front of the line may be employed in less productivity or less lower wage jobs than they might find in a strong labor market but they are still going to be employed.
Mason: Again, that sort of thing makes me say that even though the headline number is not too far off from what the CBO predicted, it would be a mistake to go from that to say okay, this is simply the predictable result.
Beckworth: Okay. So what you're saying then, again the CBO, the trend line for potential GDP continued on that path since the crisis, and compared to real GDP today there's a 15% gap, the CBO anticipated 7% of that but what you're saying is that 7% should not be just dismissed purely as an inevitable outcome. It might just be coincidental that some of the demographic forces line up with that number. Is that right?
Mason: That's right, yes.
Beckworth: Okay. And along those lines I want to mention a quote I thought was real interesting, because you do flesh out a little bit in the paper about education versus aging and on page 41 you have this quote, you have, "While it is true that the American population is getting older, it is also growing more educated. All else equal, older people are less likely to be employed and educated people are more likely to be employed. Over the past two decades, these two factors have essentially canceled one another out."
Beckworth: So in other words it's kind of hard to really know for sure what's driving that decline in the employment part of potential GDP that's gone down, is that right?
Mason: That's right. And there's another point I really want to emphasize here, which is that this is not simply a matter of data or statistics, you have to make some choices. Statistically, if you just look at the statistics, you say older people are less likely to work, less educated people are less likely to work. So you might then just include both of those as demographic variables.
Mason: If you do that then you say demographics doesn't predict any fall in employment population ratios, because those two factors are pointing in opposite directions and they're roughly equal magnitude. You can make a choice and say, "Well, I don't think education is really a demographic factor in the same way so I'm not going to look at that but I am going to look at age." Then you get a large demographic prediction, but the data doesn't tell you to do that, that's a choice you're making.
Mason: Race also, because even though there isn't a big employment rate difference by race, the age, because of the different age distributions of the populations by different races, including race in an exercise like this, will get you a larger demographic effect. I'm not sure that's legitimate, I don't think that it makes sense to say, "Well, African-Americans are just less likely to be employed, that's just a fact about them."
Mason: I don't think that's a legitimate way to proceed, but that's again a choice you have to make when you're doing this sort of exercise. Then a third choice is where you start your analysis, because there was a big fall in employment to population ratio already between the peaks of the previous two business cycles, in other words between 2000 and 2007.
Mason: That is not really explained at all by population aging. Now, if you just start your analysis in 2007, if you take that previous fall baked in, the demographic component is going to look bigger. If you say, let's start our analysis at the high point of the employment population ratio, which is the baseline for what full employment might look like, then you're going to see a larger, non-demographic component.
Mason: So there's a lot of, when you're making these... Everybody kind of does this analysis the same way, the technical side of it is not controversial but there's a lot of choices you're going to have to make about exactly what to put into that analysis that are going to change the results that you see.
Beckworth: Okay, so we have all these issues that make it complicated to know for sure why potential GDP's gone down. I think what you're pointing out is that you can't be certain that it was inevitable, that it's all structural, and that's just that 7% of the 15% decline. So some of it could be due to the weak economy, keeping people out of the marketplace, out of labor search.
Beckworth: I just wanted to kind of make this real, take this back to some of the discussions we had over the past five, six, seven years. There was a lot of talk, 2010, 2011 that well, hey, this is all structural unemployment now. We've gotten past the cyclical portion, this is all structural, and yet we saw that unemployment continued to go down.
Beckworth: We also see, for example, that the working age or prime age employment to population ratio continues to go up, this far out. So looking back, I would like to think that people who made the structural claims would have some humility and say, "Look, you guys were right."
Mason: That's exactly right, yeah. It's really striking how people who were saying, "Well, it's back up to 6%." I haven't seen any soul searching at all from the people who said that, even though if we'd taken that view seriously, we would have two, two and a half million people unemployed today if the estimates, which were up around CBO but I think also a lot was coming out of the Fed in the immediate post-recession period… gone up by more than a point.
Mason: If that had been taken seriously we would, and actually they hit that target, we would have a couple million more people unemployed today. That's a lot of misery that would have been on people's hands, and yet I don't see anybody saying, "Oh my god, I was wrong. I'm glad nobody listened to me."
Beckworth: Yeah. Indeed. Now, again, the key point in your paper that you make is if you look at the official, potential GDP measure, it's been going down and the output gap, the difference between where our economy is and where its potential is, at least the CBO's potential, it's gotten smaller. It's gotten smaller. Which textbook theory would say we're getting closer to full employment, but what you make a point to stress, it's gotten smaller because potential GDP has shrunk, not because real GDP has rapidly grown.
Beckworth: And why has potential GDP shrunk? Again, part of the discussion would be, well, it's demographics, there's some structural components. But you also look at another part of the conversation... So, we've already discussed that there's some questions about the demographic, the structural side, but there's also some big questions about productivity growth. So a big part of the decline in potential GDP is a slowdown in productivity growth. So tell us the standard story there and what's problematic about it?
Potential GDP and Productivity Growth
Mason: Well, I think the models, the textbook models that we all came up with basically say productivity is exogenous, it's determined by technology, and that's coming from somewhere but it's not something that we're trying to explain within our model, and it's particularly not something that's sensitive at all to demand conditions.
Mason: So if we see slowdown in labor productivity, okay, some of that is maybe less capital, less investment, but mainly it's, in the standard accounting, which I'm skeptical of for other reasons, but in this case it's mainly total factor productivity, which is supposed to just represent technological progress, which is coming from just the possibilities of improving our ability to control the world, which are not really dependent on macroeconomic policy. So it's just out there.
Mason: The problem... Well, there's a lot of problems with that story. One, which Jason Furman says very, very, very bluntly is this is telling us that somehow by coincidence, we had the worst recession, the worst financial crisis of the post-war period. At the same time, we had this huge slowdown in employment growth coming from demographics, and at the same time as by coincidence we had the worst slowdown in technological progress of this period. It's not impossible, but it seems like it shouldn't be the first explanation you look for.
Beckworth: Yeah, so it's highly unlikely all three of those things happened together exogenously.
Mason: Right. If you could find a story that would link them up, that would be nicer. Another problem is if you think about... If you look at one piece of it you can say, you can maybe make that make sense. But one of the things I say in this report, and other people again have pointed it out, is that you really should be looking at the whole macroeconomic picture.
Mason: A slowdown in labor force growth is a negative supply shock. A slowdown in technological progress is a negative supply shock. If you have those, what you should be seeing is inflation, you should be seeing people trying to maintain their spending, their spending growth, in the face of reduced productivity capacity of the economy, spending runs ahead of production and you get rising prices, or you get higher interest rates to stop that. Obviously, the exact opposite of what we've actually seen.
Mason: So it's hard to understand how you can have these two negative supply shocks and yet still have a situation of chronic, weak demand, or certainly not an inflationary situation. Whereas if you can tell a story in which the slow productivity growth and the slow employment growth are the result of weak demand, the result of a negative demand shock, then you don't need that highly unlikely coincidence and you don't have this puzzle of why aren't we seeing the inflation that our standard models predict and the result in reaction to negative supply shocks.
Beckworth: Well, let's talk a little bit more about productivity and how it could possibly be influenced by demand. And I know this will be probably one of the more challenging ideas for our listeners and for anyone, for me too, is to look at productivity, because as you mentioned, and most of us probably think, productivity growth is largely exogenous or at least the long term developments, right? That are driven by policy changes over many years, education, technological innovation.
Beckworth: I think those are important, but what you're saying is, and I think this is maybe where, maybe people will get confused, is I think that's still true, I do think we can't predict technological innovations. We have long term policies that affect how productive we are.
Beckworth: But, you could have a sudden drop in productivity that might be endogenous, at least within a certain range. And maybe in a minute we'll talk about your three different measures of potential GDP, and I also want to say upfront I don't think you're arguing that if we ran demand nine to nothing all the time, we'd get massive productivity gains that make us look like futuristic humans with great devices and stuff. You're saying there's a certain gap in productivity that could be filled?
Mason: Right. Eventually you hit supply constraints, of course it's not a hard wall but you start seeing rising prices, rising wages, which is not necessarily a bad thing but you should see it eventually and bottlenecks. And you probably, I mean this is maybe getting beyond the scope of the paper, but you also see a certain breakdown of the market mechanism. Very high demand economies are probably inevitably more centrally planned economies at some point, simply because the ability of the market to manage very large reallocations of resources in short periods is limited, and the classic example is obviously the wartime economy, so the US and other countries during World War II.
Beckworth: So what you're talking about here is a slowdown on productivity, it's endogenous but it's unique in the sense that there is a ceiling to how much you can generate by revving up demand. But let's talk about that, let's talk about the potential channels or links between more robust demand growth and productivity growth. You listed a number of them but let's just give me a few examples, how would it work, if we're able to get to get demand growing at a faster pace that we might see productivity pick up?
Mason: Well, a classic example is the 1990s, we think of this, this was the last period of really rapid productivity growth in this country. It was also the last period of rapid wage growth, particularly at the lower part of the distribution.
Mason: Now, most people probably see those as independent or as productivity growth being the driver in wage growth following that. But there's an argument, I actually made an interesting paper by the Economic [inaudible], but other people have made it as well, that it was the wage growth that drove the productivity improvements and not vice versa.
Mason: Here's an example of how that can work. One thing that you saw in the '90s, which was unusual, is very large productivity improvements in retail and other service industries, particularly retail, which is a sector that is generally sort of impervious to productivity improvements, certainly compared with manufacturing.
Mason: Now, the interesting thing is if you look at what specifically they were doing to the extent where we can see that you see things like more comprehensive use of barcodes, electronically monitoring shelf levels, sales-based ordering, more use of electronic fund transfers, a lot of improvements of supply chains.
Mason: If you ask how many of these technologies were available in the '80s or even in the 1970s, almost all of them were. There was the ecommerce thing which was really new but almost everything else, the technologies existed but they weren't widely used. Why were they suddenly adopted so widely in the late '90s?
Mason: Well, I think there's a very plausible argument that rising labor costs suddenly really strengthened the incentives for firms to raise productivity in ways that hadn't existed during the period of relatively weak wage growth that had come before that. I think you can point to that in a lot of other cases, that rising wages are a very strong spur of productivity boosting innovations.
Mason: The funny thing about this is if you make the same argument in the case of a minimum wage, it gets a totally different reaction. If you say, "Well, if you raise minimum wage to $15 an hour, everybody at McDonald's is going to be ordering from kiosks and there won't be anybody behind the counter anymore." Maybe, but if that's true, what you've just described is an increase in labor productivity, and somehow people who are perfectly happy with that story, when we're talking about a minimum wage, think that it's something weird or counterintuitive when you tell the same story in response to a wage rise that's induced by low unemployment, even though logically if the story applies in one case it really should apply just as well in the other.
Mason: So I think if we change our frame a little bit we realize this isn't such a strange thing. There's also, a big part of the story is probably shifting from less productivity to more productivity occupations. Again, if you go back to World War II you had a really vivid example of this with a lot of people who were drawn out of agriculture and low productivity service occupations going to work in war industries in manufacturing.
Mason: And I think it's easy to imagine similar cases here where high productivity, in a situation where labor is really scarce and wages are rising, higher productivity sectors are going to outbid low productivity sectors for scarce labor and that's going to show up as an increase in aggregate labor productivity.
Beckworth: So I think a big part of your story then is that there's going to be increased investments for a number of reasons. Some of them you've outlines. There's a number more in your paper, but your argument is, again, up to a certain point extra demand growth will stimulate more investment spending, which in turn will lead to capital formation that will make us more productivity and put productivity back on that trend path as well, is that right?
Mason: Yeah, that's a major part. I think there are channels that don't necessarily require or involve higher... But yes, definitely, higher capital expenditure, higher investment is a big part of the story. And we have very good reasons to think that strong demand and more rapid income growth and higher wages will have the effect of boosting investment.
Beckworth: One of the arguments you make for the productivity story being made, that it's demand related, is that you show a graph, which is interesting, and you look at different sectors of the economy, and you show that all of them have experienced productivity slowdowns, or most of them have experienced productivity slowdowns, and that's hard to explain with a supply side shock. You might expect a few of them but your argument is that since all of them have had a decline, let me rephrase it. Most of them had a decline in productivity growth, an easier story to tell is demand, is that right?
Mason: That's right, yeah. You would think of innovations as arriving in particular sectors because technologies are generally useful in particular sectors, and you would also think that if what we're seeing is simply the end of the exceptional '90s boom, that the slowdown would be concentrated in the sort of high tech sectors that were the big beneficiaries of that boom.
Mason: The fact that we've seen slowdown in a lot of sectors across the board, is much easier, in my mind, to fit to a story of weak demand and slow wage growth because those are conditions that are really shared across the economy.
Beckworth: You know, along those lines there's been a lot of discussion lately about the natural interest rate or some people call it r*, which has fallen a lot and it's really, really low, and maybe it's one of the reasons the Fed believes it can't raise interest rates.
Beckworth: I do believe the equilibrium interest rate has fallen, but... and I also acknowledge that real interest rates have been turning down for a long time, but if you look at these estimates of these natural rates, they take a sudden fall from 2007 to 2008. Just like what you mentioned before, all these things that suddenly happened, productivity suddenly falls, investment spending suddenly falls, employment suddenly falls. All these things suddenly fall, so one famous example, one well-known example of the natural interest rate is the Laubach-Williams measure of r*, and this is John Williams of San Francisco … and his coauthor created this measure that's often used, often cited.
Beckworth: And it goes from the end of 2007, it was just a little over 2%, and by the end of 2008 had fallen about .3%. That's a sudden drop, given the size of the real rate historically. That's more than just a trend decline, that's a sudden drop, and again if you look at that I think in conjunction with the story you're telling, it does seem to suggest a business cycle story, a sudden collapse in demand, as opposed to something structural that has permanently lowered the equilibrium rate.
Mason: That's right, that's right. I think if you look at a lot of different dimensions, the last 10 years look much more like an extended business company downturn than the phenomenon that we associate with a supply side driven slowdown. Because obviously one of the things that you get in a business cycle downturn is a period in which the neutral rate is much lower, where you need that stimulus of low rates to return you to your long-term path.
Mason: I think it's much easier to tell that story. Now, I should say I'm enough of a Keynesian that I'm skeptical of a story in which there's a unique association between different interest rates and different levels of economic activity. I think it's quite possible that if you could get the economy back up to something more like a full employment level, back into what people used to call a high pressure economy, which I guess is a term that's getting revived a bit now.
Mason: That at that point you might be able to sustain full employment with significantly higher interest rates than we have now. The failure to really restart rapid growth is sort of perpetuating the situation where the Fed feels it needs to hold interest rates low. But on the larger point, absolutely I agree.
Beckworth: Yeah, so let's transition to your discussion on what exactly is potential GDP, because one of the critiques you'll get, comments I'm sure I'll get from the show is yeah, sure demand was important in 2008/2009, but how could it go on for this long? How could demand cause problems for this long? At some point natural rate forces will kick in at some point, prices adjust. This is where the idea of hysteresis comes in, correct?
Mason: Right, that's right.
Beckworth: So explain to us why it could be the case that demand shocks could have such persistent effects on the economy.
The Economic Impact of Demand Shocks
Mason: I think the idea is that again, we can sort of distinguish and again, I really took this from Jason Furman in conversations that I had with him in the later stages of writing this, but that's it's helpful to distinguish a short run potential from a longer run potential. Short run potential is the level of output at which you do start seeing some rising inflation, rising wages, some sign of supply side constraints, supply constraints.
Mason: But it doesn't necessarily mean that at that point you've reached the limit of what demand can do because there's an important sense in which supply adjusts to demand in which people's decisions about whether to participate in the labor force or not and businesses and the capital stock adjusts in response to demand conditions.
Mason: Actually, another person who I wish I'd quoted more in here, Narayana Kocherlakota, the former Minnesota Fed chair/president, has written about this. It's very nice, a couple of very nice posts where he says if business investment is depressed because of weak demand then you may have a situation where in the short run your full employment and full capacity level is going to be lower.
Mason: But if you maintain conditions of overfull demand for long enough for business investment to catch back up then eventually your full employment level, of employment, will rise back up to where the previous trend.
Mason: So the two dimensions here, I guess, are one, the labor force dimension, that once people have dropped out of the labor force, given up, become discouraged, they don't reenter immediately, and so it takes an extended period of overfull employment of a high pressure economy of rising wages, which may involve some degree of above target inflation to draw those people back into the labor force, but once you do you then, your wage growth will level off and your inflation will probably fall back to target.
Mason: That's probably the most common version and then the other version is that to the extent that you induce higher investment and other adjustments that raise labor productivity, you again can transition from a period of overfull employment to a period of full employment and on target inflation by bringing aggregate supply back up to the level of aggregate demand.
Mason: And so distinguish that, it's useful to think about the short run potential at which you start seeing more rapid wage growth and inflation from the medium run potential which is kind of the most you can reasonably hope to get from demand side policies, which is probably higher at least today.
Beckworth: Yeah, and just to reiterate this point, what you're arguing is that there is actually some ceiling, there is some potential real GDP, and when we hit that point demand policies can do nothing, but we're not there yet, we're stuck at the short run potential GDP, right?
Mason: That's a little stronger than I would put it.
Mason: I think for policy purposes we need a target, we need a definite target.
Beckworth: All right.
Mason: But in the real world there's a continuum, and as you have a higher and higher pressure economies, you have lower and lower unemployment, you have more wage pressure, you're going to need institutional adjustments to make that work. But I'm not sure there's a hard line, you're just going to have-
Beckworth: Okay, so there's a lot of uncertainty, for sure. That's part of the reason we have this debate, right, is no one knows for sure where potential GDP is, it's this unobservable latent variable.
Beckworth: The point is I guess, I'm trying to make this where it be more palatable to most macroeconomists in the sense that at some point they'd say look, there's only so much macroeconomic policy can do, there's only so much fiscal monetary policy can do to get the economy humming.
Beckworth: You acknowledge that, right? You're saying, yeah, I agree with that, it's just that we haven't hit that point yet, and we're being deceived by misled into believing potential GDP is much lower than it actually is.
Mason: That's right, that's exactly right. And from a policy standpoint it doesn't matter that much if the true output gap is 5% or 10% or 15%, because the policy prescription is exactly the same in all three cases, which is that we should not be raising rates right now and we should be thinking seriously about socially useful ways to have the federal government spend more money or collect less taxes.
Mason: So it doesn't... In a sense we can agree that there is a limit somewhere and we don't have to argue a lot I think right now about exactly where that ceiling is, if that we are right now substantially below it.
Beckworth: And maybe this is also a product of the fact that we don't have these severe recessions very often. In other words, the term potential GDP in normal recessions, normal periods, is adequate. We don't have to get too much in the weeds like we are now.
Mason: That's right. You could argue that because demand management policy, more recently carried out by the Fed, has been fairly successful. We don't have the hysteresis problem, we don't have this divergence of short run and longer run potential because we don't allow demand shortfalls to persist long enough to have this effect on supply. But now that we have, it's more of an acute problem.
Beckworth: Yeah. So maybe one other point on this, if we go back to the late '90s that you were just mentioning. At that point would you say the economy was at its medium run potential?
Mason: Right. I think that's, and this is another disagreement that people have. I think that's a better baseline. I think if we want to ask how much can demand do for the economy, that's the period you should be looking at. I think there's a strong case that already in the 2000s expansion we never really got back to that, that there was already a persistent demand problem which got much worse, of course, after the more recent recession but that we should not regard conditions during the 2000s as representing longer term potential output or full employment.
Beckworth: Okay. There's an example I like to use and I used it recently in the podcast with Scott Sumner, I interviewed him. But I like to think of hysteresis in this light. You might be someone who goes to the gym regularly and maybe you can bench press 400 pounds, you're a young buck, you're strong, you're great at this.
Beckworth: Then maybe you get kids, you get married, you don't go as often, you go back to the gym and all of a sudden you can only bench press 100 pounds, right? Your potential has gone down for how much you can push but that potential has gone down not because you are innately weak but because you haven't been exercising.
Mason: That's right, that's an excellent example and the lesson here is really the same, which is if you do that and you say, "Ah, I'm old now, I haven't got it, maybe I never had it, did I ever really get to 400? I don't know." It's hopeless. Well then you are going to validate that, that will be a self-fulfilling prophecy, you're never going to do the work that will get you back to where you were.
Mason: So this I think is a very good analogy for where we are now. If we face this sort of short run inability, if we take the fact that we can't bench press over 100 today, as a sign of where our innate capacities are, we will never take the steps to get back to where we were.
Beckworth: That's right. Let's move to where we are today and you have a nice list on page 60 of a number of symptoms and I'm going to read them off, low and falling inflation, low and falling interest rates, slower growth of output, slow growth of employment, low business investment, declining share of wages and income, slow growth in labor productivity growth. And again, you say look, this is where we are today and this screams demand shortfall, is that right?
Mason: That's right, that's right. Everything there.
Beckworth: Go ahead.
Mason: I was just going to say everything there is what we would expect to see in a period of slack demand, a business cycle downturn. And you can tell, this is the key point, you can tell a supply side story for any one of those pieces individually but it's very hard to explain the combination of them in a story that doesn't involve persistent, weak demand.
Beckworth: Yes. So here's my question to follow up to that observation. Why have policymakers been so timid? And particularly the Fed, I think there's a limit to what fiscal policy can do. Even if you believe fiscal policy can pack a punch, the reality is politically it's hard to get anything done with it.
Beckworth: But the Fed, they have the option, they could be more aggressive if they wanted to. So what is your takeaway? Why has the Fed been reluctant to really put the pedal to the metal and allow the economy to run a little bit hot?
Explaining the Reluctant Nature of the Fed
Mason: Well, that's an interesting question. I think it's very hard to avoid political answers in this. Obviously with fiscal policy as you say, but I think on the other side too, I think there's a certain attitude which was maybe a little more pronounced a year ago, which is that acknowledging that we face a large demand shortfall and a level of GDP, a level of output that's well short of potential, kind of implies a rather negative judgment on macroeconomic management over the past decade. I think for people who were involved in that, that's a direction they would rather not go in.
Mason: So I'm not sure how major that factor is but I think that's a factor. I think that there's a question of how effective conventional monetary policy can be in this situation. Well, obviously conventional monetary policy faces some rather hard limits. So then if you're going to take this problem seriously and you're going to expect the Fed to solve it, you're asking the Fed to go well beyond the kind of policy tools that they're used to or that they're comfortable with, and I think there's very natural reasons why people are reluctant to do that, not necessarily bad reasons.
Mason: It's obviously the case that the more... that almost any form of unconventional policy that you can talk about that would potentially address this situation is taking the fed beyond its neutral, technically world, and involving it in more political judgments in stuff that looks more like fiscal policy and making more determinations, not just about the overall level of activity in the economy but the direction of that activity favoring certain kinds of firms or certain kinds of spending over others.
Mason: I think it's sort of understandable that they would rather not go in that direction and so they're going to be more attracted to a story of the economy that doesn't suggest that they should be.
Mason: Now, there are other aspects of this and I think... I sort of feel I have to bring them up.
Mason: Wage growth, there's no question that it's just a logical necessity if the wage share in the economy is going to rise, there has to be a period in which real wages rise faster than productivity. Or, putting it another way, nominal wages rise faster than the sum of productivity growth and inflation and then the Fed's inflation target.
Mason: Now, there are a significant number of people in the monetary policy world, I think more so in Europe than here, but some here who regard that as unacceptable, that wage growth above productivity growth plus 2% is a problem that needs to be prevented.
Mason: If you take that attitude, well first of all you've created a ratchet in which the wage share of national income can fall but can never rise. But secondly, you're going to be reluctant to see a period of overfull employment, a period of a high pressure economy in which wage growth really accelerates, which is just a necessary part of any period of returning to trend, even if it's temporary, you're not going to draw people back into the labor force, you're not going to create the strong incentives for productivity raising innovation unless you have a period of rapid wage growth.
Mason: And if there are people for a variety of reasons having to do in some part with where they come from and who they talk to, who find rapid wage growth unpalatable, then there's going to be a lot of resistance to that kind of policy, even if it's what orthodox theory suggests they should be looking towards. So I think if that has a political economy, we can't pretend that it's a purely technocratic institution.
Beckworth: So, does that explain the low inflation, then? I mean it's persistently undershot its target and it has a 2% target since 2012, and as many people know it implicitly had one before then. So these are the reasons why it consistently undershoots its inflation target as well?
Mason: Well yeah, I think the Fed has a... Let's put it this way, the Fed has a loss function which is not necessarily the loss function that we would agree on democratically if we took a vote on it. They put a higher weight on avoiding overshooting inflation than on undershooting or overshooting unemployment.
Mason: You can't conduct policy without some kind of judgment of that kind and based on their behavior, they put a higher priority on overshooting their inflation target, than missing on other dimensions.
Beckworth: Well, I'm dying to hear from you. What are your thoughts on QE? Would QE make a big difference?
The Impact of QE
Mason: That's getting a little beyond my comfort zone, to be honest, but I think the answer is no, I think that... Here's the puzzle to me, and I really don't understand. I read a lot of Keynes, and Keynes wrote back in the '30s and other American Keynesians picked this up in the '50s and '60s that if you had a problem like this, the solution was simply to Keynesian target a range of interest rates rather than just the overnight rate. That you just announce here's our target for the five year, 10 year.
Mason: Well, the Fed did that during World War II, that was explicitly their policy but we didn't get that this round. Instead we got these large purchases of longer bonds with the goal of holding down rates some amount, but without an explicit target. I think the failure to announce an explicit target greatly limited the effectiveness of that policy because when you announce a target then you've basically got the whole market pulling behind you. The actual scale of purchases, we know this. The scale of purchases the Fed has to carry out to set the overnight rate is trivial because they announce a target and the market moves the meet that target.
Mason: If they had done the same thing with longer rates, they could have moved them much more strongly than they did and for various reasons they didn't want to do that so they adopted this policy of buying bonds in the hopes that that would lead to a movement in rates but without an explicit target, which I think was much, much less effective.
Beckworth: So if you had been Fed chairman starting in 2007, let's say, what would you have done?
JW as Fed Chair: Responding to the Great Recession
Mason: Well, I think... That's a big question. I think that the Fed has got, I think if we're honest about it the Fed has got to get into more directed, more explicit direction of credit in the economy. We have this weird idea that okay, on the one hand we're expecting the Fed to be in charge of macroeconomic stabilization, keep output at potential.
Mason: But then we talk about all these other problems with low rates, that low rates don't actually stimulate productive investment, they stimulate asset bubbles and so on. Well, in that case either you shouldn't be using interest rates as a macroeconomic policy tool or there's something wrong with your financial markets that it channels abundant liquidity into socially useless or destructive activities instead of productive investment.
Mason: If that's the case, then the Fed is going to have to step in and say, "We are going..." Which many central banks, historically in much of the world, have done. We're not simply providing more liquidity, we're not simply lowering a target rate. We're going to direct credit to particular activities.
Mason: Which they sort of do but not in an open or explicit way and not in a way that involves any sort of discussion. Obviously purchasing mortgage backed securities is directing credit towards owner occupied housing. That's just what they're doing, but they don't say this is our policy, we are going to support lending to owner occupied housing.
Mason: If they would be willing to be more upfront and acknowledge that the only way you can manage demand is in this more directed way, then we can have a conversation. How are we going to vastly increase financing for the kind of investment that we need to deal with the problem of climate change? That would be a good place to start.
Mason: Or the infrastructure spending, everybody talks about, that Trump talks about, the Fed could make it happen. They could say here are the types... We've got, in a lot of state and local governments, we've got budget crises. Obviously in Puerto Rico and elsewhere. If the Fed could find the way and when they're motivated, they usually can find the way, to directly support those governments, maybe by actually just simply buying their securities or in other ways, you could alleviate a lot of suffering but you could also directly support demand in the economy because these local governments, the ones that are really facing these financial constraints, would increase their spending or decrease taxes very directly.
Mason: So I think at some point, if we're going to expect the Fed to actually be in charge of managing demand, then we have to also expect them to manage the direction of spending and direct credit in a much more intrusive way. And of course, it follows from that that we want probably this to be a more democratically accountable type of organization than it is.
Mason: But I think this notion that you can have a very large effect on the level of GDP while keeping your hands completely clean about the composition of GDP is just not tenable.
Beckworth: I guess I would go for the first option on that. Maybe trying something like helicopter drops, permitting increase in the monetary base before I'd go down the path of directing credit. I'm worried, I guess, if we go down the path of directing credit the Fed's going to make some of the same mistakes that the private sector, that you worry about, is making. How do we know for sure what's the best use of credit?
Beckworth: So before... That's my own perspective, I would at least like to try maybe something different, QE where the injections are permanent and maybe tolerating a little bit higher inflation temporarily, going from there.
Mason: Well, I'm all for higher inflation targets but they can't even hit the target they've got so I'm not sure.
Beckworth: Well, I hear you on that, I hear you. I just worry, I don't want to end up like the US government running a bunch of state owned enterprises like they have in China. What I do think though, with QE, I mean I have my theory on QE and listeners are probably tired of hearing this but my critique of QE is that it effectively was a temporary injection. It was all geared towards credit reallocations, as you mentioned, it was tinkering on the margins, right? It was trying to get spreads down in certain areas.
Beckworth: I think that, it's not my favorite approach, but I think it did make a difference during QE one. I think QE two and QE three were less effective, didn't do as much. But I think a better approach would have been just to have said, "Look, we're going to permanently inject the monetary base." You wouldn't have needed as big of an expansion of the monetary base to have done that, but you would have had to have done something along the lines that Krugman talks about in his '98 paper, Woodford… where you basically, you do level targeting. Level targeting requires the expectation of a permanent increase in the monetary base. QE was the opposite, it's a temporary increase.
Mason: Well, I guess we have a difference of opinion on this one. I think this sort of temporary permanent distinction makes a lot more sense in models than when we come into real policy content. The truth is, the Fed is the Fed today and tomorrow it will be different. It is just not possible for today's Fed to make a commitment to do anything permanently, because President Trump is going to appoint his own chair next year and he'll do what he wants to do no matter what Janet Yellen said she was doing.
Mason: So, the notion that the Fed can really set people's expectations for all future time, again it's easy to write down in a mathematical model. I don't think it's really descriptive of real policy questions. I'm also skeptical that the actual economic behavior is forward looking in that way but that's almost a secondary question in this case.
Beckworth: Well, fair enough. Now, your paper has received a lot of attention, as we've talked about, and one of the criticisms I saw on Twitter said, "Look, all you're doing is just pulling demand forward. All you're doing is just tapping into future spending, bringing up forward, you're not making much of a difference." I have a response to that but I'd like to hear what your thoughts are when people say things like that to you.
Mason: But that seems to assume exactly what's under dispute. If you take as your premise that in the long run there's a given supply determined path of real GDP, then by definition once you assume that all that demand policy can do is move demand around.
Mason: But that's precisely the issue that we're debating here. If you say... if you believe any of these stories about the labor force or productivity or for that matter, population via immigration, which isn't something I talked about, but is part of the story too. Any of that stuff adjusting endogenously, then that's not the world we live in. So to me, that sort of argument is assuming its own conclusions. It doesn't really seem very persuasive to me.
Beckworth: Now, I agree with that. What they're thinking of is something that we just went through where I live, there's a sales tax holiday for your kids to get supplies for school, right? And it's probably true that what I would have bought later I'm buying on that particular day.
Beckworth: But this is very different, what you're arguing is for an increase of the potential, you're saying, "Look, let's get potential GDP back up to where it should be." That's equivalent to saying, "Let's get the permanent level of income back to where it should be." In fact, I think you can invoke Milton Friedman here. You could say look, all we're doing is tapping into Milton Friedman's permanent income hypothesis. It's shrunk because of bad policy, we're going to take it back the where it is.
Beckworth: If that's the case, we're not pulling forward, actually you're doing greater consumption smoothing because we have greater permanent income.
Mason: That's right, that's right. You should not in this conversation treat it as just a premise that permanent income is fixed and given and invariant to policy.
Beckworth: Yeah. I think that's often overlooked. It is true since the Fed's working through interest rates, there's this intertemporal dimension, you're pulling resources from the future to the present. But it's also true that if you increase potential, you've increased your future income as well. So it's not just interest rates, allocating resources across time, it's a bigger stock of income across time as well.
Beckworth: Let me move on, the time we have left, move on to something that I love, and that's nominal GDP level targeting. I'd love to hear your take on that.
Parting Thoughts on Nominal GDP Targeting
Mason: Well, I'm not against it. I think it would provide an improvement on what we've got, especially the level targeting piece of it. But I think, I mean hypothetically, if we'd had that kind of arrangement 10 years ago, we presumably wouldn't be having this conversation now because everybody would sort of agree with me that we need to get back on that previous trend.
Mason: On the other hand, the reasons that people reject the trend, what if there's a permanent fall in productivity growth, do you really want to accept permanent inflation? I don't know. I think in general I'm not against it but I'm skeptical that changing the targets by the central bank in itself has that much of an effect on the real world.
Mason: I'm not convinced that people respond to their expectations of what the central bank is going to do that strongly, and I'm not sure that the central bank can commit to its future behavior very strongly, which is one reason that people don't weight that very heavily in their expectations.
Mason: And I'm also not sure that even given that target, what the Fed would actually be doing to meet the target, because again they're not meeting the target they actually have. So, maybe in that world we would be 10% below the Fed's GDP target. And they'd be raising rates because of some vague concerns about financial stability. I'm not convinced... I'm not against it, I think it would probably be an improvement but I'm not convinced it would have that dramatic an effect.
Beckworth: Well, let's assume for the sake of argument that it could hit any target it wants. Let's assume it's credible. Maybe it's a joint operation between monetary policy and fiscal policy. I think there's ways you can come up with a mechanism to make it work. But just for the sake of argument, wouldn't it make more sense to directly target spending and demand rather than a symptom of it in inflation?
Beckworth: I mean inflation targeting, we're targeting inflation which could be caused by supply shocks or demand shocks. Why not cut to the chase and go directly after aggregate spending?
Mason: I think that's a good argument. I agree with that in general. I think it is a clear formulation of what the central bank is doing to say that it is targeting, and it really is always targeting whatever the stated target is, the flow of monetary spending, the flow of aggregate spending in the economy. I think it's definitely a clearer description of what the central bank is doing.
Beckworth: Yeah. Well, interesting conversation. We have run out of time today. Our guest has been JW Mason, JW thank so much for coming on the show.
Mason: Well, thank you for having me.