Christopher A. Sims (2014) - Inflation, Fear of Inflation, and Public Debt

The insights I’m going to be talking about today are derived from rather complicated dynamic models – which I’m not going to be able to explain to you in detail, so you’ll understand them in half an hour. My aim is to give you some basic intuition about the new types of theory about inflation and public debt that I think are becoming more and more central to macroeconomics - and some of their nonstandard conclusions. And hope thereby to motivate you to actually understand the theory by going and reading the literature. The way many people have been taught to think about inflation is the monetarist view that velocity, the ratio of nominal income to the stock of money, whatever that is, is a fairly stable quantity. M, the money stock, is under the control of policy. And this implies a simple one-dimensional, one-directional causal model, connecting policy to inflation. Money growth determines the growth of nominal output: P times Y. And since it can have no long-run effect on output, it alone determines inflation in the long run. If you read Milton Friedman’s more popular articles, this is the view that you will see coming through every time he talks about inflation. But this theory no longer works for several reasons. One is that in the simple monetarist models, when they’re made formal, M is non-interest-bearing government-issued currency. Of course, most of the money stock, as it’s usually defined, is not currency; it's various kinds of deposits that are labelled as transactions deposits. So M1 includes bank created money. And in a simple monetarist theory, what you learned in first year economics text books before 2009, was that there was a money multiplier and that was high-powered money, which was non-interest bearing reserves plus currency that was controlled by the government. And there was a money multiplier that reliably translated high-powered money into M1. But in rich economies today nearly all deposits in principle pay interest. Though right now the interest is trivially small in many cases. Even chequing accounts and reserve deposits at Central Banks pay interest. Reserve deposits at Central Banks, in other words, are now yet another kind of interest-bearing public debt. They are not distinguished from public debt that bears interest by the fact that they don’t pay interest. The only thing left that’s non-interest-bearing government paper is currency. And currency is not M. Most of the statistical work on monetarist views of inflation is very explicit that M is made up of more than just currency. So I’ve said one reason the theory doesn’t work is that now everything pays interest. But of course right now nothing pays interest, or rather all short-term securities are paying almost nothing. Central Bank reserve deposits now pay interest but it’s very low. In the US short-term government debt, treasury bills, pay lower interest than reserve deposits at the Central Bank. And this has led to banks' willingness to hold reserve deposits far in excess of required reserve ratios which has made the money multiplier completely obsolete. It’s no longer true that there is something you could call high-powered money. And that there’s a strong tendency of the money stock is usually measured to grow in proportion to the high-powered money. And as I’ve said reserves in the US and almost all Central Banks now pay interest. They’re just another form of government debt, not that much different from treasury bills. So the notion that there is something called M that’s used for transactions and doesn’t pay interest, and is distinct from interest-bearing public debt - that’s essentially obsolete. So how do we think about determination of the price level? The theory that I’m going to talk about is more complicated than MV = PY. And so I’m not going to be able to convince you in the way you should insist that you be convinced if you’re economics graduate students or young economists in general. But I’ll just give you a brief idea of the intuition behind it. In replacing MV = PY is this equation, B over P, where B is nominal government debt. This is fiat debt, debt that’s government paper. It promises to pay more to government paper at maturity. P is the price level so B is the market value of outstanding public debt. So if there’s long-term debt, its market value may fluctuate even in nominal terms. So if debt is issued at 5% 15 years ago and is still on the market at a time when the market short rates have dropped down to 1% or less, that debt is going to have a much higher market value in nominal terms. But then we divide by the price level to get a real value of marketable government debt. And the equation is that that should turn out to be equal to the discounted present value of primary surpluses. Primary surpluses are revenues less expenditures. Now, the expenditures do not include interest payments on the debt when you’re calculating the primary surplus. And in conventional budget accounting for, when you look at the government deficit as usually computed, interest expense is counted. The primary surplus subtracts interest expense. This is a simplification for many reasons. The most important one in this version of the formula is that it assumes that there is a constant known real rate of interest, rho. Of course, the real rate of interest does vary over time. But as a simplified view, assuming a constant real rate of interest is much like assuming a stable velocity in the monetarist theory. So this is a simplified view of this other way of thinking about determination of the price level. And I don’t have to explain how to drive it in detail. When I present this theory to audiences that include mostly people from the financial industry, they find no difficulty at all in understanding this formula. It’s derived the same way the usual valuation equations for bonds or stocks are derived. In fact, sometimes I’ve had experienced people from the financial industry who said that seeing this kind of theory of the price level finally makes sense of determination of the price level to them. And they’ve never been able to understand MV = PY. But if you’ve had too much economics this may seem like a very strange and bizarre theory. And so I’ll ask you just to accept it for the time being. If you want to understand what kind of model produces it, there is a bunch of papers and literature that produce it. There’s lots of controversy about exactly what the implications of this model are. If you come to the discussion section this afternoon, I’ll try to give you references to papers where you can pursue the theory. But it is clear that this is less simple than MV = PY. For one thing an equation like MV = PY will still be part of any general equilibrium macro model. It doesn’t get replaced by this bond valuation equation. It coexists with it in equilibrium. Furthermore, it doesn’t simply replace the M causes PY uni-directional model, one-dimensional model, with a nominal debt causes PY uni-directional model. This is one widespread misunderstanding of this theory which is called the fiscal theory of the price level. People often think, well, the fiscal theory of the price level must be just like monetarism except with nominal debt instead of money. So when there are big deficits it must predict there’s going to be a lot of inflation. But that’s not what this theory is. That equation has 2 things in it that are under the control of policy. B, nominal debt, that’s determined by current nominal deficits and, of course, whatever debt was carried over from the past. And future primary surpluses. Those are both things controlled by policy. Furthermore, as I said, if there’s long-term debt the value of the long-term debt can fluctuate with the interest rate. So there’s another variable that’s left out of the equation as I wrote it. Here we’ve got nominal debt as if that’s a fixed number. But actually, if interest rates and expectations of future interest rates change, B can change without any deficit, because B is the market value of the debt. Besides the fact that it’s got 2 variables in it that are controlled by policy, it also has expectations of the future in it. One of the appeals of MV = PY as a nice simple theory is it's mechanical, all the variables in it are dated today. It doesn’t get to be complicated unless you really start thinking about what determines V and interest rates and start thinking about systematic effects of interest rates on velocity. But if you maintain the idea, well, velocity is going to be fairly stable anyway. You’ve got an equation that doesn’t involve thinking about the future in deciding about the price level. This theory says, what markets believe about future fiscal policy is central to determining the price level. So let’s start asking some policy-relevant questions of this theory. I’ll give you the answers. I won’t tell you exactly how you derive them, you’ll have to read the literature to understand that. But could debt accumulation lead to runaway inflation? This is a widespread fear. Many people look at quantitative easing. The policy in the US and Japan, and to some extent in Europe, of the Central Bank expanding its balance sheet by buying assets and increasing reserves as creating a risk of an explosion of inflation at some later point. Now, some of that comes from people who are monetarists and still think of reserves as high-powered money, without recognising the fact that interest is paid on them now. Which just breaks down the high-powered-money-multiplier argument. But some of it is from people who understand that if debt gets really high… Even if they’re monetarists, they may think the pressure from the fiscal authority to start expanding the money supply by buying government debt and holding down the interest rate might get to be very great. So the worry is that if deficits are large, debt gets big relative to GDP, we could end up with a lot of inflation. And the question is, is that a legitimate worry? Well, from the point of view of this theory it’s certainly something that could happen. It depends on if expansion of nominal debt is going to produce inflation. It has to be because the nominal debt is expanding despite beliefs about future primary surpluses not expanding. The future real primary surpluses must be holding fairly still while nominal debt goes up. If that’s true, then it will create inflationary pressure today. One way to think about that intuitively is: The government is putting into the hands of the public lots of debt that at current prices make the public richer. If nothing else changed that would make them want to spend more and would create inflationary pressure. If people believe though, that as this expanded debt gets into their hands, there is a corresponding expended stream of future primary surpluses that will back the debt, then they don’t feel any richer. They have bonds which are a positive entry on their personal balance sheets, but that’s offset by expected future taxes less expenditures. So they don’t feel richer. The notion that increased nominal debt leads to inflation depends on what people believe about future fiscal policy. And it’s unlikely that people’s beliefs about future policy are reasonably treated as invariant to current fiscal policy. So to get a theory of inflation out of this, you have to start asking, how the current policy decisions and current announcements about future policy affect markets belief about future fiscal policy. Monetarists - and in fact in this sense monetarists include the New Keynesians – they ignore the government budget constraint and this B over P equals discount at present value of tau equation. When they’re thinking about how inflation is determined, they ignore this part of the model. They set it aside. The equations from which this B over P equals tau over rho equation are derived just disappear from the New Keynesian models and from monetarist’s models. It’s not that the people using these models think those equations aren’t there. It’s that there are assumptions that are natural that make it possible to ignore those equations in determining the price level. The kind of assumption that makes it possible to ignore deficits in thinking about the price level is that every increase in debt today creates an expectation of increased primary surpluses in the future. So that increases in nominal debt don’t have any effect on the price level because they’re always balanced by people. When people see a big deficit today it makes them think there’s going to be increased fiscal stringency in the future. So they don’t feel richer. It might seem more plausible that if people see current deficits it makes them think, primary surpluses in the future are going to be less, or there might be deficits in the future. If that’s the kind of expectation, interaction there is between current fiscal policy and expectations of future fiscal policy, then debt expansion is inflationary. Now, another question on this policy issue is, could a Central Bank by being firm enough 'force' 'responsible' fiscal behaviour? That is maybe you could argue that it’s impossible for the fiscal authority to issue debt today without creating an expectation of future primary surpluses that back it so long as the monetary authority is firm enough. For example, that it’s committed to a fixed rate of growth of the money stock or even a fixed level of the money stock. I can’t explain it in detail; this requires going to these dynamic models. But it’s not true that there’s an asymmetry here, that the Central Bank can force the fiscal authority. Neither the Central Bank nor the fiscal authority can force the other one to do anything. It is absolutely true that if the monetary authority adheres to, and can adhere to, a fixed-money growth-rate rule. Or a rule that says interest rates are going to respond very strongly to inflation, so real rates go up when inflation goes up. If they have that kind of policy rule, then there is no equilibrium in which the fiscal authorities don’t back current debt with future primary surpluses. But it’s also true, if the fiscal authority doesn’t back current debt with future primary surpluses, there is no equilibrium in which the monetary authority maintains a constant money growth. There are no forces within the model that make one of those policies force the other policy authority to follow a different policy. It’s just that there are mutually inconsistent policies. If both authorities try to embark on these mutually inconsistent policies, the message of the theory is this situation will not last forever. The public knows it doesn’t last forever. That what actually happens now depends on beliefs about which authority is going to change policy when. And that’s all you can say. Another interesting implication of this theory is that if the public is convinced that deficits today do not imply surpluses tomorrow. That the surpluses tomorrow are stuck by, for example, a political process in which everybody can see that there is no way taxes are going to go up. And there’s no way expenditures are going to go down. We’re stuck with this particular stream of real-debt service and there’s political gridlocks, so that can’t change. If that’s true then when the monetary authority raises interest rates, it does not reduce inflation. Actually, the rise in interest rates, in that case with primary surpluses fixed, the rise in interest rates means that the conventional deficit, and hence the rate of issue of nominal debt, goes up. And the rise in interest rates is inflationary, not deflationary. There are actually time periods, and countries, where this has been front and centre consideration in monetary policy. In Brazil during the periods of high inflation, the monetary authority which already had interest rates in the 20% range, facing inflation in the 30% range, thought that it might be a good idea to raise interest rates to make the real rate positive. But they recognised that given the political situation, the rise in interest rates – because public debt was already very big, interest expenses a large fraction of the budget. The rise in interest rates would just increase the rate at which public debt was growing. And they knew this would be inflationary, not deflationary. This is a situation that’s easy to understand from the point of this theory. Very mysterious from the point of view of a pure monetarist view. But today I think in the US, Japan and Europe it’s not true that the public thinks that current deficits imply future deficits. In fact, the opposite is true. In fact, I think, even something worse than the opposite is true. People in the US, Europe and Japan are very aware that their ageing populations will require painful fiscal adjustments in the not too distant future. And that the political processes are not addressing these adjustments. The degree to which the political process is addressing these adjustments varies across countries. In the US it’s particularly depressing. How difficult it is to get any progress on resolving the issues of who is going to be hit, to what extent, by future fiscal stringency. Are the old going to have fewer retirement benefits? Are the young going to pay higher taxes, and if so how much? These are third rails in political discourse. Nobody wants to talk about them. In 2010 in the US 60% of non-retired people believe that social security, the main government retirement programme, would not exist in a form that would provide them any benefits then they retire. People are quite pessimistic about the future fiscal situation and its impact on them. With these beliefs, deficits that seem to arise out of crisis and political gridlock increase people’s uncertainty about who will be affected by future fiscal adjustments. And in my view makes deficits actually deflationary - actually, makes deficits reduce demand. People see their deficits. They already knew there were going to be budget problems in the future. They’re convinced that inflation will not occur. Part of this is that they’ve been very well convinced that Central Banks really can control inflation. And it’s been low and it will stay low forever. So these deficits are going to have to be covered. It already looked like I wouldn’t have any social security and my taxes would be high. This debt comes up so it looks like it’s going to be even worse. And furthermore I don’t know - if I’m young, is all the adjustment going to be on the young or is it going to be on old people? So the level of future fiscal effort is increased. And at the same time uncertainty about who is going to actually bear it increases. So the standard assumption in monetarist models, that lets people ignore the fiscal side in determining the inflation rate, seems to be more or less accurate for these countries that are now in slow growth, low inflation situations. And that’s a big problem. Because the way the monetary authority controls the inflation rate, if fiscal policy has this form where current deficits create expectations of future surpluses, is it moves the interest rate. So when inflation goes up, the way the monetary authority produces a uniquely determined price level and inflation rate is it promises that, if inflation goes up, it will raise interest rates even faster than inflation. So real rates will have to be higher if inflation is higher. And if inflation goes down it will lower interest rates even faster than the decline of the price level. And so real rates will go up. But all these countries are stuck at, or near, the zero lower bound. They have almost little or no room to make nominal interest rates go lower. The fiscal theory predicts that there is a class of equilibrium in which there’s a well-determined inflation rate in which the nominal rate is stuck. Contrary to what you’ll see in most textbooks, it is possible for the monetary authority to decide to just peg the nominal rate. And this does not lead to chaos or terrible equilibrium. What it does is transfer the determination of the price level from the monetary authority to the fiscal authority. If the monetary authority fixes the nominal rate, the price level is then determined by the fiscal authority. Assuming the fiscal authority does not follow the rule of making real future primary surpluses respond to the nominal debt. But instead follows a rule that says they’re committed to a fixed path of future primary surpluses, regardless of what happens to the current nominal debt. With that kind of fiscal policy you get a well-determined price level and inflation rate, in which you get something like the monetarist theory with B determining PY. But that’s not the situation we’re in. The situation we’re in is one in which the public thinks that current deficits correspond to future primary surpluses. And this theory says that if you have that combination of policies, the public believing that current deficits imply future primary surpluses, in other words future fiscal stringency. And at the same time interest rates can’t follow the inflation rate down, because they’re stuck at zero. Then, the theory says, the price level is indeterminate; it can do essentially anything. In fact, the theory says it will tend to drift downward stochastically. But the current price level is indeterminate. Of course, this is on the theory that this situation, this policy combination is going to last forever. If you say it’s only going to last a long time, then what happens today depends on beliefs about how long that time is and what policies revert to when that time is over. But that’s just another way of saying it’s very hard to say what the current inflation rate is going to be when current policies configuration seems to be in this form. And, unfortunately, this prediction of the theory looks uncomfortably like what we have seen in the US, Europe and Japan. Long periods of interest rates stuck at almost zero. And inflation, despite attempts at innovation by Central Banks, inflation not really doing what Central Banks would like to see it doing. Central Banks would like to see an average positive inflation rate of 2% or so a year. They’re not seeing it. In Europe it keeps going down instead of up. Central Banks undertake policies that amount to saying they are going to issue one kind of interest-bearing government debt, interest-bearing reserves, in order to buy another kind of interest-bearing security. Until you get into complications like risk aversion and stochastic modelling, these models suggest that this kind of policy of the government buying one kind of security for another is not going to have any effect at all. It’s not going to affect private budget constraints. So from the point of view of this theory, quantitative easing is a very weak policy. An attempt by the Central Bank to do something in a situation where what it would like to do and ought to do is get interest rates into negative territory, but it can’t do that. And fiscal policy, unfortunately, remains in a configuration where people think that deficits are going to be financed by future taxation. So is there an exit from zero lower bound? Yes, it’s easy to describe, but not easy to implement. It requires fiscal policy that’s expansionary now, without a commitment to cut future expenditures or raise future taxes to preserve current price stability. But, unfortunately, people are very convinced that that’s not the kind of fiscal policy we have. And it would require a major shift in the thinking of policy makers and the speeches they make to get people convinced that this is what’s actually going to happen. And furthermore it requires the political system to make commitments across time and stick to them, which is very hard for politicians. So I have just a couple of slides about Europe. The simple fiscal theory assumes one government issuing debt denominated in the currency of that government. A government like that never has to default. So the theory doesn’t discuss default. European fiscal authorities do not issue debt in a currency they control. Their sovereign debt has defaulted recently and is still considered defaultable by markets. At least if you look at the same kind of security denominated in Europe, as different interest rates – depending on what country issues it. So we don’t have time to elaborate the theory to cover default in multiple fiscal authorities. It gets a lot more complicated. But here are just a couple of the implications. The EMU was set up with the mistaken idea that it was possible to completely separate monetary and fiscal policy. But every monetary policy action has fiscal implications. That a commitment by the ECB to thwart speculative runs on European Union sovereign debt creates fiscal risk via a potential need for capital injection into the ECB, has become evident. If they bought a lot of sovereign debt of southern European countries, there would be a possibility this wouldn’t succeed. If it didn’t succeed the debt might drop in value. The asset side of their balance sheet would go way down. This would impede their ability to fight inflation. So revulsion against the idea that ECB actions could create fiscal risk, and hence potential cross-country implicit fiscal transfers, has limited ECB policy in the crisis. The northern European countries recognise this as a possibility, that the ECB could take actions that would actually end up taking money from Germans and Finns and give it to those irresponsible Italians and Spanish. And that has created resistance. And without fiscal backing, if there were not agreement by the European Union members to recapitalise the ECB if it really needed it, it’s possible that combatting a speculative attack could put the ECB’s balance sheet at risk. This is a very low-probability event. But it is non-zero probability. And in that case it would require a fiscal injection. And that would be an implicit transfer across governments. And no single government in the EMU can make the kind of expansionary fiscal commitment that I described would be necessary to get out of the zero lower bound. In the US it seems very unlikely that that will happen, but we know what would need to be done and who would need to do it. It’s not clear in the ECB that we would. So a combination of fear of inflation, fear that the ECB could be forced to inflate or allow inflation because it didn’t have fiscal backing, and an incomplete set of fiscal institutions could leave Europe in an environment of low inflation or deflation for a long time. So I'm done.