 When I saw the title of this panel, Monetary Rules in Light of the Crisis, I sort of did a double take, because the notion of monetary rules, by definition, that they should be somewhat timeless, the merit of a rule should not rise or fall depending on whether or not we had a crisis. But of course, the fact of the matter is that if we never had had the crisis eight years ago, nobody would have questioned the implicit monetary rules that we had been following up until that point. It's only because of the unsatisfactory economic growth we've had since then, that we're now having a vigorous discussion about whether we should have different monetary rules. And I'm sure people in this room must have just gasped when they heard none other than John Williams, who is about as central to the fraternity of central banking in this country as they get, rolled out the idea of a higher inflation target or a nominal GDP target a few weeks ago. In fact, when I saw the agenda of this year's Jackson Hole conference out in Wyoming for the Federal Reserve, I thought, did these guys steal the agenda for your conference and just try and light one up you? Because the theme of new monetary frameworks, which is with a theme of the Jackson Hole conference, is really central to what we're gonna talk about on this panel. And I'm very pleased that we have three terrific economists who are just perfectly suited to address this question, most of whom you probably know either through their academic or their popular writings, but I'll very briefly introduce them starting directly to my right. Peter Ireland teaches at Boston College. He has written a lot about monetary policy, interest rates, inflation. He's recently recently about nominal GDP targeting. Peter gets special points in my book because he's married to a Canadian. Next to him is Miles Kimball. For a long time, you probably knew him as associated with the University of Michigan, but he is now just before the trade deadline. He was not by the University of Colorado. You might also know him from his blog, Confessions of a Supply Side Liberal. And on the far side, David Beckworth, who many of you know because he's now here associated with the Mercatus Institute at George Mason University. He was previously a Western Kentucky University. He's also, runs a terrific blog that I read very regularly called Macro and Other Market Musings. And he also, I believe, really big deal to me is selling NGDP target t-shirts. I'm still a little bit annoyed that I haven't got mine yet, David, so I'm just taking this opportunity to remind you once again that I haven't got my t-shirt. So we're going to start, I think, with the far side date. We're going to move down. Each of them is going to give a presentation and then we're going to have a round-table conversation about it. All right, so the title of my paper is The Fed's Dirty Little Secret. It's getting late. We need to catch you, title, keep you awake. So this is what it is. And as we get into it, it's not really such a dirty secret, but I've kind of framed it this way to kind of motivate what we're going to talk about. And you're going to have to wait to find out what the dirty little secret is a little bit later in the talk. I want to begin and motivate this idea that's a dirty little secret by asking, why wasn't QE more effective? There's a long literature, not long, but there's a sizable empirical literature that's tried to establish what did QE accomplish. There's been some evidence that maybe it tinkered with yields. It's mixed evidence on the real and inflation effects. But overall, QE has been disappointing relative to what was expected, the hopes, the claims, FOMC claims that it'd be a stronger recovery. They talked about QE shoring up aggregate demand growth and it just didn't happen. Moreover, it didn't even help nominal growth that much. It's one thing maybe to say, look, there's other things going on driving the real economy, but you would expect the Fed to have some meaningful control over nominal measures. Inflation's been at 1.5%, the core PCE measure, since about 2009. Nominal GDP, I have a picture here, fell below its trend growth and never recovered. None other than Brad DeLong found this astonishing. He noted to him it seemed like it would be normal or the way we do macro policy to return nominal demand to its trend growth path and it didn't happen. Some have argued, well, maybe this is just a consequence of a financial crisis, right? Reinhardt and Roghoff, of course, Roghoff, excuse me, are the most prominent advocates of that, but there's been some other research that has said, look, it's not just the case that financial crisis must always lead to weak recoveries. It's conditional on the policy response. A number of papers have come out and argued that and so that raises the question, why didn't QE at least shore up nominal economic growth, like nominal GDP or nominal income? And I'm making the argument, the reason it did not is because the Fed cannot credibly commit to a permanent injection of the monetary base. It needed to do that in order to get us off the zero lower bound in a quick, efficient manner and I motivate this and talk about this in the paper, both from the quantity theory of money perspective, kind of a look back at some of the old monetarist discussions of this, but also from a New Keynesian perspective, some of the research that dates back to, well, Krugman's worked from 98 and forward, they look at what does it take to get off the zero lower bound and they say a credible commitment and in the paper, I note that that may even mean you don't actually have to have a monetary base injection, just a credible commitment to it. And so in the absence of that, we weren't gonna have a great recovery. The monetary base injections, because the Fed couldn't commit, they were gonna be temporary and I'll explain in a minute why they had to be temporary. And by making the monetary base injections temporary, and I know this was talked about with John Taylor and David Laidler and I'll document the Fed's intent that they are temporary, but because they are temporary, they consigned QE to the, oops, let me go back, slide. To the irrelevant results of, let's see, where is it? Yeah, to the irrelevant results of Krugman and Woodford and Egretson. So Krugman, through his credit in 1998, kind of rediscovered this idea that I would argue that the monetarist had known all along that when you're at the zero lower bound, something that you have to do is you have to have a permanent and sustained increase in the money supply. Now he did this from a New Keynesian model, New Keynesian perspective and following him, there was Egretson and Woodford and their Brookings paper and a subsequent spate of other research that's come out that's confirmed this. And what they show is, given that central banks operate with interest rates when you're pinned at the zero lower bound, you need a permanent increase, only a temporary, well, permanent increase leads to a temporary increase in the expected inflation rate. And the idea is if you have this permanent increase in the monetary base, you're going to raise the future price level and that future price level increase will raise expected inflation today. And it has to be enough to push the real rate down. And so the Fed's dirty little secret, I'm gonna kind of cut to the chase here, is that QE was never going to spark a robust recovery. Given that the injections were always gonna be temporary and to the extent that they had to postpone the shrinking of the balance sheet, they sterilized them with interest on reserves, QE was never going to generate the recovery that they hoped and claimed that it would. I document some of the claims that were made by Fed officials, what QE would do. But by design, it was not going to spark a robust recovery. And this is not to say it had no effect. On the margin, spreads may have been tinkered with some, but in terms of a quick robust recovery, it just didn't happen. And I think that's because these irrelevant results of Ericsson and Woodford and Krugman and others has been borne out. And so what I do in the paper, and I forsake a time I can't go through all of it, is I go through and show how this understanding that was well understood back by the early monetarists was understood by the New Keynesians and the Fed went ahead and did it anyways. So as if they ignored this whole literature that said, look, if you're at this point, you've got to go to permanent injection. So I'm gonna just give an example from each camp, one from the kind of old monetarists, one from the New Keynesian perspective. And I wanna be also clear that picture is a little deceiving there. There was no grand conspiracy at the Fed. That implies that they knew better. I think it was more they were constrained by their commitment to low inflation. And that's kind of the other part of this paper, is why did they make it temporary? And this is their commitment to inflation. And in fact, I document in the paper that they claim to have a flexible inflation target, but it seems to have become more of a rigid, just straight up inflation target. And that has constrained their activities. They're still fighting the last war of the 1970s when it's a new battle front that's facing them. But moving on, let me bring up Brunner and Meltzer in an 87 paper. And they make this key distinction between a transitory injection versus a permanent injection, just to demonstrate they understood this difference. And they say the adjustment of reserve positions to transitory change by buying and selling federal funds or by borrowing or repaying loans to the central bank has negligible effects on interest rates and asset prices relevant for household and business decisions. They continue a perceived permanent change in the monetary base initiates very different responses and has different costs. The adjustment to a perceived permanent change is reinforced by changes in price expectations. Just to kind of show through a little flow chart, a little model what the differences would be. Here we have the equation of exchange. And we're gonna invoke this idea. I mean, what the monitors are really doing is they're invoking this long run neutrality of money idea that in the long run, a permanent increase only affects nominal variables. Real effects may be affected short run, but not long term. So we have the injection, the M term here goes up. Moving forward, is it gonna be permanent or temporary? What's based on the expected path going forward? If it's going to be permanent, then people expect prices to rise in the future, money demand goes down, and velocity goes up reinforces that, and so nominal spending picks up. On the other hand, if you think it's going to be temporary, so the summation of all future money supply gross, we're gonna sum to zero, which means at some point in the future, it might as well be pulled out and potentially we have some deflation, money demand goes up, and we have an offset velocity drops, offsets increase in the money supply, so there's no effect. And that's the argument they were making. So a question that comes up, is this still relevant? This idea that some of the early monitors had, is the quantity theory money relevant? And of course this has a long history, going back to at least the Hume. And in the paper, I do just a little vector auto regression experiment just to see, look, can we still invoke this idea and justifying kind of the idea that escaping the zero lower bound requires a permanent injection? So I took a VAR and I estimated over the period 1960 to 2007 quarter four, and identified an exogenous shock to the monetary base using long run restrictions. And I also controlled for money demand shocks, or demand for monetary base that would endogenously grow. And this is what comes out. So this is a 1% shock, and what is a typical response? This is by quarters. So you see there's a 1% shock, it still evolves to about 3%, 20 quarters out. A nominal spending tends to follow that. Here's the price level on the bottom one. And by the way, these are permanent shocks, I'm talking about in this first column here. And this leads to an increase in the price level, you'll note the price level takes time to actually catch up. It's not really significant until about two years out, which is consistent with this idea of the quantity theory of money, that you have a permanent injection, eventually workers with the economy prices go up. On the other side, I provide the results from a temporary monetary base shock. And you see it creates no effect other than the 1% increase in the base shock is reversed going forward. The rest of the results show that the money supply also goes up. And again, this is an exogenous monetary base shock. I wouldn't expect M3 to go up portionally for some kind of endogenous response. Rural GDP is a temporary effect and goes down and there's a liquidity effect in the federal funds rate. All right, now let's move to the New Keynesian perspective. I'll give an example of that in a minute. But from the New Keynesian perspective, they view monetary policy and it's the way the world operates and many of what we do, and modern macro is kind of a New Keynesian perspective, but monetary policy is viewed through the perspective of the current and expected path of interest rates and be precise relative to the natural rate level. So something that would fall out of Woodford's book and I'm missing an expectation operator on the right hand side, so I apologize, but the RA term, that's the actual real interest rate and the RN term is the natural real interest rate. The sum of the expected differences between those drive the current output gap today and this falls out of the Michael Woodford's New Keynesian textbook and so what this says is look, the current state of the economy, the business cycle driven by the output gap is based on where rates are expected to go relative to the natural rate level. So what happens if you hit the zero lower bound? What do you do? What if the natural rate goes negative and that real one has a hard time going down there because of the zero lower bound on nominal rates? So in terms of the Fisher equation, if the nominal rates pin at zero, what has to happen is expected inflation has to go up in order to push that real rate down until it reaches the natural rate level. That's kind of the New Keynesian perspective and so this issue came up in the late 90s, Paul Krugman and his famous Brookings paper looked at this issue and he came to this following conclusion that a permanent increase in the monetary base is necessary to get that burst of inflation up high enough to bring the real rate down, market will clear and have a recovery. So he says, a monetary expansion that the market expects to be sustained will always work. Whatever structural problems the economy might have, if a monetary expansion does not work, if there is a liquidity trap, it must be because the public does not expect it to be sustained. I'm gonna throw in one more. I'll throw in Michael Woodford. He echoed that same sentiment and I have a large list of New Keynesian quotes that show that the permanent point as well as the monetarist, Nuke in the paper to see all of them. I'm throwing another one out. This is actually from an op-ed in the Financial Times that Michael Woodford is really critical of QE and this quote I thought was interesting. He said, the economic theory behind QE has always been flimsy. The problem is that for theory to apply, there must be a permanent increase in the monetary base. And so he goes on and the last part he mentions that the Fed has not promised to allow inflation to rise above its normal target level. And to be clear, they're not arguing for a 1970s type inflation. They're arguing for a one-time burst that would return the price level to its trend path and get us there. And one of the examples they invoke where this was not done and an example of the irrelevance result was Japan. So here's the example of the original Japan QE. It was mentioned earlier. The monetary base rose on the 75%. It was pulled back out in 2006 and so you see hardly any effect on nominal GDP or on the price level. Woodford makes the point that the fact that it was so big makes it non-credible. There's no way that could be permanent. A 75% increase in the monetary base would lead to a proportional rise in the price level if it were to be permanent. That's just not credible. The Bank of Japan wouldn't let that happen. So by becoming so big, when QE is undertaken, as it has been, it leads to big balance sheets that's been discussed already. And those balance sheets are so big they simply cannot credibly commit to an expansion of the monetary base. We'd need to run away inflation. However, they need a little bit of a permanent increase to get the escape from the zero lower bound. Now look at that top figure there. I want you to compare it to this chart here. It looks rather similar. Well, this chart here is the Federal Reserve's own projection of its balance sheet. This comes from the Federal Reserve's annual report on its balance sheet. And every year since 2010 it's produced something that looks like this. The black line shows the actual assets of the Federal Reserve and the gray line shows its projection. This is actually from the 2014 report that 2015 came out recently. But as you can see, the projected path of the assets, the Fed sees going forward for the balance sheet is gonna put it right back on that trend growth. And you can think of that trend growth simply as being the growth in a monetary base given regular money demand growth over time. So I wanna respond to the comments made earlier. Both Ben Bernanke and Jenna Yellen have committed to this. The FOMC and the June 2011 meeting and September 2014 meeting, they all lay out exit plans that say there's going to be this reduction of the balance sheet going forward. I've got three minutes left so I'm gonna skip this next slide and why has this happened? The argument I make is that the Fed has been committed to inflation targeting and it's become more rigid over the past eight years. This is its core inflation and instead of being 2% it's average one and a half. This is a forecast from its projections, from the meetings, and this is two years out. So two years out, FOMC officials are projecting that at most it'll be 2%. And by two years out, the Fed has some influence over inflation. Here's a three year forecast, some of these it's available. To put this in perspective, Jenna Yellen gave a speech where she laid out what her estimate of the short-term real natural rate was in 2009 and she put it down as minus 5%. So if you think about what that would imply in terms of that Fisher equation, it would imply expected inflation of 5%. And there's simply no way the Fed would have tolerated that much of an increase in expected inflation to get the real rate down to the natural rate level. I also provide a counterfactual, what if the Fed had tried to return normal income to its trend path and here are the different paths and you can see inflation would have been anywhere from three to three and a half percent, which again is far above the actual path taken. So this leads to the QE, what I call the QE paradox, central banks turn to QE when there's large aggregate of man shocks that push it to the zero lower bound and they turn to QE as kind of a work around solutions. They're waiting to get around it. However, QE creates a large balance sheet, almost inevitably leads to these large balance sheets, which means that the monetary base increase cannot be credibly made permanent, it's too big. But we need that, we need some small amount of a permanent monetary base injection. QE guarantees that that won't happen. So QE actually defeats the very purpose it's created for. The balance sheet gets so big that it leads to the expectation that it won't be made permanent and therefore defeats the purpose of getting aggregate demand back up. So my proposal, I think I'm running out of time here. My proposal is to have a rules-based nominal GDP level target and I have a real radical part that might get me thrown out of this conference. And as is mentioned before, a level target allows for some flexibility and short-run inflation so you can't escape it. You don't actually have to do the explicit permanent increase. There's kind of an implicit commitment to that and I explained how on the paper, but I also call for this target to be backed by the U.S. Treasury Department. And this is to add credibility and also incentivize Fed officials to act properly. So what would happen is if the Fed did not hit its target, the Treasury would step in. For example, if it was below target, the Treasury would put a bond at the Fed. Fed would give it money and it would distribute it in an automatic rules-based fashion and it would do the opposite if it was going the other direction. And I'm gonna have to stop there. I'm out of time, but the idea is if you look at the consolidated balance sheet of the U.S. government, that's what's really involved in a permanent expansion and monetary base. And what I wanna do is apply this rule to it in a way that allows that to happen in a predictable, manageable, rule-based approach. All right, thanks. There we go. All right, so I realized that professionally I grew up with the idea of rules. So that debate about rules or not having rules seems something that mostly predates me, but it's very interesting to hear this discussion. And I think the kind of legislation that John Taylor talked about seems like a great idea. You know that you choose a rule, if you depart from the rule, you explain it. And that helps you get to a better rule. It's not that you choose a rule for all time, but to have to explain yourself when you depart from a rule that you've formulated is both good science and good policy. So this is what I wanna argue is that there is a lot of room for improvement in the kind of rules that we have. So I'm gonna call that next generation monetary policy. I wanna argue that monetary policy is really far inside the production possibility frontier. There are a lot of very interesting things we can do to improvement. And one of the most dangerous things would be complacency. So I've heard some people with some experience within the Fed say that there's a certain complacency in some quarters of the Fed acting like, hey, we did pretty well coming out of this crisis. It was, I mean, things could have been worse. And the particular individuals who were in the heat of the battle, I honor them for making things not be worse. And yet the standards we should apply for monetary policy going forward should be much, much higher standards than the kind of thing we've seen in the past. It's one thing to forgive the mistakes of the past, but it's another thing to say that that's okay for the future. It isn't. Now the next thing I wanna say is that the value of experimentation shouldn't be underrated. We shouldn't dismiss monetary policy ideas just because they're new. It's just not true theoretically. On the one hand, you say, okay, we know how things work. If we do the same thing as we always have. But if you do a formal model, the virtues of experimenting with new things can really be enormous. I think there's a danger of complacency, not just by central bankers themselves, but by folks working academically on optimal monetary policy. They can get into a rut. And one of the dangers there is simply writing papers justifying what central banks are already doing. And I think that's a dangerous route to go. I mean, you should, you know, it's one thing, I mean, it's one thing to assume that the folks in private markets have developed some wisdom there. And there's a certain amount of wisdom in the central banks, but we haven't been doing central banking well for that long. So our assumption that there's such enormous wisdom that we should be explaining the wisdom they already have rather than challenging what they're doing, that seems, I think we're better off at this point still challenging it. Okay, so what are some of the things? So, and I'm gonna emphasize interest rate rules here. The, you know, the Taylor rule is essentially description of the Volcker and Greenspan policy. If you free up the parameters, it provides a scaffolding for talking about interest rate rules generally. The virtue of interest rate rules is we understand that we understand how interest rates work really well. And it's interesting that at Jackson Hole this year, Ricardo Rice was talking about a return to interest rate rules. So I wanna set out kind of what I think is some of the exciting research agenda for interest rate rules. Eliminating the zero lower bound or any effective lower bound. Well, I won't read all of these things. So I'll just do them on this individual slides. The first thing, which is really foundational is to eliminate the zero lower bound or any effective lower bound on interest rates. This is the most basic thing we need to do in order to get back to the normal interest rate rule policy that we're used to. And people talk about negative interest rates as if there's something new fangled and weird, they're not. It's a regular interest rate policy exactly as we understand best in theory. And the key thing, though, is that you have to lower all the interest rates, including the paper currency interest rate. And one of the things, so I have a full schedule this fall going to East Asian and European Central Banks explaining how to do this. It's really remarkably easy to get these pieces of paper to earn a negative rate of return. If you have the electronic dollar or euro or yen be the unit of account. And then you have, you know, basically you take, you go off the paper standard, but you still have the paper currency in existence. You don't need to get rid of paper currency to have its interest rate be negative. One way to think of this is the following. If you, you know, the zero lower bound comes because we're worried that people will earn a zero interest rate on paper currency minus some storage cost. But think of the round trip, you withdraw paper currency, store it, then deposit it back in your reserve account. All you need to do to interrupt that zero rate of return is to charge the banks paper currency deposit fee when they put the paper currency back in the reserve account. So it's actually remarkably easy to take away the zero rate of return that the financial markets would get on paper currency. And, you know, there are details to this. You can do this in a way that tends to, I mean, you can shield regular households from negative interest rates in their checking and savings accounts and so on. But anyway, there's now a growing literature on this. I have a paper with Ruchira Garwal on, oops, oh no, what did I do, on breaking through the zero lower bound that's an IMF working paper. I have a lot of links in how and why to eliminate the zero lower bound, a Reader's Guide, which is on my blog. Ken Rogoff has a new book, The Curse of Cash. Marvin Goodfriend gave a talk about taking paper currency off par at Jackson Hole this year. The one difference is I would do it with a very sedate crawling peg and he was contemplating a floating exchange rate between paper currency and electronic money. But I would say I've been advocating getting rid of the zero lower bound since 2012 and the shift in attitudes in the last four years has been really remarkable from being a totally out there unthinkable policy. It's now very much something that's being discussed. So this is not, you shouldn't assume that the zero lower bound is a policy choice, not a law of nature. I think it's a very bad policy choice. Okay, but take away the zero lower bound and we still have the issue of what should our monetary policy rules be. This is an area that I hope to get involved in the research, but I'm only beginning in the research, so this is a research agenda rather than settled research results. But let me therefore ask some questions. So the first question I wanna ask is what if we double the coefficients in the traditional Taylor rule? So one of the talks here was talking about doubling one of the coefficients. What if you double both of them? I would argue that central bank mandates appropriately emphasize inflation and to a lesser degree unemployment. I think interest rate stabilization is less important than that. If we moved rates twice as much, could we close output gaps and stabilize inflation twice as fast? I think this is a very important research issue. We shouldn't assume maybe changing interest rates twice as much would really get the job done of stabilization faster and allow us to move on quicker to the supply side issues because we're more quickly back to the natural level of output. I think that's a very, very important question. Notice that this is much easier to do if you no longer have a zero lower bound because you need to be able to cut rates by quite a bit as well as raise them by quite a bit, but this is two-sided. You wanna raise rates by a lot to cut off excessive booms quickly. Second question I'd like to ask is shouldn't we have data dependence in both directions? Shouldn't it be that as it is, you can look at the Fed funds rate target and it does a stair step up and then has a plateau and then a stair step down and so on, but there's this tradition which I think is a bad tradition that you have a big penalty on changing directions. Well, if incoming data says, hey, we ought to go in the other direction and we ought to go in the other direction and putting a big penalty on changing direction seems to me to be a bad idea. And notice you don't, if you have no zero lower bound, you don't need to get umph for monetary policy beyond the expectations implicit in already in having a more approximate random walk. I wanna say, you don't have to go whole hog here. I mean, in theory, it could be a good idea to have kind of a jagged diffusion process almost for the Fed funds rate. You don't have to go that far to get a benefit from this. I would just argue we should have a lower penalty on changing directions. And interestingly enough, if you look back in the kind of Voker and Greenspan eros, they change directions more often than in the more recent era. And there's kind of this myth that commitment has to lead to interest rate smoothing. That's a technical topic, but I think that's basically a myth. The, okay, next question is, what about the relative importance of output stabilization versus inflation stabilization? Fortunately, sometimes this doesn't matter. There are many shocks that hit the economy where closing the output gap and stabilizing inflation are exactly the same thing. This is a remarkable thing, theoretically, that's called the divine coincidence. But there are other kinds of shocks where there is a trade-off between output stabilization and inflation stabilization. I think the big ones have to do with some change in a relative price. This is kind of relevant to what's going on, what the Bank of England is having to decide there where they're expecting a long run, a persistent change in the exchange rate and wondering how to deal with that. I wanna argue that probably many of the formal models are overestimating the cost of the variability of inflation relative to the output gap. And this is for interesting reasons. The cost of the variability of inflation in the formal models have to do with leapfrogging prices causing people to go to the wrong store. And the cost of that depends on how sensitive people are to going to a store that's not the store best suited to them just because it has an old price that's especially low. And so the more price-sensitive people are between stores, the more you have to worry about stabilizing inflation even in the short run. Those numbers, though, for the price elasticity of demand come from, actually, so I've done some research with Shushanta Basu and John Fernald who are the source of some of these numbers. So they measured the degree of returns to scale at 1.1 and then that translates into a price elasticity of demand of 11. But I think that 1.1 returns to scale is probably an underestimate because when we did the work together, we would find returns to scale for non-durable, say as being 0.6, which really is incredible, you're gonna have at least a constant returns to scale of one. So there are biases because in the booms, you have lower quality factories and maybe workers whose quality is lower than you think who are coming online. So the returns to scale is probably bigger and there's a relationship between the returns to scale and the markup ratio and therefore to the price elasticity of demand, which then implies that probably people aren't quite as sensitive to prices in going to the wrong store. And you can make another argument that if there are barriers to entry and sunk costs that also the price elasticity of demand is lower and that makes output stabilization more important relative to inflation stabilization than some of the models are assuming. So this is a very technical issue, but very important and relevant to a lot of things. It's also relevant to the NGDP targeting debate because do you wanna have a price level target or do you wanna have a nominal GDP target could easily depend on this. Okay, another question is what about adjusting for risk premium? So Mike Woodford and Kertia Woodford say that if you have an increase in the risk premium, you ought to drop the safe rate by about 80% of the drop of the increase in the risk premium. I think that's a good idea. And what this is doing is it's getting reasonably close to say we wanna almost target the commercial paper rate rather than the Fed funds rate. And this also goes in the same direction as what you'd want for financial stability as well. So I think that's an important thing to add to the rules. Now, again, that's something you may need to have eliminated the zero lower bound to do if you have a large rise in the risk premium, but this would have helped a lot back in 2009 and 2008. Finally, let me talk about coordination with financial stability. I wanna argue first that if financial stability concerns are affecting your monetary policy, you're not using your other tools to get financial stability enough. Other than, particularly that's true because other than possible aggregate demand effects, there are very little social costs to having a very high capital requirements. And on this, I'm very much in the same camp as Anadad Mahdi and Martin Hellwig. In particular, we should have very high capital conservation buffers. There's no reason to allow banks to pay dividends or buy back their stock until they have much, much higher capital requirements than they have now. And I wanna, let me end by, well, let me show you this. If you have both low interest rates and high equity requirements, then you can get both higher aggregate demand and more financial stability. Those are two policies that work better together than separately. So finally, I just wanna say there's a lot of potential to interest rate policy. We, I think we ought to aim for something like, or hope for something like closing output gaps and stabilizing inflation within about a year limited mainly by the lag in the effect of monetary policy. We don't need to have QE, which a lot of people have talked about the disadvantages of. We don't need fiscal stabilization beyond automatic stabilizers. We don't need helicopter drops. We can go back to the great moderation and hopefully better than that. We can get closer to staying at the natural level of output all the time and return to focusing on the supply side because we've got the monetary problem more nearly solved. And we can't get there with complacency or timidity. We have to really boldly fold forward. Okay, so in this paper, Mike and I not only advocate for a rules-based approach to monetary policymaking, but in addition, we choose to cast our preferred monetary rules in terms of quantity variables, measures of the money supply that really haven't appeared in mainstream monetary analyses for years or maybe even decades. So question one is, how did we get to this point where it's possible to analyze and even make monetary policy without any reference whatsoever to any measure of the money supply? I don't blame John Taylor for this. Instead, I think it has to do with a broader professional consensus that came together actually around the time that the Taylor rule was first presented in the early 1990s, but much broader than just that. The consensus was that previously stable empirical relationships between measures of money and other key macro variables like output and inflation weakened or maybe even broke down altogether. At some point after 1980, in light of that professional consensus, one of the themes that Mike and I wanna bring out in this research is the idea that measurement matters. If you take care to work with appropriately constructed monetary aggregates or properly adjusted measures of the monetary base, you do see these empirical relations continue to appear even in data after 1980 and they do seem strong enough to continue to serve reliably as the foundations for a rules-based approach to monetary policymaking. Here is an illustration of just that. What this graph shows are year-over-year growth rates in our preferred broad monetary aggregate, the DeVizia MZM aggregate. DeVizia MZM includes all of the assets included in standard simple sum MZM, but the aggregate itself is constructed according to the economic aggregation principles developed by William Barnett and much of his work. Now the data are noisy to be sure, but what you see with the DeVizia aggregate is that pro-cyclical patterns in money growth do seem to appear even in samples after 1980. Notice, for example, the distinct declines in DeVizia money growth appear to pre-sage or foreshadow each of the last three cyclical downturns in the United States. Here likewise is our preferred measure of the adjusted monetary base. For the monetary base, the measurement issues are slightly different. Here the problem, so to speak, is that the Fed Reserve has been paying interest on reserves and when it started paying interest on reserves, it worked to shift the demand curve for reserves potentially very, very far to the right. A lot of the increase in reserves supply brought about by quantitative easing may have done nothing more than accommodate that shift in demand and to the extent that it did, QE never should have been expected to have any effect or big effects on broad money growth, let alone output and inflation. So to correct for these demand effects in this paper, Mike and I follow some recent work by Jack Tatum and we simply subtract off from the Fed Reserve, Bank of St. Louis is measure of the adjusted monetary base excess reserves. Since it was excess reserves that were minimal before the financial crisis, they've bloomed since then, but the excess reserves have been locked up, so to speak, by the Fed's simultaneous decision to pay interest on them. So again, as you can see from the graph with this adjustment made, pro-cyclical patterns in money growth continue to appear even after 1980 sharp declines in money growth appear to pre-sage each of the last three cyclical. In this next set of tables, we aim to quantify the relationship seen visually in the graphs we were looking at just a moment ago in the following way. We take our two measures of money, and in this case, nominal GDP, and we pass all of these three variables through a time series filter intended to isolate their cyclical components. Then we look for the strongest positive correlation between the cyclical component of nominal GDP and those of quarterly lags of our measures of money, and we report that strongest correlation together with the lag at which it can be found in the table. So two takeaway points here. First, notice how moving from row two, a sub-sample before 1980 to row three, a sub-sample after, the lags needed to find the strongest correlation between money and nominal GDP, they lengthened considerably, and that's something I'll come back to when talking about our final results at the end. But second, and perhaps even more important, notice how focusing even on the most recent sample of data since 2000, the correlations between our measures of money growth and nominal GDP are as strong as, if not stronger than they ever have been. That's important because many observers have associated the period of zero interest rates in the United States with the Keynesian liquidity trap. But if that analogy were exact, what we should expect to see are large movements in money with no relationship to subsequent movements in nominal spending. Instead, exactly the opposite appears in the data. Here are similar statistics, but correlations with a measure of aggregate prices, the GDP deflator, and you can see the patterns are exactly the same. The only difference is that the lags needed to find the strongest correlations between money and prices are even longer than those for nominal GDP, and again, I'll come back to that point at the very end. So in order to explore in more depth the dynamic relationships between money growth on the one hand and nominal GDP and prices on the other, we adopt in this paper the framework of the P-STAR model. Some of you may remember the P-STAR model was a single equation, empirical model of price level determination with explicit quantity theoretic foundations developed at the Fed Reserve Board in the late 1980s. The academic style paper was published by Hallman Porter and Small in 1991, and a first footnote to that paper, credits Allen Greenspan for coming up with the idea. The reason I mention this detail is just to highlight the fact that if you go back in time to the late 1980s or even the early 1990s, you could still find in those days papers in leading academic journals that did take a quantity theoretic approach to monetary policy analysis, and you could still find leading policymakers at the Fed and other central banks around the world with an expressed interest in developing quantity theoretic alternatives to monetary policy evaluation. So even though the quantity theory has been set aside for a while, doesn't mean there's no point in bringing it back, especially in the aftermath of an episode where many of us feel that monetary policy may have gone off track in one way or another. Having a quantity theoretic alternative is useful for no other reason than uses a cross check like David Laidler mentioned earlier, just to make sure that monetary policy conducted in a more conventional way doesn't go off track again. And that's really the spirit in which Mike and I offer up this research to you and others today. Whoops. So the P-STAR model is, as I said, it does have explicit quantity theoretic foundations. It's built directly on the quantity equation. It rearranges the quantity equation, like at the top of this slide, with P isolated by itself on the left-hand side, that invites us to think about policy-induced movements in the money supply then feeding through, affecting prices and other nominal variables economy-wide. The P-STAR model then assumes that there's some long-run equilibrium level of velocity, V-STAR, towards which actual velocity tends to converge over time, and real GDP converges to potential over time, so that if one calculates P-STAR, that's what gives the model its name, the price target dictated by the current level of the money supply, once V and Y converge to their long-run levels. What the theory expects you to see is a convergence of the actual price level towards P-STAR over time. Now, the original P-STAR model was developed as a model of price-level determination, but we can, and do in the paper, extend the P-STAR model or adapt it to handle the case of nominal income targeting as well. As a matter of fact, with nominal income targeting advantages, you don't have to rely on any estimate of potential output. All you do is to calculate the nominal income target, X-STAR, dictated by the current level of money and your current estimate of V-STAR, and then as V converges to V-STAR, you expect to see actual nominal GDP converging to X-STAR. Now, in their original formulation of the P-STAR model, Hallman-Porter and Small use simple sum M2 as their measure of money, and assume that V-STAR was a constant. Here we do something slightly different. Instead, we allow for time variation in V-STAR, driven by low-frequency movements and interest rates, payments and changes in the payments system, and so on and so forth, and we estimate that time-varying V-STAR using the trend component from a one-sided version of the Hodrick Prescott filter, one-sided so that the approach can be implemented in real time, using data that are actually available when policy decisions need to be made. Now, going back one last time to that theme, measurement matters, here's a graph of simple sum M2 velocity, and it shows what, again, many of us remember. Before 1990, simple sum M2 velocity was pretty close to constant. Then, in the early 90s, it shifted upward, and that's what threw the original P-STAR model off track and led to its abandonment. But look what happens when you calculate divisia M2 velocity instead. If you were looking at the appropriately constructed monetary aggregate from the start, you never would have gotten the idea that velocity is constant. Instead, you would have done from the start what we're doubling back and doing now, which is to come up with a way of estimating, adaptively, in real time, a time-varying value of V-STAR. So here, going back to our preferred divisia MZM measure, the blue line is actual velocity, the red line comes from our one-sided HP trend. You can see that there are gaps, but the gaps are small and they get closed pretty quickly. That suggests to us that the approach shows promise. Here is the analogous figure for the adjusted monetary base. And finally, here is one of the gap variables that we construct with our P-STAR model. This is just an example, nominal GDP targeting with divisia MZM, but the other gaps constructed with the adjusted monetary base are for the price level. They look quite similar. You can see those in the paper. So again, how do we construct the gap? We begin by constructing or calculating the nominal GDP target dictated by the current period's level of divisia MZM and our current estimate of V-STAR. And then we calculate the percentage point difference between the nominal GDP target and actual nominal GDP. And we interpret periods where the gap is negative, like in the early 1980s, as periods when monetary policy was contractionary, exerting a downward influence on the growth rate of nominal variables, and periods with a positive gap, like at the depths of the financial crisis, is a period when policy was expansionary, reflating the economy, so to speak. And to test the theory, we run regressions of the same form used by Hallman, Porter, and Small. Here, little p is the log price level, so delta squared p is not inflation, it's the change in the inflation rate. So the change in inflation rate is being regressed on four quarterly lags of itself, plus the lag value of the price gap. The null hypothesis, which the theory tells us that we want to see rejected, is that C is equal to zero, we expect, in other words, according to the theory, to obtain adjustment co-efficiency that are positive and statistically significant, implying that positive gaps put upward pressure on inflation, negative gaps downward pressure on inflation. Here is the analogous regression for nominal GDP. Again, change is a nominal GDP growth on four of their own lags, plus the lag value of the nominal GDP gap. Here at last are our results. So this is phenomenal income targeting with Divizia MZM. Notice, first, the adjustment coefficients positive throughout. Notice also that moving from an early sub-sample before 1980 to after, the size of the coefficient goes down, reflecting slower adjustment, the longer lags that we saw in the reduced forms earlier. Finally, notice how the p-values allow us to reject the null hypothesis that C is equal to zero at a very high level of statistical significance, except possibly for the most recent period. There, the rejection is simply at the 10% level. For that reason, I just wanted to show you some quick checks for robustness. Do the same tests for the same recent period, but with Divizia M1 or M2 instead of MZM, you rejected an even higher level of significance. So this is a robust finding throughout. Same thing with the monetary base. Again, positive estimates of C, the magnitudes decrease before versus after 1980, reflecting slower adjustment, statistically significant coefficients throughout. So we view these results as quite positive, suggesting that the Fed Reserve could achieve or could implement successfully a nominal GDP targeting strategy based around its ability to influence either the monetary base or a broader Divizia monetary aggregate. The results turn out to be a little bit different for price-level targeting, though. Price-level targeting for Divizia MZM. We're able to reject the null hypothesis that C is equal to zero in favor of the alternative that C is positive for the full sample and for the early sub-sample before 1980, but not for the late sub-sample since 1980 and for the most recent period, the coefficient actually turns out to be negative. Results are a little bit better with the monetary base, but not much. At least the coefficients are positive throughout, but basically it's insignificantly different from zero for the more recent period. So just to wrap up and just to be clear, we do not interpret the negative results for price-level targeting as implying that inflation is not always and everywhere a monetary phenomenon. We believe that it is. Instead, we refer back to the lags and the reduced form correlations that I showed you earlier and conclude that these lags are too long and probably too variable as well to be fully captured or adequately captured by a simple time series model of the kind that would be needed to run a successful targeting strategy. But again, the results for nominal GDP seem more encouraging. The links between money growth either as measured by the monetary base or a broader divisive aggregate nominal GDP seem strong enough they remain strong even through the financial crisis and the great recession to make a nominal GDP targeting strategy built around the quantity variable a real possibility. Thank you. Thanks very much, Peter. So I'll start it off with just a few questions while my first question is actually for you, Peter and David and Miles, if you have some thoughts on it, I'd love you to offer them as well. And the essence of both of your papers is that we can come up with a money-targeting rule that then restores output and nominal income onto some optimal path. And the big question in my mind is that what is the mechanism by which expanding the money supply, which the central bank still has a lot of control over, translates into aggregate spending? I mean, that's the part that's always missing for me and that seems to be the essence of the liquidity trap. You can print and print and print money but you can't actually force people to spend. Walk me through that. How does actually just doing that, especially when you're stuck at the zero-bound, actually generate spending? That's an excellent question. I mean, I guess in response to your skepticism I would first refer back to the correlations that I showed on the slides. I mean, as an empirical matter, it does seem like, even over the most recent period, changes in money, either a broad-division aggregate or the adjusted monetary base, have been followed by changes in nominal GDP. On a less formal level, I would note, for example, that if you take a look at broad money growth since the financial crisis, although the Fed reserve through QE has dramatically expanded the monetary base, the amount of broad money growth the Fed's policies have produced has not always been excessive by any means and at times, in fact, money growth has dropped. In 2010, when the Fed reserve pulled back from QE1, broad money growth fell precipitously and so the quantity theory would suggest that slow money growth during the recovery may be a big part of why we have sluggish real GDP growth and slow inflation since then. So, my first response would be to say that even if we don't yet have a fully-formed model, like the New Keynesian model, that links changes in money to changes in nominal spending. In the data, those connections do seem to be there. David? Two points. First, the whole idea of level targeting is that the central bank has committed to make up for past mistakes, so it's a more serious commitment to correcting past misses of the target. But as part of my proposal, I think there is a credibility issue if you're worried that why would people take the central bank seriously? That's my whole Treasury backstop justification that if the Fed did nominal GDP-level targeting and it didn't bring about a recovery in nominal demand, the Treasury steps in and it's the threat of doing it. I don't think it actually would have to do it, the threat of the Treasury stepping in and backstopping the target. I mean, systematically doing helicopter drops and again in a rule-based, automatic fashion that would make this very credible. You wouldn't have to worry about it because you knew no matter what happened both the Fed and the Treasury were committed to this target. I think of these quantity rules as being helpful partly because they're a way of telling people you're going to do big enough interest rate movements to get the effects. I mean, as far as the theoretical issue, it's really remarkably easy to explain why interest rates will have the effects they have on the economy. You take every single borrower-lender relationship in the economy and if the interest rate goes down in that borrower-lender relationship, you have several things going on. You have a countervailing wealth effect. If the interest rate goes down, now the borrower is better off and the lender is worse off, but borrowers tend to have a higher propensity to consume the lender, so that's stimulus. And then you have a substitution effect, which just means it looks cheaper to spend now relative to later, so that's some extra effect. That's true for every borrower-lender relationship. Sometimes people say, oh, lower interest rates are going to make people save more, but they're forgetting that those savers have somebody on the other side of that transaction who's borrowing from them, and that has a wealth effect going the opposite direction. So once you have that principle of countervailing wealth effects, it's really very clear that interest rates have the effects they do, except in the very, very rare case where you have borrowers with a lower propensity to consume than lenders. If I hear you correctly, Miles, you're essentially saying that the money-targeting frameworks that David and Peter are talking about can be useful as commitment devices. The ultimate lever is still the interest rate, which takes me back to my original point. So long as the interest rate is still stuck at zero, I don't see the mechanism by which just telling people that you're going to print money gets them to spend, and this is a practical question, because I'm a journalist, so part of my job is to report on what central banks do, and the public reads my article, and they respond to it. Well, back in the time of Volker, when Volker said I am going to keep the interest rate in double digits until inflation rolls over, I can conceptually understand how many people, you know, because unemployment kept going up until inflation went down. In this world, I don't see that connection. Let me respond to that. Again, if we had the nominal GDP level target, and let's say we implemented via a Taylor rule, so we were using interest rates, and let's say for some reason we hit the zero lower bound, again, I don't think we would with a level target, but let's say for the sake of argument we did, and so they're out of ammunition is your point. What do they do? What would happen is the Treasury would step in, it would sell the bond to the Fed, get the currency and start sending checks to households. That's the mechanism. And it's that commitment, that credibility that would, in my mind, automatically get the public to start spending velocity would go up. And so when you've got the backing of the Treasury and the Fed, it significantly raises the credibility. So there is a mechanism, and the expectation of that mechanism by itself should be corrective. But that's not really monetary policy. That's fiscal policy, isn't it? It's almost like a monetary... It's an NGDP target for the fiscal authority. I might be throwing out the room for it, but it is a little bit tapping into fiscal policy, but I would stress again, though, it's a commitment more than anything. If you know the Treasury is there, you don't really have to tap into it. A new version of the Treasury, the famous Paulson Bazooka. A little less radically, if you are focusing on monetary aggregates the way Peter is talking about, that might be a way to get people to expect that you would keep the interest rate longer at zero than they otherwise would. And so that could have an effect. I think it's better to have the interest rate go down a lot for a short period rather than long periods of zero interest rates are very bad for savers compared to short periods of negative rates. But it's still better to... If you do decide to keep the zero lower bound, which as a policy choice, then talking about the monetary aggregate could be a way to get people to believe that you'd keep the rate at zero for a long time. If I could try one last thought experiment that might help, imagine if you have $100 in your pocket right now, your interest rates are zero right now. And I give you an extra $100. You now have $200. Now, maybe you're not going to go out and spend all of that money right away. And actually, what would increase the chances of your not spending the money right away would be, if you thought the Fed would reverse the actions that increase the money supply in the first place, so there wouldn't be an increase in the price level, then you could just hold on to the money. It would be as good as having the money in the bank or buying government bonds or stocks or whatever. But if you knew in advance, as in David's framework, that the increase in the money supply was permanent, so that not only did the money in your pocket double, but the money in everybody's pocket doubled, and so sooner or later, prices are going to be twice as high as they are today, then you're going to feel more of an urgency to spend the extra $100 before the prices rise. So that would be a traditional quantity theoretic explanation of how alone an increase in M, irrespective of what's happening with interest rates, could affect nominal spending. Well, if I could just, but I just want to respond to that, because the mechanism by which you get that $100 into my pocket is critical, right? Because that essentially is what David's recommending, as I said, as we agree, that's fiscal policy. That's not really a monetary rule. I mean, when the central bank operates on the monetary base, they're out there buying assets and creating money, which ends up in the banking system, right? The banking system, in order for that to actually span broad money and credit aggregates, has to find somebody to borrow, right? But if, you know, risk aversion is so high that nobody wants to borrow, that is the essence of the liquidity trap, unless the fiscal authority overcomes that risk aversion by going ahead and doing the spending. Again, it's clear from David's analysis and other analysis, it's clear throughout monetary history that, I mean, we say that, you know, inflation is always and everywhere a monetary phenomenon, but it's just as clear that fiscal considerations play a role in what happens to the money supply in the long run. Hyperinflations are at the opposite extreme. They occur not because a central banker says, gee, maybe there's a Phillips Curve and we can really exploit it by having, you know, 10,000% inflation per year. It's usually because there's a budget crisis and it's usually because there are, you know, it's in the aftermath of a war where there are men with guns that want to be paid, and the only way to pay them is to print money and hand it over. But from a monetarist's perspective, the key difference between fiscal and monetary policy would be that fiscal policy involves the total nominal value, the nominal value of total outstanding government liabilities, including government bonds, as well as money, as well as Fed Reserve notes and reserves. So I would contend that even if it was pure monetary policy, so that instead of giving you $100, the Fed Reserve bought a $100 government bond. So you're without a $100 government bond, but you now have $200 in your pocket. If somehow the central bank convinced everyone that the money supply had permanently doubled, everyone would expect the price level to double, and that expectation of faster inflation is what would bring about the change. Greg, one last thought. I have in the paper the case of Israel during the Great Recession. It approached the zero lower bound and just barely missed it, and the reason is because it did have a permanent monetary base injection. I did it through massive foreign exchange intervention, but I showed in the paper the graphs there's a clear one-time jump in the path of the monetary base, and Israel had one of the best performances during the Great Recession. Israel is the only one that did better, and it effectively followed the prescription of some of the proponents, Lars Finston's proposal. How do you permanently increase the monetary base? He says go out and intervene in the foreign exchange market. Bennett McCollum, similarly. So it does work. There are cases where it's been done. It does work. Interesting. Miles, a question for you. So you've talked about how interest rates are really the superior instrument, and we should be using them more aggressively and creatively, actually. Once again with the zero normal bound, interesting ideas about how to overcome that, which amounted negative interest rates in the withdrawal of cash. You talked about Ken Rogoff's new book, and we actually had an excerpt in the Wall Street Journal of that book a few weeks ago, and I watched the email and the tweets and the comments on the website, and they were running about 99 to 1 against Ken, mostly calling this like just a terrible, evil, egregious variant of government theft and invading our pockets, vacuuming out our money for all sorts of evil purposes, robbing us of the last bit of privacy. So I suggest that whatever the economic arguments, which are substantial, are, is that you've got a big political problem there, and I would add that negative interest rates seem to be having extremely negative and perverse effects on bank equities. Every time in Japan or Europe, the prospect of negative interest rates comes up. There are stocks tank, and that can't possibly be good for the monetary transmission mechanism. So my question to you is, what is the answer to these very practical problems of actually getting to the theoretical benefits that you've outlined? Well, this is exactly what my, the talks I'm giving central banks on, and so that there's three hours worth that I'm convincing central banks to listen to, but basically, you know, first of all, I think the politics are very interesting to talk about there are some very basic things you want to do that aren't being done yet. One is you want to be very clear that people with small checking and savings accounts aren't going to see negative interest rates in those accounts, and that's really remarkably easy to do just with a tiered interest on reserve formula that's linked to the, you know, the banks voluntarily providing evidence of the amount of these small accounts that they have, and so you assure people they won't face negative interest rates in their checking and savings accounts, but they will see lower positive rates for, you know, getting an auto loan and mortgages and so on, so that helps. This effective subsidy for the zero interest rates on small accounts is also basically money you're throwing to banks. The other thing that helps the banks a lot is to lower the paper currency interest rate, which is something that no central bank has done yet. So, I mean, a lot of these are the zero lower bound in action. As it is now, what is the zero lower bound made of? The zero lower bound is not yet made of people storing large amounts of cash. I think we're a ways away from that. What the lower bound is made of is concerns about bank profits, and you do get concerns about bank profits if the banks have to worry about people putting their money in cash. So, it's not sort of large whole scale storage of cash, but banks worrying about that. And so, it is important at some point to lower the rate of return on cash. The other thing to say about the politics is, I mean, look at the politics for central banks of what we actually did. I mean, the best politics for a central bank is to get the job done and get quick economic recovery. I mean, why is the Fed unpopular and their chance of end the Fed? Two things. One is the Fed's role in the bailouts, but that's not the only one. The fact that it took us so long to get recovery is a big part of it. If you'd had quicker recovery, then, you know, of course people would have complained about the negative interest rates along the way, but with quick recovery, that would have all blown over by now. As far as the other practical problems, I mean, a lot of these really are strains from not lowering the paper currency interest rate because if you lower all the interest rates in tandem, including the paper currency interest rate, banks live on spreads as long as the spreads are normal, the banks should be fine. Even in terms of what's actually happened, it's not the bank's profits haven't gone down yet. It's fears about what will happen to bank profits in the future, and so it's banks being afraid, oh, no, maybe they'll leave the paper currency interest rate at zero and we're going to have problems in the future. So I think a key remedy is to have the central banks express a willingness to have the paper go-off par and then it's not so hard on bank profits. I'd love to get some questions now from the floor. If you have a question, I think we have a microphone right over here, yeah. I was going to start out by saying state your name and affiliation, but I don't think John Taylor or Stanford wants to do that. John Taylor or Stanford. So my question is related to Peter's presentation mainly and also Greg's question. If you took what I call a New Keynesian model, a good one which would have a money demand equation in it in terms of we threw many of those out in the past, but I have some good models like that. You could simulate a money growth rule. In fact, we used to do that constantly and then we changed the interest rate rules. But in that context, you can have good performance assuming you have reasonable stability of money demand or a V. So it's really not surprising in that context. And I think that answers Greg's question too because if you simulated an increase in money or an increase in money growth, you can trace through the long-term interest rate effects, the short-term interest rate effects, the exchange rate effects, and there's a whole wide range of financial influences. The transmission mechanism is not just a funds rate. It's whether it's stuck at zero or not. So it seems to me that's the way to think about this. And that brings me to a question is, what is the policy recommendation that you come from this? Is it fixed MZM? Is it feedback MZM or whatever? I think that's kind of the question I'd like to get the answer to. Right. Thanks, John. Well, there's a two-part answer to that question. If you wanted to take the approach seriously and offer it up as an alternative to what the Fed Reserve is doing now, what you would do, it would be a kind of flexible monetary targeting where you would figure out what your target path for nominal GDP was, and then given your current estimate of the long-run velocity, you would then trace out the target path for the broad money supply or the adjusted monetary base. But you would also, along the way, because you're allowing for time variation and velocity, you would be willing to adjust the target in response to changes in velocity. So in other words, you would be adjusting the growth rate of the base or adjusting the growth rate of a broader monetary aggregate in order to keep nominal GDP on target. Another way that the framework could be used, though, is, as David Laidler referred to earlier, in a manner that's similar to the way that the two-pillar approach was supposed to work at the ECB, where you use it as a cross-check. You let the FOMC do something like Bob Hetzel recommended, a more systematic macro forecast strategy built around control of the federal funds rate as always. But you ran this thing in the background, and if you saw money growth going off track in one direction or another, you would pause and ask why. Can I just add parenthetically, or is there something about your presentation and your remark that I've been thinking about? Going back to Canada in the early 80s, the Bank of Canada was following a money-targeting rule, just like the Fed, and they also abandoned it. General Buie, the governor at the time, famously said, we didn't abandon M1. M1 abandoned us, which was another way of saying that velocity became completely unstable. And if I carefully look at your slides, it does look like velocity fell a lot after 2008, and it hasn't gone back yet. So we could be waiting a very long time for that long-run velocity relationship to, you know, reassert itself. And in the meantime, I'm not sure that we're actually getting to your target. Well, to answer that, I mean, but our V-star is tracking velocity downwards. So this isn't old-fashioned monetary targeting where you would set M, and then, and even though velocity has shifted up or down, you do nothing. Velocity falling would mean you need to increase the growth rate of the money supply even more. So what you're getting at is kind of what are the practical implications of this approach now. I am sympathetic to the comparison between rates predicted by the Taylor rule and the current level of the federal funds rate very low, and worried that that might mean that the Fed is falling behind the curve. But one of the things that makes me less worried than I think many people, others are, you included, John, is the observation that, you know, when you combine money growth, like, of about 6% per year, which is what we're seeing, combined with declining velocity, it looks to me like monetary policy is either about right or maybe even a little bit too contractionary even now. So again, using it as a cross check, you know, you could imagine Janet Yellen with an approach like this saying, like, look it, I know that according to the Taylor rule, interest rates should go up, but I'm worried for the following reasons. Inflation is below target. Money growth seems insufficient given the decline in velocity. We want to wait a little longer. The difference is that then you're constrained in the future if money growth does begin to accelerate and velocity turns around and you say, well, now's the time to act. Next question. Dennis Campbell, Independent Investment Advisor. Let me add to Greg's skepticism about the transmission mechanism by introducing the element of debt. Isn't it quite possible, even likely, that in a world of high individual, consumer, and corporate debt, there is no, there's an interest in paying off debt, not consuming more businesses don't particularly want to invest in such a world. Therefore, monetary policy becomes largely ineffective. David, why don't you handle that one? Well, I will refer you to Kevin Shitty's paper at the Brookings. One of the advantages of nominal income targeting is it provides for better risk sharing. So such debt burdens do not create problems for the economy. That's the first thing I would say. And I have this in my paper as well, that nominal income targeting shifts risk in a more equitable manner across both creditors and debtors. The other thing is it's not so much, I like to view not so much as a creditor-debtor issue all the time as money demand. You want to decrease the demand for money when there's been a shock. No matter what drives it, a debt shock, some other external shock, when money demand goes up, you want some kind of offset to that. And you can do that by raising expectations of the future price level. If you want to address the question of debt. Okay. Next question. I'm Derek Timi with the FDIC. So my question is, to the extent that printing money and mailing it to people is an option for conducting monetary policy, why wouldn't that be more frequently employed as an alternative to doing it through the banking system? That seems kind of... I do have a blog post, helicopter money is not the answer. But you know, mostly, you know, once you've eliminated the zero lower bound, I think the attractions of helicopter money wouldn't be so great. I mean, I think helicopter money is a much, much more radical approach than having a non-zero paper currency interest rate. I just want to clarify the question. Were you saying, wouldn't it be a great thing to use it more often? Or are you saying, wouldn't it be worrisome if we started using it a lot? I think you're saying it would be abused. Is that the issue? No. So I'm really saying I think it could... I don't see why that wouldn't be more effective than just lowering interest rates by buying bonds or something. In other words, why don't we just turn over stabilization policy to fiscal policy through one of these? Let me be clear again. You can think of as an increase in automatic stabilizers, something that you don't go to Congress that just automatically kicks in, and the details need to be worked out more. But my idea is that the Treasury, maybe once a year, some automatic mechanism where it meets, if it's below the nominal GDP is below targeted, it kicks in if it's above target, it kicks in in an opposite way. But again, my belief is that just having that mechanism available would mean that velocity would automatically adjust and you wouldn't have to actually tap into it. Next question. Tracy Miller, Mercatus Center. My question is for David. It's your basic... your trend of the bank reserves. So your trend shows bank reserves going down, starting in 2016, from the current $4 trillion level down to like $1.5 trillion. So my question is, is most of that going to be through things like mortgage-backed securities that actually mature, or does it involve having to actually sell off Treasury securities, which you could raise questions about whether that's going to happen. Okay, so just to be clear, that's the Fed's projection, not mine. Right. But under their exit proposal, their exit strategy in 2011, 2014, it would largely come through just not reinvesting their principal payments. So the debt would mature and at some point it would shrink. In fact, between QE1 and QE2, this was a problem. The Fed's balance sheet did shrink and it made them nervous and so they started reinvesting. So it actually... Bernanke's talked about this. It actually could take place relatively quickly. The Fed has been hesitant to do it because they're concerned about financial conditions and the disruption it might make. But Janet Yellen at Jackson Hole, she reiterated her commitment to doing this in her speech. So for now, the Fed seems committed to doing it, although there's been a lot of push. It was interesting that Ben Bernanke, he had a blog post after the conference and he's pushing for now keeping the balance sheet permanently larger. As Fed Chairman, he made several speeches emphasizing it's going to shrink, it's going to shrink, it's going to shrink. So he's taken an interesting twist and I don't know where this is all going to end up, but even if it is permanent, the interest on reserves is effectively going to create the same effect as if it had been reduced and that is effectively sterilized. Sure. John and I would say that Simon Potter in his speech at Jackson Hole said, we're pushing that barrier out to 2018. Well, you're right that they were promising 2016 to start normalization. Simon Potter now says 2018 and I intend to believe him because he's the guy who will do it. That's right. If you look at the annual reports where I drew that graph from, it's been since 2010 and as I think I mentioned, every year it's been pushed out. So the Fed's committed to it and the market believes that otherwise inflation would be skyrocketing. So the Fed's committed to it and it may take a long time before we get there. We have one last question from the floor. Thanks, Rich Dancker. I had a question on the same topic, I guess mainly for Professor Beckworth. Do you think there's a possibility that the temporary nature of the stimulus the Fed's balance sheet is actually harming growth? In other words, is the fact that velocity is in a free fall and that people are perhaps delaying spending purchases until prices might fall when the Fed chops the balance sheet back to normal? Is that actually hurting real growth? Is there a possibility of that? Everyone, the literature you cited seems to say that the temporary aspect is at least ineffective but I haven't heard anyone necessarily make the case that it might actually be causing growth to do worse than it otherwise would without such a large balance sheet. I do think that it's prevented a robust recovery at least in nominal income growth and maybe in 2009, 2010, 11 it would have also supported real growth at a faster pace but this is kind of QE paradox by doing QE as a way to get around the zero lower bound they're effectively preventing a recovery in nominal expenditures because the balance sheets are so big they cannot credibly commit to making those permanent. I didn't mention this but in terms of how much would they need to make permanent so in 2009, at the bottom of the business cycle it would have been based on where the using currency as kind of a baseline for where kind of the truly permanent level of the basis they would have injected 67.5 billion so small change compared to what actually has been done even as recent as the end of 2015 just at the back of the envelope they would have had to have increased the monetary base 150 billion above that baseline trend very small, you don't need a lot of permanent increase in the monetary base to have the effective raising nominal demand to its pre-crisis trend path. I'm going to finish off with one quick question of my own because I'm a journalist so I'd like to ask each of the panelists should the Fed raise interest rates this year given the way that you look at the world and your models? I'll start with you, Peter. My own, the short end, no. Okay, well you actually laid out the reasons why money growth is slowing down it seems that policy is too restrictive already, right? I mean I think I'd rather say what I said about them raising their rates the first time which was that they should not raise rates until they say they would quickly lower them again if need be so I think that the whole next rate raise becomes much less portentous if rather than saying oh we've got this policy where there's this huge penalty on changing directions no if we find out we were wrong we'll go back the other direction and so I would like to take the whole portent out of this big deal about whether they raise the rates away by saying they ought to be data dependent in both directions Okay, I assume that we've removed that portent and they're saying this is for as long as the data is good, should they do it? Should they raise rates? They know a lot more than I do about exactly where the rate should be I'll say no, I mean they don't want to get ahead of the recovery as Scott Sumner said earlier you want to have robust nominal growth before you attempt to raise rates otherwise you will just repeat what happened in Europe in 2011 I wouldn't have thought it coming in I'm a pretty dovish panel I just moderated thanks very much, this is really great Hang around Thanks Greg, that was a great session One of the challenges of being an MC at an event like this is that you're not really supposed to ask questions or intervene and my handlers are very strict about making sure that I don't do that because as those of you who know me may guess I've got plenty of times when I would like to say something however, I'm now told that I have a few minutes to close up boy would I like to just tell you what I think and have it be the the last word I won't do that but I thought I should add something to all that's been said today there's so much to mull over I hardly can pretend to say anything more profound than what's already been said but one thing occurred to me that I think we should think about and that is that it is after all difficult to assess the potential merits of monetary rules using evidence from a rule less past which is what we've had of course for most of the post war period and before that for that matter most of the time I wanted to just play at devil's advocate by taking the monetary rule that probably has been most readily dismissed today and saying a word or two about that that's Milton Friedman's good old K percent rule it was said earlier that we didn't abandon that the fed claimed it didn't abandon money growth or money rules M1 but that M1 abandoned it I don't think that's actually true and Leland Leland Jaeger who I wish could be here has written about this what if the fed had in fact implemented a straight Friedman money growth rule when he first started advocating such a thing and when the demands function for money was or appeared to be stable even measuring money the old fashioned way well we might not have had the reputation of the mid to late 70s or a lot less of it and then perhaps we wouldn't have had the appearance of money market mutual funds the disintermediation of banks as a result of their competition with those mutual funds with regulation Q all of those things are big factors in the destabilization of money demand or velocity or whatever you want to call it and of course once that happened the demands were no longer something anybody would argue seriously for including Milton Friedman who was not very who was very quick to abandon what he had formally recommended and to start recommending other alternatives so there's a kind of sort of backwards Lucas critique that I'm appealing to here right yes sometimes if you try to exploit what you think of the relationship you destabilize it and then it ends up being a very bad thing to do we learned that with the Phillips curve we learned it once I hope we learned it again but sometimes it's just the opposite if you don't act on a relationship that's there then the relationship goes away and you can't act on it anymore so I just wanted to add those two cents to the discussion so what difference it makes but I think it does suggest that there's a real challenge in trying to assess rules where we really have to rely mostly on hypotheticals and what ifs that are after all rather difficult exercises to undertake so thank you all again