2023 Financial Markets Conference - Old Challenges in New Clothes: Outfitting Finance, Technology, and Regulation for the Mid-2020s

Welcome and Morning Keynote with Mervyn King

The Monday morning keynote was the opening event for FMC 2023 and featured welcoming remarks from Atlanta Fed president Raphael Bostic and a virtual address from Lord Mervyn King, Baron King of Lothbury and former governor of the Bank of England. He was introduced by Julia Coronado, founder of MacroPolicy Perspectives and a clinical associate professor of finance at the McCombs School of Business at the University of Texas at Austin.


Raphael Bostic: Good morning, everyone. Welcome to a southern state for a morning event. It's really good to see you all, and I want to welcome you to our latest edition of the Atlanta Fed's Financial Markets Conference. Another year has passed. A lot happened this year, so there's a lot to discuss. I'm not going to talk too long. We can just get into it. For those of you who have been here before, I want to say welcome back and thank you for coming. For first timers, I hope that you all find this to be a good experience so you'll come back again.

This is one of my favorite events in the course of a year. I always leave this conference very energized. There are a lot of smart people here. The conversation is pretty detailed, and it really makes me have to think hard about a lot of the policies and the issues that we're wrestling with. I hope that you have the same experience, and that you leave here with things that can inform your work moving forward.

I want to actually just thank you all for being here and being engaged. The things that you do really help us do our job better and understand how our policies are working and what things we might need to look at and look for as we think about our policy actions moving forward.

Over the years, we've talked about new things, and this year will be no exception. We have sessions on Web3, digital currencies, we're going to talk about nonbank financial institutions. There are lots of things going on.

Then for me, another treat here is that I just actually get to sit back and think and really just wrestle with the ideas, take it in, and listen and learn and get away from some of the daily grind. I have to say "some of" because I never really get away from it, but this is great anyway.

Now, the theme of the Financial Markets Conference changes from year to year, but the core purpose of the conference does not. What we're trying to do here is examine the current state and future of financial markets, and how their evolution affects both regulatory and monetary policy. There may not be a more important time to be having this conversation. We have seen so many complications, so much uncertainty in the economy, and it made our job much more difficult. We need to get as much information as we can to understand what has happened, and how we should best respond.

As the last month and a half has really shown, we actually have to be creative about thinking about what might happen or what could happen. Through this pandemic—and especially through the banking turbulence we've had more recently—I've just had to learn that it's an expectation that I'm going to be surprised. Every morning when I wake up, not going to be sure what's going to happen, but we're just going to have to take that and be ready to figure it out and figure out a course of action.

We continue to weather events that expose the perils of perpetual change, particularly in financial markets, and they've really highlighted some of the vulnerability of our regulatory guard rails. I think about the recent bank failures and the fall of Silicon Valley Bank, and the first thing that comes to my mind is just the stunning speed at which money can move when you have social media and instantaneous access to funds. Those apps actually have changed how banking happens, and how bank dynamics can play out. You know, $40 billion leaving a bank in a day is a whole lot of money, and $100 billion that was set to flee really just even put that into perspective.

For me, it really said one thing: that speed has changed a lot. It means that we've got to think about our regulatory infrastructures, because they have not been built for that. If the report from vice chair Michael Barr made clear, and I hope you all have seen it, our processes are actually even being more deliberate, not less, as we came into this. But we're going to change because we have to. The world is making that an imperative. In my district, I'll just say we were already something out of the "yellow to orange" alert. We may be in a more perpetual state of red, because 36 hours is not much time to be able to figure things out and take decisive action, so we're going to have to just make sure that we do that properly.

Now, through the pandemic there's been one word that actually may characterize it. It has been very much on my mind, and that's "uncertainty." Uncertainty stalks the economy, and it also stalks the financial system and the macro economy. We think about the debt ceiling situation today, stresses in the banking system, questions about how much credit is going to tighten, and what that means for the real economy. We have the ongoing war in Ukraine, and we also have questions about how fast inflation will moderate over the next several months. We have to think about all of these things and how they interact and intersect with monetary policy pretty significantly, and I'm really pleased that we ... this is the perfect time, as I said, to start because I'm going to get a lot of insight and guidance from you as to how we should be thinking about this as we move forward.

In terms of the agenda, it's a great program. I'm sure you've looked at it. I'm not going to spend a lot of time talking about it. I'll just highlight a couple of things. The first session, our opening policy session, will look at how monetary policy should have played out as inflation spiked in 2021. We have a distinguished panel that will say a lot about how policymakers, more broadly, miscalculated how persistent elevated inflation would be.

I'm sure you all remember the "T" word, which I'm not going to say, but it did sort of animate a lot of policy conversation moving forward. I really think it's going to be interesting to think about discretion versus rules-based policy in this context, and how much we should be thinking about these things. We'll also talk about the increasing importance of nonbank financial institutions, and these institutions are actually quite interesting because through their actions they can actually expose banks to additional risks, which means that financial stability becomes more complicated when you have nonbanks present. We have to really think about this hard as we think about a regulatory structure.

Tomorrow afternoon, the final policy session will focus on mitigating risks to financial stability in a restrictive policy environment. This discussion will not only focus on financial stability concerns that have already emerged, but it will explore some new concerns that might develop and how central banks might react.

This is really going to be a really interesting conference. I want to thank all of you for being here, but I also want to close my statements by just thanking our staff. Putting these things together is never easy, and they always seem to hit a home run on this. I'm just incredibly privileged that I get to work with such a tremendous group of people, so I'm going to just thank my staff for making this happen. You guys are absolutely wonderful.

Let me also thank the staff here at the Omni Resort. They always do a great job of hosting us, and always make these things memorable so that people do want to come back, and that's a very good thing.

All right. I hope everyone enjoys the conference. Please engage; don't leave any questions unasked. This is a time for us to really be in dialogue together, so that's all I'm going to say right now. I'm just going to now turn the podium over to our moderator for the opening keynote conversation, Julia Coronado.

Many of you know Julia. She is a frequent economic commentator and is founder of MacroPolicy Perspectives and a clinical associate professor of finance at the McCombs School of Business at the University of Texas at Austin. We are so pleased to have Julia with us. She gets asked to do a lot of things, so really glad that you have agreed to join us today. She will be introducing our opening keynote speaker, Lord Mervyn King.

Julia, the floor is yours.

Julia Coronado: Thank you, Raphael, President Bostic, and welcome, everybody. I'm privileged to kick this off. The speaker needs no introduction. Lord King was governor of the Bank of England during the global financial crisis and has just a wealth of experience as an economist on the staff before that, really diving through many cycles. He will be giving us some prepared remarks, and then we're going to have some Q&A. I hope you all are familiar, you have your apps ready. You will submit your questions through the app. I will see them and be able to curate them when Dr. King has completed his prepared remarks.

So, without further ado, please join me in welcoming Dr. King.

Lord Mervyn King: Julie and Raphael, thank you both very much, and good morning, everyone. I'm sorry I can't be with you in person, especially given the temperature difference between London and Amelia Island.

The past few years have not been happy ones for central banks, and the problems have arisen in the two areas that constitute their main responsibilities: controlling inflation, and preventing bank runs and failures. In the advanced economies, we have failed to prevent a surge in inflation that's brought back memories of the 1970s, and central banks now face a dilemma. Do they continue to tighten monetary policy to bring inflation back to their 2 percent target, or do they end further tightening because bank failures and the response of other banks may reduce money and credit growth and hence inflationary pressure? Resolving that dilemma will be the main challenge in the management of the global economy in 2023, and the recent failures of regional banks in the United States have prompted suggestions and debate about the reform of deposit insurance.

Those challenges add to an already crowded agenda for the upcoming G7 Summit in Japan. In my view, they will require serious rethinking of existing policies. Many of those policies are national rather than international in nature, but the intellectual mistakes and their corrections are common to all advanced economies.

I want to focus this morning on why these two objectives of price stability and financial stability have proved difficult to achieve in the period since 2020. Common to both policy areas is the same problem: namely, that central banks have fallen victim to a style of academic research which has given the impression that models can predict the future and hence tell us how to set policy. In my view, the main challenge for policymakers is to recognize that the forces driving the economy are always changing, or to use the technical phrase, the world is nonstationary.

This proposition was put most forcefully back in 1939 in John Maynard Keynes' review of Jan Tinbergen's pioneering study of econometric relationships. The attractions of estimating such relationships, and the number of articles and PhD theses that could thereby be completed, overwhelmed the question of whether it actually made sense to assume stationarity. We live in a world of constant change, a world of radical uncertainty. None of this might matter if models were used in an appropriate way, but sadly, they are not.

Models can produce useful insights into the way the economy works, but they are not literal descriptions of the world. They cannot be. The world is just too complex. Central bank policy, whether monetary policy or banking regulation, must be set in the world and not in the model. Most of the recent failures of policy have reflected a mistaken view that policy can be set in the model. Instead, key insights from models need to be combined with an attempt to understand what is going on in the world. Asking the question "what is going on here?" may seem trivial, but it isn't. It is the essence of coping with the need to make decisions in a world of radical uncertainty.

Let me make this proposition more concrete by talking about the mistakes in monetary policy in the past few years. After 30 years of low and stable inflation, the major western economies lost control of inflation during the pandemic. CPI inflation in the United States is now down to 4.9 percent, having peaked at close to 10 percent. It's likely that inflation during 2023 will fall sharply across the G7 economies. Nevertheless, policy mistakes allowed inflation to rise to its highest level for several decades. What went so badly wrong?

Part of the answer is the sharp rise in food and energy prices following the Russian invasion of Ukraine, but that's not the whole story. Food and energy prices have fallen back. Excluding food and energy prices, core CPI inflation in each of the US, Euro Area, and the UK is running between 5 and 6 percent. Central banks were slow to realize that the rise in inflation was more than a transitory deviation from target.

We are all familiar with Milton Friedman's dictum that inflation is always and everywhere a monetary phenomenon. Monetarism became discredited for two main reasons. First, the relationship between monetary aggregates and nominal incomes proved nonstationary. This told us, or should have told us, less about the role of money and more about structural shifts in banking and the financial system. Second, Friedman and other American monetarists tended to focus on the monetary base rather than broader monetary aggregates which couldn't be controlled directly, but by the central bank.

As we have seen with QE [quantitative easing], base money is relevant to the determination of aggregate nominal incomes only insofar as it affects broader measures of money. As a result, academic research turned its back on decades of monetary theory and decided to develop a theory of inflation without any reference to money at all. Once again, the attraction of writing down such models overwhelmed the question of whether they made any sense.

Unfortunately, inflation is a nominal variable, so any theory of inflation has to be related to nominal variables. The challenge of how to close a model by pinning down the price level or inflation in the medium term was solved by the assumption that inflation was determined by expectations, and that expectations were determined by the official inflation target. In other words, the model assumed that inflation in the medium term would always return to the official inflation target of 2 percent. Milton Friedman's dictum had been replaced by a new dictum that inflation was always and everywhere a transitory phenomenon.

But a satisfactory theory of inflation cannot take the form "inflation will remain low because we say it will." It has to explain how changes in policy, whether QE or changes in interest rates, affect the economy.

For a long while, central banks were successful in keeping inflation close to the target, so nothing disabused them of the strong assumption they were making until the pandemic came along. Following a sharp reduction in potential supply, the immediate consequence of the measures taken to prevent the spread of COVID, central banks around the industrialized world decided to expand demand by a substantial program of money printing through quantitative easing.

Quantitative easing is an expansion of the money supply, although most central banks are reluctant to describe it as such. Unlike its use after the banking crisis a decade or so ago, which was aimed at preventing a fall in broad money resulting from a contraction of commercial bank balance sheets and hence deposits, this time QE created a substantial monetary overhang. Growth rates of broad money accelerated rapidly. In the case of the United States to the highest levels since the end of the Second World War, at an annual rate of 24 percent in the first half of 2021. Aggregate money demand exceeded aggregate supply valued at the current price level.

I'm not suggesting that policymakers respond, or should respond, in any automatic fashion to changes in the growth rates of monetary aggregates. But I do think it would have been sensible to ask in 2021 what is going on here, with broad money growing at 24 percent of an annual rate?

The case for substantial monetary expansion in March 2020 was framed as a response to "dysfunctional markets." But the monetary injection was not withdrawn once financial markets were operating normally. Indeed, QE was expanded further in 2020 and 2021. This was unnecessary.

The lockdown response to COVID-19 led to a significant fall in potential supply, most of which was expected to be temporary. A big hit to supply, albeit temporary, is not the circumstance in which it's sensible to expand aggregate demand. The actions taken to deal with the pandemic reduced the supply of goods and services. Central banks increased the supply of money. Taken together, this produced the time-honored recipe for inflation of too much money chasing too few goods.

The rise in inflation reflected an absence of common sense, a much-undervalued attribute in policymaking. There were at least three different ways to look at the problem: an excessive growth of broad money in relation to real GDP, the one I'd stress, an excessive fiscal and monetary stimulus to aggregate demand, and interest rates well below the level implied by any version of a Taylor rule. All three approaches imply that policy was overly expansionary.

The good news is that after these policy mistakes of 2020 and '21, 2022 was the year when central banks corrected the errors. They raised interest rates and stopped printing money through QE. The result is that monetary growth and nominal demand are no longer expanding rapidly. In the United States, for example, broad money is now declining. In due course, not only will headline inflation fall back as the effects of higher energy and food prices drop out of the 12-month measure, but the domestic component of inflation will also come down.

How quickly this will occur is unclear, but relying on a model which ignores money and other variables relating to the banking and financial sector can lead to errors both on the upside, in terms of inflation outcomes, and also on the downside. Setting policy in 2023 will involve a difficult balancing act.

Let me turn now to the stability of the banking system. Two days ago, the G7 finance ministers and central bank governors met in Niigata in Japan. I don't know how many of you have read their communiqué, so let me quote for you:

"Financial institutions' recognition and full and prompt disclosure of their losses play an important role in reducing uncertainty, improving confidence, and restoring the normal functioning of the markets. In addition, authorities should ensure measured and flexible responses to market stress, including arrangements for dealing with weak and failing financial institutions, both domestically and cross-border."

Very sensible, but unfortunately that quotation does not come from the communiqué of two days ago, which said remarkably little about inflation or banking. Instead that quote comes from the communiqué of the G7 ministerial in which I took part when it met in Tokyo in February 2008. Fifteen years, and thousands of pages of complex regulation later, the communiqué issued on Saturday ignored the lessons of the past.

We've failed to prevent another set of bank failures. Once again, regulators have been making it up as they go along. In the case of Silicon Valley Bank, uninsured depositors were fully compensated, and in Switzerland Credit Suisse shareholders benefited at the expense of the contingent convertible bondholders, an inversion of the normal hierarchy of creditors.

Two features of those bank rescues stand out. First, the absence of a clear ex-ante framework for the provision of central bank liquidity to any institution suffering from a bank run. Second, the willingness of governments and politicians to intervene in the resolution of a failing bank and change the rules in an ad hoc manner.

On the first, a bank run can occur not only from uninsured depositors but from the failure to roll over short-run wholesale financing. That was demonstrated in the financial crisis in 2008. Banks are fragile; they are institutions that borrow short and lend long. As such, they are vulnerable to any loss of confidence, whether justified by the underlying reality or not.

Over several centuries, we've come to rely on the benefits of maturity transformation to finance investments by the issue of short-term liabilities that are used as money. If we want to retain those benefits, then central banks, in my view, must stand ready to lend against any very short-term runnable liability used as money. Ambiguity over whether a central bank will stand ready to provide liquidity makes a run more likely, but a commitment to provide liquidity cannot be open-ended. That would be to underwrite excessive risks taken by banks.

The answer is an ex-ante framework in which banks are prohibited from issuing more runnable liabilities than the central bank is willing to lend against the collateral which the bank can offer. The willingness to fight a fire by the provision of liquidity must be tempered by measures to limit the size of the fire. The terms on which a central bank will provide liquidity need to be designed carefully and spelled out in advance. In my book The End of Alchemy I described such an approach as the central bank operating as a "pawnbroker for all seasons."

The basic principle behind the scheme is to ensure that banks will always have access to sufficient cash from the central bank to meet the demands of depositors and others with claims on very short-term debt. For each type of asset a bank is willing to offer as collateral, the central bank would calculate the "haircut" which it would apply when deciding how much cash it would lend against that asset. Adding up over all assets, it would then be clear how much the central bank would be prepared to lend with no questions asked. The regulatory rule would be that no bank could issue more runnable liabilities than the central bank was committed to lend.

The central bank would promise that the haircuts would remain fixed for a lengthy period, probably several years. That would then give a clear, contingent credit line to banks. Interestingly, the new Bank Term Funding Program announced by the Federal Reserve in March, which provided for lending against the par rather than the market value of bonds used as collateral (a negative haircut on market value), but nevertheless, that is very close to the way the "pawnbroker for all seasons" would set haircuts.

The problem with the Bank Term Funding Program is that it sets haircuts retrospectively. It would be better to incorporate such an approach into an explicitly ex-ante framework which would leave no room for ambiguity about the availability of central bank liquidity to support short-term runnable liabilities. Such a framework would likely have limited the speed at which Silicon Valley Bank expanded its deposits before it got into trouble, and Credit Suisse's travel on a self-inflicted path of multiple scandals might have been caught earlier by a central bank more willing to impose a haircut on some of the dubious assets than was its regulator in dealing with bad behavior.

The need for a clear ex-ante framework for the provision of central bank liquidity really stems from the fact that it's impossible to anticipate the scale of contagion from one failing bank to other banks, as the problems of First Republic demonstrate. People have forgotten that when BCCI was closed by the Bank of England in 1991, an idiosyncratic case of fraud if ever there was one, the concern about the fate of other small banks led the Bank of England to provide support secretly to other small banks. The fact that a bank run can occur today in a few hours rather than a few days is actually less important than the fact that the maturity of its short-run liabilities is much shorter than that of its assets.

The approach I have outlined would ensure that no run could bring down a bank because there would always be cash available to cover all the runnable liabilities. It doesn't make much sense to guarantee all deposits in a bank that fails and yet maintain that the upper limit on deposit insurance remains the same for all other banks.

The right time to introduce this scheme is today, when the expansion of QE means that the deposits of commercial banks with the Federal Reserve are at an abnormally high level. The regulatory reforms that followed the financial crisis, and there were thousands of pages of them, were actually little more than sticking plaster to a system that really requires a much simpler and less costly but more comprehensive approach to the provision of central bank liquidity.

In conclusion, recent failures in respect of both price and financial stability are the result of self-inflicted intellectual wounds. I don't blame central banks or any individuals in central banks for this. The failings, in my view, have been intellectual and are stemmed from academia. For central banks, bringing inflation back to the 2 percent target will be a challenge. That challenge will be all the greater if we continue to rely on models that assume that any deviation of inflation from target is always transitory. Central banks need to monitor developments in money and credit growth and in the financial system more generally. It's very interesting to me to see that part of your conference is going to deal with nonbanks.

Bank runs will continue to occur until we put in place a clear ex-ante framework for the provision of central bank liquidity against collateral, which means that everyone will know that no bank will run out of cash if it is subject to a run. We cannot really rely on resolution frameworks, necessary though it is to have them in place. We can't really rely on them solving problems if governments continue to intervene in that process.

To go back to what I said at the very beginning, radical uncertainty means that models provide us with important and valuable insights, but they are not literal descriptions of the world. In particular, the complex world of money and the financial system does not fit easily into a tractable model. But they do matter. We must always ask the question: "What is going on here?"

Thank you for listening. Now, let me invite Julia to moderate the question-and-answer session. Julia?

Coronado: Thank you, Lord King. We have a number of questions. If you want to ask a question, please submit it. I have several questions to begin with. Let me just start by asking you, Lord King, you seem to still see a lot of risk in the banking system. In other words, you're not seeing SVB as an isolated case that's been taken care of. You still see potential risks in the financial system.

King: Well, there are always potential risks in the financial system. The problem with the phenomenon of bank runs is that they can occur even if there is no underlying genuine risk. It's a loss of confidence. Rumors can spread and this is the great risk and cost of what I called in my book "the alchemy of the banking system." What we can do is put in place a sensible ex-ante framework for the provision of central bank liquidity, which removes that risk and takes it off the table.

I don't think it's a question of what is the underlying genuine risk of the financial system? We're certainly not in a position that we were in in 2007, when both the liquidity held by commercial banks themselves was at a very low end, and also the amount of equity which they had issued relative to total assets was very low. We're not in that position, and in that sense the banking system is a lot safer, but the problem with any bank is that a bank run can occur, and no one could be confident of whether it will lead to contagion to other banks which is why, despite the argument saying that Silicon Valley Bank was an idiosyncratic case, which was unlike other banks and in many ways, it was. That still did not prevent contagion from other banks.

[Audio/connection issue]

Coronado: ... calls by some of the liquidity facilities, and steps that have been taken?

King: Because I just got noise—an echo—coming over the line, I didn't hear the question.

Coronado: Okay, let me start again. You mentioned that it's basically a game of confidence, that SVB showed that deposits can move quickly, and that means other banks aren't immune even if they're different from SVB. Do you see things like the mark-to-market losses on bank balance sheets from higher interest rates, and/or that speed of deposit mobility, as current risks to the system? Or do you think that the actions that central banks have taken largely address those risks, and contain those risks?

King: Well, the risk is certainly there. Only the Federal Reserve has introduced an explicit scheme which would allow banks to obtain liquidity from the Fed by using their collateral in the form of government bonds at par value. Of course, a central bank is different from the rest of the private sector in the sense that it can hold collateral for a long period. Therefore, it's perfectly feasible for a central bank to lend against the par value of government bonds. But where that's not in place, or where there is no scheme which convinces people that this will happen, then that risk is clearly there. It's now time to introduce a proper ex-ante scheme, which would spell out precisely the terms on which a central bank would provide liquidity. Doing it ex-post is far too risky and leads people to make things up as they go along, often over a weekend.

Coronado: Am I correct in understanding that you advocate for blanket deposit insurance?

King: Well, to be precise, I'd abolish deposit insurance as a scheme. I'd replace it with my scheme, which is one in which the central bank would prevent banks from issuing more runnable liabilities—short-run, runnable liabilities—over and above the collateral which they are willing to lend against, and with the haircuts attached. You wouldn't need a separate scheme for deposit insurance in those circumstances because if there was a run on a bank and deposits started to disappear the central bank would simply provide the cash which would enable the bank to pay out to the depositors the full amount of their liabilities to them.

What is very risky is to leave in place a system in which commercial banks lend long term on risky investments and finance it by short-term borrowing, which everyone believes to be safe, and the central bank has to validate that belief in order for the monetary system not to collapse.

Coronado: One question related to this that got many votes. Does the solution of a bank run always have to go through the central bank? Are there no potentially new private sector solutions that could be considered?

King: My view is not in the sense that one of the lessons of the financial crisis in 2008 was that in a big crisis the only source of liquidity is a central bank. The regulation that was put in place—I was actually chair of the Governors and Heads of Supervision meetings that agreed to regulation of liquidity in the new Basel III regime after the financial crisis. What became very clear was that it doesn't make any sense to think about liquidity regulation and the definition of high-quality liquid assets that banks have to hold, without integrating it into the provision of central bank liquidity.

Many of the so-called high-quality liquid assets were unavailable because in countries like Australia, they didn't have as much government debt that they had issued, so it was very difficult for commercial banks to get hold of so-called high-quality liquid assets. Secondly, that in a crisis, things that people assumed would be liquid turned out not to be liquid. We saw that again in 2020. The only way to solve this problem in any convincing fashion is actually to recognize that in a financial crisis, only the central bank can create liquidity.

The choice is either that we have to restructure the entire banking system, by saying that banking has to be essentially based on the ideas that many people introduced in the 1930s known as the Chicago Plan, which doesn't actually then enable the banking system to finance real investments. That's going too far, in my view. The question is, can we by adopting the framework I've described enable the banking system to play its role of generating maturity transformation, but not so much maturity transformation that the central bank is unwilling to provide liquidity to the short-term liabilities that the commercial banking system has created?

Coronado: You've also emphasized, from a macroeconomic perspective, the importance of nominal variables and considering money supply again alongside other factors. Another question that received a lot of votes: How important do you think the 2 percent inflation target is going forward? Is it still the appropriate target, given your observation that inflation is a nominal variable? Do you see reasons to change the target?

King: No, I don't. Again, a lot of academic research which has suggested that the solution to our problems is to raise the inflation target is completely misplaced because it's all based on the idea that the inflation target always has 100 percent credibility. If you raise the inflation target in their models, what you do is lower the effective real interest rate and that's supposed to stimulate demand, increasing growth and raising ultimately inflation if you're worried about inflation being too low.

Again, one of the observations I'd make about the lack of common sense is that many academics argued until 2020 that the big challenge to monetary policy was the need to raise inflation. It was too far below target, inflation had to go up. We saw in 2020, in August, the Fed introduced the asymmetric average inflation targeting framework. This was just hubristic, because the average shortfall of inflation below the 2 percent target on the core PCE measure over the five years prior to that was one-half percentage point a year.

Now, I was involved in the introduction of inflation targets back in the early 1990s. The Fed was rather late to that game. But when we started with inflation target, if you had said to us, "You know what? Inflation might be half a percentage point below the target," we would have said, "This is nirvana. This is fantastic." We'll never be able to get inflation precisely that close to the target and we shouldn't panic if we do. The big contribution of inflation targeting was not to be part of a mechanical model where you had a fixed rule that you always raised or lowered interest rates if inflation was a certain amount above or below the target. Those rules that were introduced have enormous value as posing the question, "Why do we think the current interest rate is correct, if such a rule would suggest it ought to be very different?"

You can't be that precise, and inflation targeting was really all about central banks being much more transparent about the arguments they were using to set interest rates, so the outside world could challenge those arguments and there could be a proper debate. It was a willingness to have a narrative about the economy, the evolution of the economy, which changed over time and central banks would update that narrative according to what they saw in the economy. But it was never the idea that you'd abandon a close and detailed examination of all the data on the economy in order to just follow one simple rule.

It's very important that we just do not fall victim to these arguments that raising the inflation target is the right thing to do and will make sense. The models are not descriptions of how the world, how households, consumers and financial markets, would react to a decision to raise the target. You have to set policy in the world, not in the model. This is one of the damaging implications of falling victim, in my own phrase, to assuming that the academic research is all you need to understand in order to set monetary policy.

Coronado: Along the lines of the "look out your window" approach, you mentioned monetary aggregates. Are you seeing those as a leading indicator? Are you looking at the change in money? Is it the level of money in the system that's important to you? All of the above? How are you seeing those monetary signals, in terms of thinking about the outlook for inflation in the economy?

King: When you see extreme changes in the growth rate of money—broad money, not the monetary base, but broad money—then you should ask the question, "What is this telling us?" Now, there may be times when it's telling you that there is structural change in the banking system or the financial system, and you shouldn't worry too much about it, but you ought to understand and try to convince yourself that you shouldn't be worried about it. You shouldn't just refuse to publish data on broad money and refuse to look at it. It should always be on the dashboard that central banks look at.

You can argue that in 2021 we had an excessive monetary injection. That's now pretty much gone away. Indeed, the concern now would be that have we simply unwound that monetary injection because broad money is now falling in the United States, or are we getting to the point where we're having too rapid a fall in broad money? Is that telling us something? I just don't think you can describe this in terms of a mechanical response, but it should be something that prompts a conversation.

If I take the bank of England in the last few years, if you do a word search on their monetary policy reports for the last few years and you search for the word "money," you won't find it. There's no mention at all. Now, that's unnecessary because such a report ought to take into account the possibility that money matters and is a useful leading indicator. If you want other people outside the central bank to engage in a debate, you have to give them the information about what is happening to money and why you think it either does or does not matter.

That's the important thing: to have a credible narrative. That's how you build up credibility. But simply saying, as I've heard, many, many economists say, "Well, we know that money doesn't matter at all. We can ignore it. We'll just look at measures of inflation expectations," is widely hubristic about the meaningfulness of measures of inflation expectations. Simply, it's irresponsible to ignore major developments in monetary aggregates that we've seen in recent years.

I'm not suggesting that they're used as a mechanical indicator, indicating a precise quantitative response of interest rates to what we see, but it certainly should feed into the decision. My final point on this is that I am somewhat worried that with the demise of the belief in those models—which experience illustrates that just assuming that any deviation of inflation from target must be transitory because that's what the model says—once you abandon that, what central banks are in danger of doing now is setting policy according to the latest inflation number.

Because of the lags in the transmission mechanism between changes in policy and their impact on inflation, the real risk now is that central banks, by focusing just on the latest inflation numbers, will be too late in correcting the course for interest rates. Therefore, it is important to have a feel for those indicators that are likely to be leading indicators rather than the actual inflation numbers themselves.

Coronado: Thank you. What are your thoughts on the potential that structural forces are changing? Things like demographics, things like the state of globalization, maybe shifting to fragmentation. Do you see structural forces acting on inflation as changing, or are we just recovering from a shock? What's your assessment of these questions?

King: The way I like to think about it is that they are always changing, and in unpredictable ways. The rapid move towards globalization and the introduction of China into the world trading system had an enormous impact on the world economy, still does in many ways. That wasn't easy to anticipate or predict.

What's happened in the last few years with the pandemic, with the attempt to improve resilience of economies by ensuring duplication in sources of supply—these things are also changing. They weren't easy to predict, but we shouldn't think of all these things as black swans. They're not; they were all imaginable. I mean, we've had pandemics in the past. We'll have them again in the future, so it's not a black swan. It's just events where it doesn't make any sense to pretend that we can judge quantitatively when or how often they are likely to occur.

Thinking about the world as nonstationary, as a world in which things are always changing, and in which what we have to do is to look carefully through the window in your description, by walking around in other people's description, looking at the world, creating a narrative about what's going on, is fundamentally important. But the insights from economics which are relevant to that are ones which are pretty timeless. If you think that too much money is chasing too few goods, you're likely to get inflation. That's not a literal description of what's going on in the world, but good economic models don't try to do that, we can't forecast remotely easily.

What we need to be able to do is not to pretend that we have a view that inflation or growth is going to be 2.6 or 3.1 next year. We should never think in those terms. We should ask the question, "What could go badly wrong? What are the risks? How can we combat those risks, and take measures to alleviate their occurring?"

Coronado: So, last question. Some modelers in the audience might be smarting from your remarks. The question is: Given policy lags, policymakers have to be forward looking. How do they do this other than by using models?

King: Well, first, you should never use one model. You should be prepared to use a whole set of models, compare the results, and then ask yourself, "Which one do I think is pointing to the biggest risk that we face?" What makes absolutely no sense, in my view, is to pretend that you've got a model that will tell you whether inflation next year is going to be 3.2 percent or 2.7 percent. Every forecast made in January in the last few years has been wrong within months. We've got to accept ...this is not a criticism of models or modelers in any way. It's a criticism of the way some models are used, which is to pretend that I know exactly where the economy or inflation is going. We simply don't.

That's why forward guidance in monetary policy also played a potentially damaging role because we simply do not know where we want to set interest rates. So why pretend that we do? That doesn't mean to say that models aren't very helpful in understanding deep features about how an economy operates. The most valuable model that was produced was by Ricardo in the 19th century when he said ... guess what? He said, "If you think you are living in an economy which is better and more productive in every dimension than other countries', it nevertheless still makes sense to trade with them."

The idea of comparative advantage versus absolute advantage was counterintuitive, but it has turned out to be a very important insight which economists obtained and which most non-economists find puzzling to begin with. The actual description of trade between England and Portugal, which Ricardo used, bore no relationship whatsoever to the reality of trade between those countries at the time he produced his model. The model produced valuable insights which we carry around in our heads today when we confront the world. But it wasn't a way of forecasting what would happen to trade in the 19th century.

Coronado: Okay. With that, join me in thanking Lord King. Thank you for joining us today.