Claudio Borio on the Future of Central Bank Operating Systems

Despite overestimated concerns surrounding balance sheet size and reserve demand instability, central banks should consider the possibility of returning to a scarce reserve operating system.

Claudio Borio is the head of the Monetary and Economic Department at the Bank for International Settlements, or BIS. Claudio is also a returning guest to the podcast, and he rejoins Macro Musings to talk about central bank operating systems and the challenge of large balance sheets at central banks. David and Claudio also discuss the basics and uniqueness of the scarce reserve system, the arguments in favor of an abundant reserve system, the politics of large central bank balance sheets, the possibility of a tiered reserve system, and a lot more.

Read the full episode transcript:

Note: While transcripts are lightly edited, they are not rigorously proofed for accuracy. If you notice an error, please reach out to [email protected].

David Beckworth:  Claudio, welcome back to the program.

Claudio Borio: Well, thank you. Thank you again for the invitation.

Beckworth: It's great to have you back on. It's been a few years since you were on here, and I'm delighted to have you on today, because you had a speech earlier this year titled, *Getting Up from the Floor.* Very clever title, and it's tied to this notion of a central bank operating system called a floor system, and I think our discussion today is very timely, because there has been a growing interest in central bank operating systems. I want to just highlight a few articles and some discussions that we've seen recently on it. In The Wall Street Journal recently, there was an article titled, *Why Central Banks Should (but Might Not) Keep the Market Flooded with Money* by Jon Sindreu.

Beckworth: And this is an article about the operating system, and the fact that there are these growing conversations. Another great article by Daniel Hinge from the Central Banking publication, it was titled, *Corridor, Floor, Other: Are Operating Frameworks Fit For the Future?* He goes into a great discussion of the various options, and going forward, what would work, and we'll come back to that because he has some fascinating insights about tiered operating systems.

Beckworth: And then, let me just mention a few other news stories. Reuters reported recently that the Swiss National Bank lost 12 billion francs in the third quarter, and last year was its biggest loss. The ECB is reviewing interest on government deposits to curb losses, another news story. And of course, the Federal Reserve is well-known right now for its losses on its balance sheet, as well as mark-to-market losses, which are not recognized.

Beckworth: So, there's a lot of conversation and interest in this, so I think it's great to have you on to talk about it. I want to go to your speech, *Getting Up from the Floor,* and begin with a quote from it. You say, "I consider the sea change we have witnessed in monetary policy operating procedures since the great financial crisis, and I do so from the viewpoint of someone who studied them before, and sees current developments with a mix of surprise, and some discomfort."

Beckworth: So, you set this paper up very nicely, and I thought we would begin first by talking about the two different systems, the abundant reserve system, and a scarce reserve system, or what some would say a floor versus a corridor. So, maybe, let's begin with the system that existed pre-2008, before the sea change that you've talked about, the scarce reserve system. Talk about how it operated, and what was unique about it?

The Basics and Uniqueness of the Scarce Reserve System

Borio: Well, thank you again for the question. Indeed, this is an issue which is being increasingly discussed by central banks around the world. I know that all of these things sound awfully technical, but I have to say that my own views about monetary policy, more generally, actually go back precisely to the days in which I was secretary to a group of central banks that was discussing these issues on a regular basis, because they had actually realized that they were not entirely sure how each of their systems was working at the time.

Borio: It made me realize that, a lot of the stuff that I had learned at university about monetary policy, and how interest rates were set and so on, was honestly completely wrong. I had to put it out of my mind and start all over again in this new world. And here, it's precisely in this world of monetary policy implementation that one can fully understand where the power of central banks comes from. But after that very short general introduction, how did the system behave in the past? Effectively, the systems all across the world were what I would call— except for one exception— were basically scarce reserve systems. That means that the central bank was keeping a tight control over the amount of reserve balances that banks hold with the central bank, in order to ensure that there was an opportunity cost of holding them.

Borio: That's, in that sense, in which they are scarce. There is no normative statement about valuing their being scarce. It's just a technical term about making sure that there was an opportunity cost of holding reserves, that the overnight rate, which was the operating target for central banks, was above either zero, or above the deposit facility rate that existed at the time. Two key features of this system… In the way that people typically think of monetary policy, they are not fully appreciated. The first is that, central banks were, indeed, using open market operations, in order to manage that liquidity, to keep it short, if you like, in general. But they were doing that in a way so as not to influence, through their asset purchases and sales, the overall structure of interest rates in the economy.

Borio: The only thing that they were trying to do was to influence the overnight rate. The second key point is that, because the demand for settlement balances, pure settlement balances, is extremely interest inelastic, it's almost lexicographic. First of all, you try to have a certain amount of reserves that you need for your payments purposes, and then you care about anything else. Because of that key feature, a lot of the heavy lifting in monetary policy implementation was done by signaling. The central bank would signal where it wanted to see the rate in a variety of different ways, and because, of course, the central bank ultimately has the control over the supply, is a monopoly supplier of reserves, and banks demanded those reserves for settlement purposes, well, you could basically set the interest rate where you wanted.

Borio: But, the whole idea was that, there was a high opportunity cost of holding reserves, so that banks would transact with each other at the end of the day, in order to clear any sort of balances that they had, excess or shortfalls, and that the central bank would basically try and clear the market at the end of the day. So, it was a way of making sure that the central bank would have a very small footprint in the market. A lot of the work, again, was done by signaling, and the liquidity market operations, open market operations that central banks were carrying out, were very much in the background and were designed not to have an influence on interest rates generally, other than through the amount of reserves in the system.

Borio: Now, there are different-- You could have reserve requirements, or not reserve requirements, but the basic feature of reserve requirements was that, they had averaging provisions, so that banks would have to fulfill them only on average over the so-called maintenance period. That basically said that they were largely indifferent about the amount of reserves they held during the maintenance period. That would allow more flexibility in the management of the amount of reserves, without that having much of an impact on the volatility of the overnight rate. 

Beckworth: Yes, and to be clear, most central banks prior to 2008 operated under this system, correct?

Borio: I know of only one or two that may not have operated that system, but, yes, all of them. Let's say that all of them did, for simplicity.

Beckworth: Yes. Again, just to summarize, in this system, you would have a policy rate between the deposit rate at the bottom and the lending rate at the top. In the case of the US, just to make this concrete, in the case of the US, which is a little bit different, I think they've called it an asymmetric corridor system, pre-2008, but the upper rate would be the discount window lending from the central bank, and the bottom would have been 0%, because that's effectively what you would get on your reserves if you did park it at the Fed. But, today, we have explicit rates.

Borio: That's why, actually, I don't like the term "corridor" so much, precisely because if, for example, you look at the paper that I wrote in 1997, which was an overview, a very detailed overview of implementation procedures across central banks around the world, many of them didn't quite have a corridor, or an obvious corridor. And in many cases, the lending facility at the top was never binding, because the management of the reserves was such that they wouldn't really need to go and borrow. The system, as a whole, was never short, so that they wouldn't have to borrow or go into the central bank. So, I think that the key word here is "scarce".

Borio: It's the idea that you keep the market relatively tight when it comes to the demand for reserves, so that the only demand for reserves that matters is the demand for settlement purposes, where I would also include, in this context, any demand that may be the result of reserve requirements with averaging provisions. If you had reserve requirements without averaging provisions, then they wouldn't play any role whatsoever in the implementation policy. It would just be a pure tax.

Beckworth: So, two more points on this system. One, as you mentioned, banks are lending to each other overnight, and they're typically— it's uncollateralized lending, right? They're just lending to each other, and they're checking each other out?

Borio: It could be collateralized, it could be uncollateralized. In the older days, it tended to be uncollateralized, for a number of reasons. Following the GFC and the like, there was more of a shift towards collateralized lending.

Beckworth: But this was useful, because, as you note in your paper, it led to price discovery, interbank markets, overnight interbank lending, which is a market that doesn't exist really anymore between banks. So, that was an important one, price discovery. Another, I think, big insight from this is that, under this system, the central bank balance sheet was linked closely to the stance of monetary policy. They moved in unison, whereas the next system we're going to look at, the abundant one, they're de-linked. In fact, people will argue that this is an advantage of the abundant reserve system, that you have an extra degree of freedom versus the corridor system. But, I actually think it's an advantage to have them linked. Any thoughts on that observation?

Borio: Oh, I have many, so I don't know. 

Beckworth: Well, shoot away, yes.

Borio: No, the first point, and I think it's extremely important, and unfortunately, it gets lost in the debate these days, is that, one should make a very clear distinction between the size and composition of the asset side of the central bank balance sheet, and the composition of the liabilities. Put differently, the size and composition of the assets of the central bank, well, that will generally depend on things like, I don't know, do I want to have some kind of influence on longer-term interest rates? Do I need reserves? It's not so much the case in the US for FX reserves purposes, this is foreign currency, the foreign currency component.

Borio: Now, [on] the liabilities side, if you take out reserve requirements— and actually a number of central banks are operating without reserve requirements. In that case, the amount of reserves in the system was tiny, because for settlement purposes, given that you can always trade at the end of the day with other banks, you didn't really need much. And moreover, you could count on the central bank to make sure that the amount of reserves in the system, as a whole, was roughly right. All you needed to do was transact with other banks, and for that, you didn't need to have any additional, much precautionary reserves, and so, the amount was very tiny. Look [at], for example, what was the case in Australia, and it still is actually there, or in Canada, and so on. That's minute, it doesn't count. Then, of course, you have cash, and cash [that] central banks meet is an autonomous factor.

Borio: They meet on demand, okay? And that, of course, was constraining the size of the balance sheet. But what I'm basically saying, is that you can have a very large balance sheet and still operate a scarce reserve system. This is actually what happens in many countries around the world today, mainly emerging market economies, where what they have is— the reason for the large size of the balance sheet is not large-scale asset purchases of domestic assets, domestic currency assets, during the post-GFC period, to have an even more accommodative monetary policy stance, but rather foreign exchange reserves as a war chest against possible risks, and also as an amount that they can use strategically or tactically for foreign exchange intervention.

Borio: But, what I'm saying is that, those central banks have very large balance sheets, and yet they're operating a scarce reserve system. All you need for that are instruments that are there to withdraw reserves from the system, and there are many different ways of doing that. You can issue your own paper, which is the case of the Swiss National Bank, or, say, the RBI. You can use FX swaps, you can use repos of various kinds, and so on and so forth. But, so, I think it's very important to distinguish size and composition of the assets, versus which system for implementation that you want. Is it a scarce reserve system, or is it, on the other side, we haven't yet talked about it, but an abundant reserve system?

Beckworth: So, did you hear the argument made for the abundant reserve system, that you had this extra degree of freedom, kind of ignoring all of these things you've just said?

Borio: Sometimes that argument has been put forward, but once you recognize that you can still withdraw reserves by using, say, your own paper, and so on, it doesn't quite-- You can de-link the two decisions.

Beckworth: Okay, well, let's jump into the abundant reserve, or ample reserve system, and talk us through that. How is that distinct from the scarce reserve system?

The Abundant Reserve System vs. the Scarce Reserve System

Borio: Remember that I mentioned that the amount— the demand for reserve balances, for settlement purposes, which is the core, and the only one that really matters in the scarce reserve system, is very interest inelastic. In fact, think of it as vertical, effectively. So, what you do in a— what I would call an abundant reserve system, or you can call it ample reserve system, I know that in the United States, people are making now a distinction between the two. But, the key point here is that, you basically increase the amount of reserves to the point that banks hold more than they need for settlement purposes, which means that the interest rate will either go to— if you don't have a deposit facility, it would go to zero, or it would go to the deposit facility.

Borio: Then, depending— again, here are complications, it could actually go below that, if there's some kind of opportunity cost to holding these reserves, or if there are different participants, like in the United States, that have access to the central bank with— in different terms of access to a central bank balance sheet. Forgetting about those complications, the key point is that the demand for reserves becomes, also, a store of value demand.

Borio: So, banks no longer hold reserves only for settlement purposes, or for the reserve requirements and so on, they also hold them because they're an attractive asset to hold relative to other assets in the system. So, if you think of, a little bit, I'm exaggerating, but for simplicity, think of a demand for reserves for settlement purposes, where the interest rate doesn't come in, and a demand for reserves in the abundant ample reserve system, which is also for a store of value, which therefore, depends on relative returns and risks associated with those returns.

Borio: So, we effectively move from the previous world, to what one would call, or is the Tobin-Brenner world, which is actually how economists incorrectly thought monetary policy was set in the old days. Whereby, we are basically increasing the amount of reserves in the system, [and] you are moving down a demand curve until somehow you are basically hitting the right spot that you wanted to hit. But, the point here is, you basically go from a positive opportunity cost to a zero, or even negative opportunity cost of holding reserves.

Beckworth: Okay, so the speech that you give talks about the sea change that happened between these two systems. We've just described them. So, what's the history here? Why did we have this sea change? Why go there? Is it just because of large-scale asset purchases? Did regulations after the great financial crisis also play a role? Why are we still there more than a decade later?

Borio: The only reason was, effectively, because of the large-scale asset purchases, and this was relieving pressure on other instruments that you would need, in order to absorb the liquidity from the system. And that was seen as a plus, so simplifying the operations, and that's the reason why we stayed there. Now, of course, people are, as you know, as you well know, they are trying to assess the pros and cons of the two systems. The reality is that, it was basically the great financial crisis that caused a number of advanced economy central banks to shift to this system. Once you're there, you say, "Well, why don't we stay there?" And, indeed, there are arguments put forward to suggest that it's a useful system, and there are also arguments to suggest that, and that's where the regulation comes in, given the changes in regulation that have taken place, that it would be harder to operate the previous system.

Beckworth: Okay, let's go through four arguments for the abundant reserve system that you outline in your paper, and you give responses to each of those. I'll briefly summarize them, and we'll go back and look at them individually. The first argument for it is, the floor system does its job, it works. The second one is, this is the new normal, it's hard to go back. We just touched on that one. The third one is, there's a de-linking between the balance sheet and policy rate, this extra degree of freedom argument. And the fourth is, there's no rush. We have a big balance sheet, why not wait to deal with it later? Let's go to that first one, the floor system does its job.

Arguments for the Abundant Reserve System

Borio: If by doing its job, you define it as allowing the central bank to have a reasonable control over operating targets, it does., I would say that, on average, what you would have had in the scarce reserve system, that depends on the system, and so on, you may have seen more volatility of the overnight rate. Again, that was partly endogenously dependent on how much volatility the central bank was prepared to tolerate. You could have more volatility in the overnight rate. By that, I mean the operating target of the central bank. What you have on the other hand, in the abundant reserve system, very often, is that you have longer, more protracted, persistent deviations of it, but very small, typically, deviations of the overnight rate from… or your operating target, from the deposit facility rate.

Borio: But now, with this twist that anyone who was thinking of these systems in the past would find mind-boggling, that the deposit facility rate becomes the ceiling, as opposed to being the floor, to the overnight rate, partly because of the leakages that we know exist in some of these systems. So, yes, it does its job. Is it easier to operate it? That's the other question. Well, you don't have to forecast autonomous factors, that definitely makes it easier.

Borio: But, on the other hand, as you've seen in the United States, the reason why the reverse repo facility was introduced was to set a floor… finally, a floor to the overnight rate. So, now, you're operating two different rates, two in the system, and you have to keep track of what happens to the amounts in the reverse repo facility and so on. Yes, you don't have to forecast autonomous factors, but even that wasn't that difficult in the previous system.

Borio: So, I think [those] arguments about it being simpler, or more complicated, are not really decisive in that sense. Plus, there is the question… but people actually argue that you could not operate a scarce reserve system these days, because the environment has changed so much that, for example, the demand for reserves has become much more unpredictable. And a key point of the speech that you mentioned is to argue that the unpredictability is, itself, a function of the system itself, because what I mentioned earlier is really a critical point that people have not really focused so much on, which is that, you're shifting from a demand for reserves, which is purely for settlement purposes, plus reserve requirements, to add a demand for reserves, which is also for a store of value.

Borio: That means, for example, that the amount of reserves that you hold to meet, say, the Basel III liquidity regulation, which is actually silent as to whether you should hold it, in terms of bank reserves, or short-term government paper, or government paper in general, that means that you would try and hold many more reserve balances, as opposed to government paper, because there is no opportunity cost of doing so.

Borio: Therefore, that will fluctuate, the demand for reserve balances will fluctuate with anything that affects, exogenously, the relative yield of government paper relative to bank reserves for whatever reason. But that would not be a factor if you actually went to a scarce reserve system, because the opportunity cost of holding reserves would be sufficiently high, but the only thing that would matter would be the demand for settlement purposes.

Borio: This unpredictability is largely endogenous. Another reason why people say it's unpredictable is because they say, "Well, the interbank market is not functioning," but the interbank market is not functioning because banks don't need to borrow and lend to each other. Another reason could be… and say, “well, some autonomous factors are now more important than in the past.” Take, in the United States, the fact that the government—What is it called in the US?

Beckworth: The Treasury General Account? 

Borio: Exactly. So, it is the amount of the balances that the Treasury holds with the central bank, and can be quite volatile. That's true, but one reason why it's holding them is because, after you had shifted to the other system, the Treasury was getting too little returns on its reserves, so it preferred to put its money back with a central bank. Before, it was holding it with the banks. When it was holding it with the banks, then that was not a problem, in terms of forecasting the autonomous factors, as it has now become. So, you see that many of these things that people take as exogenous, with respect to the system, are, in fact… could be thought of, at least to a considerable extent, the result of the system itself.

Borio: If you were to go back to the previous system, then the predictability of many of these factors would return to what it was before. The difference is, indeed, that now you've got regulatory prudential requirements that require banks to hold more reserves, and that's an additional settlement reason for holding reserves. But that, at the end of the day, is going to be largely interest rate inelastic— kind of lexicographic ordering— banks will hold that, and then will think about other holdings. And so, even that can be that unpredictable. I am pretty sure that you can work out… it would be like a rule of thumb. You could even ask banks, "How much do you want to hold reserves for these liquidity requirements?" And then, given that, you could factor that into your forecast of the demand for reserves, which must be the way in which all of those banks around the world that have huge balance sheets, and are still operating these systems— mainly emerging market economies— are doing things.

Beckworth: Great points, and Bill Nelson, who's at BPI, used to be at the Federal Reserve, likes to point out that the number of people employed at the Federal Reserve, looking at the balance sheet, has actually grown dramatically, versus what it was before. Actually, they could get by with fewer people forecasting the balance sheet pre-2008. And just to illustrate your point about how difficult is it, the prior system versus this system, I think it's quite a challenge in the current system to predict how much is just enough reserves before we fall back into a corridor system. So, 2019, right? The Fed was trying to manage its ample reserve system. It was doing QT, tax revenues were flowing in, and oopsie, we fell back into the other system that we were trying to avoid. That itself is a challenge, so, I like that point. I like the fact, what you said, that a lot of the stuff is endogenous to the system itself. You change the system, you change some of these issues.

Beckworth: Let's go to the second point, which we've touched on already, but that is, this is the new normal. It's hard to go back. I think a lot of this is tied into the regulatory issues as well, which you just touched on. But, if we did go back, do you think it's possible to talk to bank supervisors, to talk to bankers themselves and say, "Hey, you don't have to hold all of these reserves."? Bill Nelson, again, he makes a great list. He shows how the amount of minimum reserves that the Fed estimates has grown over the years. It's gotten bigger, and it's bigger, and he attributes some of that to this role that regulation is playing. Could we reverse that if we went the other direction?

Borio: Well, how much banks would need to hold for regulatory purposes, that's a question that, at the end of the day, one should ask supervisors. And I know that, across countries, supervisors have different practices, and that, to go back to what I mentioned earlier, if you just focus on the international banking regulations, the Basel regulations, Basel III, in the liquidity requirement, bank reserves and government paper are treated as perfect substitutes.

Borio: Now, that is not the case in other country-specific, jurisdiction-specific requirements, or supervisory practices, as for example, in the United States. But, the point I'm basically making is that, for example, when one is asking banks, "How many reserves would you like to hold today, or do you need to hold," blah, blah, blah, that’s not a fully well-posed question, because the question should be, if you want to know how much they would need to hold under a scarce reserve system, you should ask that question. How many reserves would you need to hold if the overnight rate was at, say, I don't know, 5%, and the deposit facility rate was at, I don't know, 4%, 3%, whatever, how many reserves would you hold then? That's a properly asked question. 

Borio: But if you basically ask banks now, unconditional, "How many reserves would you like to hold?" Given that, for the reasons I mentioned earlier, their opportunity cost of holding reserves is very low, or zero, or if anything, they would basically say, "Well, I would like to hold lots of them." So, this is endogeneity of the demand for reserves with respect to the system. And I think the key point, the difference between demand for settlement balances, including reserve requirements with other provisions, and any supervisory requirements, that settlement balances demand, and the demand for, as a store of value, is extremely important. In one system, it's just settlement balances. In the other system, it's also, and at the margin, a store of value. So, the features of the demand are completely different.

Beckworth: Okay, the third argument given for an abundant reserve system is a financial stability claim, that you can use the balance sheet… Again, this to me is where the de-linking between the stance of policy and the balance sheet size comes into play, that argument comes into play. “Well, we can inject liquidity if it's needed.” What's your response to that?

Borio: Well, here, the question is whether-- It goes back to the first issue that we discussed. Would you like to have a system in which the central bank is a backstop, or would you like to have a system in which the central bank is the mass market maker of first resort, so last resort versus first resort? At the end of the day, in an abundant reserve system, you flood the system with the reserves, which means that banks don't need to go to the interbank market any longer.

Borio: There is no pooling, effectively, of reserves any longer, and a result of that, they're "self-sufficient" up to the point. Obviously, it will never be self-sufficient if they have a run, that's for sure. So, what do you want to have? Do you want to have a system in which the central bank provides liquidity as a backstop, or do you want to have a system in which everything is injected into the system already, beforehand, and you're relying on banks being self-sufficient?

Borio: Basically, what we said before, you're killing the interbank market. Personally, but I can see that people can have very different views here, I think that having a system in which the central bank is a backstop, and a system in which the first line of defense against demands on liquidity is an interbank market, that to me sounds [like], on balance, a better system.

Beckworth: The last argument for the abundant reserve system is that there's no rush, we have large balance sheets, and many central banks. Let's wind it down first, and then we can have this conversation. Why should we be thinking about it now?

Borio: As I said earlier, I think that it's useful to distinguish a policy decision with respect to the size and composition of the asset side of the central bank balance sheet, and distinguish that question, which is a policy question, from the separate policy question, which is whether you want to have a scarce reserve versus an abundant reserve system. One of the key issues is, what happens to the interbank market? And, whether we want to provide an asset to the system which is only available to banks, which is bank reserves, versus an asset which, in principle, could be available to anyone, which would be, for example, if you were to withdraw liquidity from using some other central bank security, a reverse repo, or an FX swap, or whatever.

Beckworth: How would we make a transition back to a scarce reserve system? Let's say that we all agree that that's the target. What steps would you recommend for the ECB, for the Fed, these advanced economies with large central banks and large central bank balance sheets?

Transitioning Back to a Scarce Reserve System

Borio: In the speech that you saw, because of this discontinuity in the demand for reserves, from being just a settlement medium to also store of value, and the difficulty in predicting exactly how, as you're approaching this boundary, the overnight rate is likely to behave as a result of possible shifts in the demand for reserves, my preference would be for what I call a big bang solution. By big bang, I mean that there is a discontinuity. You would like the central bank to be fully in control of the shift, both of the timing and the modalities of the shift. 

Borio: So, the best way of doing that is, basically, to say, "Okay, on day X, I'm going to shift from system X to system Y." And at that point, you basically absorb the reserves in the system. Now, you don't exactly know where this, call it the scarcity point, is going to be. But, because you don't know it, then you put some safety belt, if you like, which is basically a corridor, a narrow corridor, so that either banks can deposit any excess with you at a given rate, or they can borrow from you. But then, you have to have an effective way of banks borrowing from you, no stigma. That goes back to-

Beckworth: Another issue.

Borio: -another issue. And so, you basically… so you start with a narrow corridor. Then, over time, you basically decide how far to increase the size of the corridor to a point where you feel that you have the right amount of interbank activity in the system that provides you with information about the underlying riskiness of the banks, and so on and so forth. And that will allow… if, once stress emerges, because there might be episodes of stress in the system where banks would like to have, in aggregate, more reserves, that would allow you to provide less of an increase in reserves, because at least part of that is going to be dealt with through redistribution within the system.

Borio: This is one of the concerns that I have is that, if the muscle atrophies, then, when the central bank needs to provide liquidity, because there are going to be… there may be tensions in the system, it will have to start from a higher level, and it will have to provide more, because there will be no redistribution of liquidity within the system. Now, of course, under stress, redistribution is never perfect by any means, that's why the central bank has to come in and provide liquidity to those that need it.

Borio: But, one thing is to say that redistribution is imperfect, or indeed even highly imperfect. Another thing is to say that, "Well, I'm starting from a system in which redistribution is going to be zero, in which case I will have to do all the work." Then, there is the usual question of how you exit. And, by the way, the Central Bank of Norway was one of the central banks that was actually having an abundant reserve system even before the crisis, and so on. But, over time, they've decided to go back closer to a scarce reserve system, precisely because they saw that the amount of reserve demand was growing over time. And this is a kind of hysteresis that you could have in the system.

Beckworth: I'm glad you mentioned the Central Bank of Norway, because that central bank along with, I believe, the Swiss National Bank, and then the South African Reserve Bank— and I got this out of Daniel Hinge's article that I mentioned earlier from the Central Banking publication— he mentioned that all of those central banks have gone to a tiered reserve system, which I think you're alluding to that. And that is where only some of the reserves are remunerated at the central bank and others are not, or they're at much lower rates. So, it creates the incentive, very similar to a corridor system or scarce reserve system, that you go to other banks and ask them for reserves. So, as a transition, what do you think about going to a tiered reserve system?

The Possibility of a Tiered Reserve System

Borio: I think that's a possible transition. And you could think of what the Swiss National Bank has done so far. It might stop where it is, but of course, it could move further. But, the key thing that they did was precisely to develop an instrument, central bank bills, to absorb liquidity in the system. Going back to what I said earlier, the distinction between the size of the balance sheet and the composition of your liabilities. I mean, the FX reserves that the Swiss National Bank has, and therefore, the size of its balance sheet is, I think, in excess of 100% of GDP.

Beckworth: Wow. 

Borio: Those are the numbers that we're talking about. And yet, it now has a tiered system, whereby, it remunerates up to a point, and after that point, [which] is basically zero, it tries to keep the amount of reserves close to that point, so there is some activity in the interbank market, which is extremely important for them in order to have a benchmark rate that actually has transactions. You remember the LIBOR scandal, and we have moved away from that. So, that's a possibility. My preference would be, once you are there, you can either go that way, or you can basically go the way that I've described.

Beckworth: Go all in, yes. 

Borio: Yes. I don't think that at the end of the day-- I can see why they did it, but I can also see that it wouldn't be difficult, basically, to skip that step altogether.

Beckworth: The next question I have for you is, how was this received, your speech, and these suggestions? Because I've talked to some central bankers on this podcast, one regional Federal Reserve President. In fact, this was during the period where the Fed was officially stating that we're going to be an abundant reserve system. I said, "Why are you doing this? Why make this official?" And he basically said, "It works. It's good enough." For me, that was a complacent, not very thoughtful answer, because you want to think through, what's theoretically best? What's the benchmark? And then, go from there, not, "It works well enough." What is your sense? Are central bankers complacent? “It works well enough?” Or, are they reconsidering these arguments?

Have Central Bankers Become Complacent?

Borio: As I said, if you look around the world, there are different views about the pros and cons of these systems. I would say that, from an economic perspective, probably the most fundamental question in this context is, what are the merits that one sees, or doesn't see, in having an active interbank market? If you think that market—and I know that some people believe that, if you think that market… “well, in any case, what is that market for? It's just the redistributing ‘exogenous shocks’ to payments within the system. Why do we need it? What's the economic function of that?” Now, if you take that view, then I think that some of the issues that we discussed before, they do not appear to be important. My personal perspective is, that interbank market performs a useful function. 

Borio: It is the first line of defense to deal with liquidity shocks. It provides useful information to supervisors. It provides useful information to market participants themselves. It reduces the need for central banks to provide liquidity, actually, sometimes, when real stress emerges. And so, for all of these reasons, I would prefer to have a scarce reserve system. I obviously understand those who, on the other hand, basically think that this is not an important issue. And, as a result, they say, "Well, why don't we stick to the system that we have, which doesn't require us to forecast the demand for reserves, as the previous system does." What I think it's important to distinguish, though— and let me go back on this, because it's a key point— we should decouple the question, and I repeat it, the size and composition of the asset side of the balance sheet, with respect to the system that you want to have. I think that's important.

Borio: My concern is that there is a little bit of a risk that the tail wags the dog, because, oh, well, banks want to have lots of reserves. Remember, that's partly a function of it also being a store of value, so we cannot cap the asset side of the balance sheet. No, I think that, actually, you should go the other way around. You should think, “what is the appropriate size and composition of the balance sheet of the central bank?” And that also has to do with issues that [are] related to losses, central bank losses, and so on and so forth. What is, more fundamentally, the role that the central bank should play in an economy? Should it have a minimal footprint, or should it have a larger footprint, and if so, why? You decide that, that's one thing, and then you decide which of the two systems you would like to have.

Beckworth: Okay, the one thing we haven't touched on that you bring up in your article is the political economy, or the politics of large central bank balance sheets. And I mentioned earlier that the ECB is looking at alternatives to its current operating system. One reason they're doing this is, there's political pressure, because it appears, at least, that the ECB is subsidizing banks, these large interest payments going to banks. And that has put some pressure on them, and that could put pressure in other places. In the US, there have been people who have called for the Fed to use its balance sheet to fund certain projects. So, this all brings into question, then, the central bank’s independence. We want to think about what the large balance sheet means for central bank independence and its ability to do its job. Is that argument gaining more momentum, more appreciation among central bankers?

The Politics of Large Central Bank Balance Sheets

Borio: Let me unpack that question, and again, let me go back to the point that I was making before. Unless you have the tail wagging the dog, you should decide on the size and composition of your balance sheet [based] on reasons that have nothing to do, or very little to do, with the scarce versus abundant reserve system. And there are issues, in terms of the risks that the central bank takes, the losses that the central bank takes, the possible losses that the central bank exposes itself to, and the political economy consequences of that.

Borio: Now, we know [that] those losses do have implications for the fiscal policy, for the fiscal position of the government, because these are revenues that the government is not getting. But from the point of view of the operation of the central bank itself, central banks can operate, also, if they are making losses. The problem is that making losses focuses so much attention on the central bank, that it may prevent the central bank from doing the right things when it needs to do those things. So, that's one point.

Borio: When it comes to the distinction between reserves and other liabilities… so take the size of the balance sheet as given, take the composition of the asset side of the balance sheet as given. So, If you were to withdraw the reserves, rather than leaving them there, in order to move to a scarce reserve system-- Now, the losses, in terms of the central bank balance sheet, wouldn't be that different.

Borio: What would be different is that, rather than having exposures that are indexed to the overnight rate, they would be indexed to whatever rate… the maturity of the instrument that you're using, in order to withdraw reserves. Now, the political economy of the losses is going to be somewhat different, because in one case, bank reserves can only be held by the banks. If you're increasing the interest rate on bank reserves, you're just "transferring money" to the banks. If, on the other hand, you have the same size of the balance sheet, you have the same interest rate, but you are withdrawing reserves through an instrument which is available to everyone, like a short-term government paper, then you’re not perceived to be subsidizing the banks.

Borio: Now, the other advantage of that is— which is not a political economy argument, it's a pure economic argument— is that in some cases, if you have very large bank reserves there, deposits with the central bank, because those reserves are only available to the banks, then you can induce a collateral scarcity. And that was an issue, in particular, within the Euro area. Large-scale asset purchases combined with central bank reserves—very large bank reserves— equals collateral scarcity, because no one can use bank reserves except for the banks. Had you actually withdrawn the amount through short-term government paper, that could have acted as collateral, short term central bank paper.

Beckworth: So, you've actually written on this interaction between fiscal policy and financial stability, and as it plays into central bank balance sheets. So, in the time we have left, I just want to, maybe, park there for a few minutes, and talk about the implications of the Fed's balance sheet for fiscal policy. One question I have is, does it really matter that the Fed, or other central banks with large balance sheets, become a more important player in public debt management? So, the Fed now manages, roughly, a quarter of the outstanding US government liabilities, it changes the maturity structure of it. Is that something we should be thinking about, or is that more of a political consideration?

The Fiscal Policy Implications of the Fed’s Balance Sheet

Borio: Again, there are two different issues. One is the impact of large central bank balance sheets invested in government paper for the fiscal position of the government itself, and in particular, for the vulnerability or the sensitivity of that position vis-à-vis the level of interest rates. Something that we argued extensively— also in the annual economic report a few years ago, but we have repeated it a number of times— is that you had this illusion, whereby, people actually thought that governments were lengthening the maturity of their liabilities, particularly after the GFC, to take advantage of the various negative or small term premium.

Borio: As a result of that, they were also reducing the sensitivity of fiscal positions to higher interest rates, because it was all fixed term, long term. That's what you saw on the pure government portion of the balance sheet, on the liabilities side. Now, what people were not focusing on, which is really what you've just said, is that a lot of that government paper was bought by the central banks in some jurisdictions. And those central banks had issued, to finance that government paper, bank reserves. And those reserves are effectively indexed at the overnight rate. So, if you take the consolidated public sector balance sheet, central bank plus the government, these large scale asset purchases amounted to a very large scale open market debt management operation, whereby, the state was withdrawing, actually, its long term paper in the market, and replacing it with overnight paper, which is effectively bank reserves.

Borio: So, that was not reducing, that was increasing the sensitivity of the government fiscal position to higher interest rates. That was not showing up as higher borrowing costs for the government. It was going to show up as lower revenues for the government, because the central bank would not remit profits to the government. So, in fact, we did some back of the envelope calculations, and, at some point, something like 30% to 50% of government debt, at least long-term government debt, was effectively overnight as a result of these operations… 30%, 45%, 50%, depends on exactly how you count it.

Beckworth: That's sizable, yes.

Borio: That is sizable, and it's something that we wanted to draw attention to. So, the large scale asset purchases, they have fiscal implications, in terms of the sensitivity of the government’s fiscal position, to higher interest rates. They increase that sensitivity. Then, there is the question of what implications they have for the operation of a central bank, and that’s largely within the typical range that we're talking about. That's largely a political economy.

Beckworth: Well, with that, our time is up. Our guest today has been Claudio Borio. His speech is titled, *Getting Up From the Floor.* We'll have it linked to in our show notes. Claudio, thank you so much for coming back on the program.

Borio: Thank you. Thank you for having me again.

About Macro Musings

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