Bill Nelson on the Future of Central Bank Operating Systems

Bill Nelson is a Chief economist and an executive vice president at the Bank Policy Institute. Bill returns to the show to discuss the changes at many central banks around the world from a supply-driven floor system to a demand-driven floor system and how the Fed has been resistant to this change. 

Subscribe to David's new Substack: Macroeconomic Policy Nexus

Check out our new AI chatbot: the Macro Musebot!

Read the full episode transcript:

This episode was recorded on March 6th, 2025

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: Welcome to Macro Musings, where each week we pull back the curtain and take a closer look at the most important macroeconomic issues of the past, present, and future. I am your host, David Beckworth, a senior research fellow with the Mercatus Center at George Mason University, and I’m glad you decided to join us.

Our guest today is Bill Nelson. Bill is a chief economist and an executive vice president at the Bank Policy Institute. Bill previously was a deputy director of the Division of Monetary Affairs at the Federal Reserve Board, where his responsibilities included monetary policy analysis, discount window policy analysis, and financial institution supervision. Bill also worked closely with the BIS working group in the design of liquidity regulations.

Bill is also a frequent guest of the show, so we encourage you to gobackandcheckouthispreviousprograms, if you haven’t already. Today, Bill joins us to discuss an emerging trend among advanced economy central banks. The movement away from the so-called supply-driven floor system to a demand-driven floor system, and what this means for the future of central bank operating systems. Bill, welcome back to the program.

Bill Nelson: Thanks, David. It’s always a pleasure to be here.

Beckworth: It’s always great to have you here. This is one of my hobby horses, the operating systems. The other one is, of course, the targeting of the Fed, what it does, but this is always important and an area that doesn’t get enough conversation, so it’s always a pleasure to have you on the program to talk about it.

Nelson: Thanks.

Trend Toward a Demand-Driven Ceiling System

Beckworth: Now, this broad trend that’s emerging is one that I don’t think gets enough attention, and I know you’ve been looking at it closely. In fact, we’re going to discuss an article that you’ve written about the Bank of England, some of the changes it’s made there. Just briefly, what I’m aware of, the Bank of England, the ECB, the Bank of Canada, Reserve Bank of Australia, the Riksbank, all of these banks are making changes away from this large balance sheet, lots of reserves, and moving more toward what they would call demand-driven or maybe more market-driven, and I know that’s not the perfect language. Maybe start us off and give us the broad contours of this development.

Nelson: Sure. Happy to. As we’ve often discussed, I always think it’s important to start out with the balance sheet of the central bank and thinking about what they’re doing. Broadly, the central bank’s assets are securities and loans, and their liabilities are effectively currency and reserve balances, where reserve balances are the deposits of banks at the central bank. The way the central banks conduct monetary policy is they adjust their holdings, their assets, to move around reserve balances. Currency doesn’t change. Then they also change the interest rate that they charge on their loans and the amount that they pay on their deposits on the reserve balances.

All of those different steps are taken to move interest rates to where they want them to be to manage the macroeconomy. This is what you and I enjoy, the plumbing, as it were. After the Global Financial Crisis, and particularly during COVID, central banks expanded their balance sheets massively, by either making emergency loans or longer-term loans or buying securities. As a consequence, that pushed the amount of reserve balances up to atmospheric levels.

So far up that interest rates that any institution would be willing to pay for each other in the money market, all got squished down to, or below, the amount that the central bank pays on reserves. Because why lend to somebody if you can just leave the funds on deposit at that interest rate? That was generally referred to as operating a floor system. In the new vernacular, it’s also called operating a supply-driven system. The quantity of reserves that are out there are driven by the choices of the central bank about how big to make its assets.

Now, what’s changed is that many central banks are now describing themselves as moving to a demand-driven system, or even they will say a demand-driven floor system, although I think that’s a misnomer, because the terms just get a little complicated. What they’re doing is they’re shrinking the size of their emergency loans, or easier to think about, their securities, like in the US is doing, through QT, letting the securities run off. Or for many of these central banks, they’re also actively selling the securities, to shrink, reduce that quantity of reserve balances, until it actually gets below the level that the banking system needs and wants normally.

So that banks have to come in and borrow. I said the other asset of the central bank is borrowing, lending. When banks borrow from them, at discount window loans, that raises the necessary size of the balance sheet to the level so that the reserves that are created are what are demanded by the banks. This is historically what would be actually called a ceiling system. Instead of being at the floor of your corridor, created by your interest rates, you’re effectively going to be in equilibrium at the ceiling, where people are borrowing. The ceiling of your corridor.

All of these central banks are actually setting their corridor, the difference between the lending rate and the deposit rate, at very low or even zero levels, with the consequence that it looks a lot like a floor system in the sense that you would expect market rates to be near the amount that the central bank is paying. It’s really being nailed down by that lending, rather than by the deposit taking.

Beckworth: Would a better title be, demand-driven ceiling system?

Nelson: It would, actually.

Beckworth: Okay.

Nelson: I think that the other central banks might be a little bit concerned about using that term, because they don’t want to piss off the Fed, and the Fed is a big believer in the floor system.

Beckworth: Sure.

Nelson: I think everyone feels burned by having had such a big balance sheet.

Beckworth: Yes. I’ve also heard, liability-driven, demand-driven. The key idea is now that the quantity of reserves will not be determined by QE or what the central bank is doing, but rather by banks going to maybe a repo facility or the discount window and borrowing from the central bank. That’s the key difference.

Nelson: Yes. One thing that this highlights is that even though in the United States, borrowing from the central bank is often viewed as a sign of distress or something that should be rare, most central banks, that’s a perfectly normal asset that they maintain. The Bank of England, for example, intends for lending to banks to be their majority asset, and it’s historically been the main asset of the ECB as well. People do things in different ways in other countries.

Beckworth: Right. We need to be humble about that. Not every system is like the US, and the ECB is a huge central bank, has its own ways of doing things. We had Isabel Schnabel on last year, I believe, and that was a fascinating conversation. She has really spearheaded the move, the discussions. In fact, I think they spent almost a year just thinking about this question. What the Fed does now for the framework review, they do that for the operating framework review on their side. In fact, that leads me to this next observation. I mentioned the Bank of England, the ECB, the Bank of Canada, the Reserve Bank of Australia, the Riksbank.

Will the Fed Follow Suit?

What I did not mention is the Fed. All these banks are inching away or moving toward this demand-driven ceiling system, if you want to call it that. Are there any prospects of the Fed one day getting on this bandwagon and joining them? I guess this goes back to, what’s the motivations for all these central banks doing it, and does this motivation affect the Fed at all?

Nelson: It’s a little hard to say, and there might be a little hopeful thinking on my part, in my view on that. The Fed is also shrinking. It’s, as is broadly known, engaged in QT by allowing its securities to roll off as they mature. It’s in the process of that happening. It doesn’t know when it’s going to stop, because it doesn’t know how low that level can get without meeting some resistance. They used to talk about this as, “Well, we’re going to stop when it’s about $200 to $300 billion above where we need to stop, so that we don’t actually need to conduct monetary policy on a day-to-day basis. That’s such a large extra quantity of reserves that normal shocks to those things will just move things around, and we won’t need to do anything.” 

Right now, my sense and maybe hopeful belief is that they’ve just decided that they don’t really need to make a decision about this until they see some indications that they’ve gotten down to the level where they’re going to meet some resistance. They don’t really seem to be talking about it. As we might discuss, President Logan from the Federal Reserve Bank of Dallas recently gave a speech indicating continued full-throated support of a floor system, but even now, the Fed doesn’t really talk so much about maintaining such a large buffer.

Beckworth: Yes.

Nelson: I think that everyone wants to be smaller. This idea that it’s cost us to be gargantuan. The Fed is now 23% of GDP. It used to always be 5%. I think that this issue of, do they really need to be that big? I think everyone’s asking that question, including within the Fed.

Beckworth: Yes. I wonder also, if they’re more mindful of this issue because of the political pressures they now face, too. A point that George Selgin has made on here many times, I think you’ve made this as well, is put aside all the technical pros and cons of floor versus corridor, at the end of the day, probably the most pressing issue could be political pressure to use the Fed’s balance sheet for political purposes, right?

Nelson: Right.

Beckworth: If you can tap into that balance sheet—and one of the advantages of a floor system, at least one of the claims that has been made, is you separate the stance of monetary policy from the size of the balance sheet. You’ve got like two independent levers, two instruments. If a politician, if a president becomes aware of that, they’ll go, “Hey, are you telling me I can play around with the size of the Fed’s balance sheet and not affect the stance of policy? That’s like a free lunch, right? I can build the wall. I can do any number of things.”

Before, some people on the left were also talking about paying down student debts. I’m just wondering now in this environment, if a Fed official is not worried about the possibility of political pressures facing them?

Nelson: It seems likely, to me. I guess one chairman, I don’t remember exactly which one—I’ve worked for four of them back in the day—indicated, as you just said, that operating under a corridor system, which we haven’t discussed yet, but that’s a system where the central bank just provides the amount of reserve balances that the banking system needs, in an effort to maintain interest rates in between that corridor, between the two. If they’re operating policy that way, if they’re under pressure to buy more securities, they can correctly say, “Oh, sorry, I would, but I can’t, because if I did, I would lose control of monetary policy.”

That’s not true, really, so much in a floor system. You can always get a little bigger. Now, the ability of what the Fed can purchase is limited by law, to treasuries and agency securities, but laws can be changed. There is a concern that this just opens up this idea that that balance sheet is there to be used for all kinds of purposes. When the Green New Deal was published, they published an article saying their financing plank was the Fed will print money.

There was a nominee for the OCC who wrote a book saying, “Yes, the Fed should just give everyone an account and put money in the accounts of the worthy and of good businesses that do proper things.” It is, unfortunately, I think, a temptation. The siren song of the Fed’s unlimited balance sheet, which is a risk, and George has written a lot eloquently on that.

Beckworth: The siren song, I like that framing. It’s an apparent free lunch, but it’s not, because at the end of the day, there’s only so much fiscal space a government has, whether the Treasury uses it, the Fed uses it. It’s an illusion. It looks like there’s something just sitting there, easy to use, won’t affect inflation, won’t affect the stance, but we know otherwise. 

Bill, we’ve been talking about this broad trend toward a demand-driven ceiling system, if that’s the proper language, and we’ve talked about the Bank of England, the ECB, Bank of Canada, the Reserve Bank of Australia, Riksbank, all being a part of this. Something that you’ve looked at closely is the Bank of England’s movement in this direction, and specifically you have an article titled, “Will the Bank of England’s New Policy Implementation Framework Work?” You say we should look closely at all of these banks, but in this case the Bank of England, because there might be lessons learned for us. Should we one day make the move?

Bank of England’s New Policy Implementation Framework

Nelson: That’s right. I think that’s right. The Bank of England’s system, which they’re moving toward rapidly, they’ve announced and it’s on, is one where they’re letting their emergency loans roll off that they made during COVID. They’re allowing their securities to shrink. The result is that reserve balances are declining to such a level that they anticipate bank borrowing to pick up. The way they lend to banks under this program is they offer seven-day loans at bank rate, collateralized by gilt collateral by their government securities. 

Then they’re also offering weekly auctions of six-month loans. Those come in three types backed by different types of collateral. The spreads on those will be wider, thus liquid the collateral. The least liquid collateral is loans prepositioned at their discount window. It’s not called the discount window, but at the central bank. The most liquid is gilts, is government securities. That’s how they’re going to be providing assets. On their liability side, their deposit facility, where banks can leave funds, pays bank rate as well. You can see that they have a very narrow corridor, really, with bank rate on their weekly auctions of loans, but collateralized by super high-quality collateral, and then paying bank rate the same rate on their deposits.

That’s where they’re going. I indicated they’ve got three objectives. I think they’re very traditional, but also very good choice of objectives. One is to have good interest rate control. One is to make reserve balances or liquidity from the central bank cheap so that banks hold a lot of it for financial stability purposes. Then the last goal is to encourage interbank trading. There’s some tension between those goals, and I’m not sure where they’re going to be able to satisfy all three.

Beckworth: It’s a new trilemma of sorts.

Nelson: That’s right.

Beckworth: We have the macroeconomic trilemma, we’ve got the operating system trilemma here.

Nelson: That’s right.

Beckworth: You can’t do all three, maybe two of the three. That’s interesting. What should we be looking for then? As this progresses, what would be something you want to see that would tell you, okay, it’s working or it’s not working?

Nelson: One of the concerns that I express, and it’s probably my least concern, but nevertheless it’s a concern, is that the system could be relatively volatile. That’s because they don’t really articulate a way to deal with mismatches between the supply and demand for reserves that just can come up on any given day for reasons that were unanticipated or incorrectly forecast. There’s no real meaningful daylight lending tool in particular—not daylight, but for the day. If they come up short, there may be volatility.

On the other hand, they interact with the markets more, and they have late-day repos, so it’s probably something that they can manage. I think that they can move rates to where they want them to be. I do think that if their lending rate is close to or equal to their deposit rate, then banks will choose to hold a large quantity of reserves, because the cost of holding those reserves is very low or even zero. If you’re a bank and you’re making a decision, “Should I keep this money that I have? At the end of the day, I’ve got some extra money. Should I keep it on deposit at the bank or should I lend it into the interbank market?” If the rates in the interbank market are basically the same as the amount that you can get at the bank, then you don’t have a very strong incentive to make a little extra by lending that into the interbank market. 

That’s especially true right now, just because these large balance sheets approaches have killed the interbank market. There isn’t much of an interbank market. It has to be revived. For it to be revived, there has to be a clear profit motive that the banks need to perceive to overcome that and get the market going again. I worry that they’re not going to be offering enough daylight between what banks will expect to earn when they lend into the interbank market—in the US, that would be a sale of Fed funds sold, or a loan of Fed funds, or a repo—that’s going to be at a rate high enough to get the interbank market to recover. 

That’s so important, because as all of these central banks have recognized—Claudio Borio recognizes, anotherguest of yours, nicely in his discussions of these—it’s a better world, certainly one that I think the Mercatus [Center] would support, where if your first line of defense against illiquidity is the interbank market. The first line of defense is your own resources. Then banks have each other, that they can borrow and lend from, to deal with moderate to medium stress.

Where turning to the Fed is something that you would normally do, under these circumstances, only after you’ve used the interbank market or if the conditions in the interbank market are temporarily tight. We want to have the markets first and the government only as a backstop.

Beckworth: There needs to be a meaningful spread between the overnight market rate and the deposit rate at the central bank. It doesn’t have to be really big, but big enough to incentivize banks to economize, to shop around, to go for it. I guess that’s the impression I got. By the way, I use this new AI tool, Deep Research. I don’t know if you’ve heard about it. It’s basically like having your own RA, and it will go out and it will take 10 minutes. I had it research all these banks, and one of the common things that it came back with is, all of them are pointing toward or leaning toward creating some positive spread between the overnight market rate and the deposit rate.

In the US, it would be, say, the federal funds rate and then the interest on reserves. Some positive spread, again, not necessarily large, but large enough to make it matter. Then I guess in the case of the Bank of England, if there were volatility, then the ceiling would catch it, right? If things shot way up, you would have these repo facilities.

Nelson: They don’t really have a good daily—

Beckworth: That’s your concern.

Nelson: —end of the day. It’s one of my concerns, in the lack of an interbank market.

Beckworth: Okay.

Nelson: Volatility isn’t necessarily such a bad thing if it smooths out, but it’s usually not something that you aim for.

Beckworth: Yes. You think that the Bank of England is dynamic, it will adapt, it will recognize, “Oh, we need to tweak this, or maybe make these facilities a little more robust, a little more enticing.”

Nelson: Absolutely.

Beckworth: Yes. I think that’s kind of the point. If there is a narrative, stepping back, a bigger story to tell here is, these banks that we’ve listed, they’re inching toward something different. This, again, demand-driven ceiling system. Along the way, they’re going to discover some things work, some things don’t work. What you’re suggesting is, well, one thing that they probably will have to do is create some positive spread between the overnight market rate and the deposit rate, and they’re going to have to make sure they have robust facilities available in case the rates do spike. So the borrowing rate is at the very top.

They’re going to feel their way through this, and maybe we’re going to get back to something where there are smaller balance sheets. It may not be all the way to a corridor system, but somewhere in between. I’ve mentioned the Norges Bank. It went through this 2012, 2014 already? It’s been some time.

Nelson: Yes, it’s been quite a while.

Beckworth: Yes, they’ve gone through this, and they ended up at a tiered reserve system. Is that right?

Nelson: Yes, that’s right.

Beckworth: They didn’t go all the way, but they felt their way, and they were like, “Oh, I like this. It provides some incentive for banks to go to each other, and then also if they need to, they can come to us.” 

Okay, so the Bank of England is making strides, so is the ECB, and these other ones. Again, what’s not on this list of central banks I went through is the Fed, and I know you are eager for the Fed to be a part of this journey, this transition, and you said, “You know what? I’ve been making so many calls for the Fed to make a journey back to something like a corridor system, or something closer to it, but I have never set down my markers for exactly what I want it to be.” Lo and behold, Bill, you did that. You went out and you made a nice list of things that you would like to see in this ideal new framework.

Bill’s Suggestions for the Fed’s Implementation Framework

Nelson: That’s right.

Beckworth: Let’s go through that list and then maybe at the end of that, talk about, well, how do we get to it? Because there’s going to be some kind of transition. It may not be pleasant or easy, but we’ve got to get there.

Nelson: Right.

Beckworth: Your paper, your article, where you make this outline, is titled, “Remodeling the House After Tearing up the Floor”—I love that: tearing up the floor—“Suggestions for the Fed’s Implementation Framework.” You have some objectives. In fact, they’re very similar to the ones you mentioned a minute ago for the Bank of England. Not quite the same, but similar: good interest rate control, cheap central bank liquidity, support financial stability, efficiency, active interbank market, Fed monetary policy independence, and low credit and interest rate risk as a part of this.

You have these objectives. I’m going to start throwing out things that you state here and then you can explain why this is important. The first one, which is really interesting, is you want to make sure that there’s a target, in fact, a point target for the federal funds rate, not a band. You definitely don’t want to do like SOFR or some repo rate. Tell us about that point.

Nelson: Right. Two issues. Basically, very much on the narrow issue of the target, how they articulate it and how they achieve it. Well, not quite yet how they achieve it. Before 2008, before, I think, December 2008, the FOMC for a couple decades, would announce what their target was for the federal funds rate. What’s, I think, not well understood now that a lot of time has passed since then, is that they didn’t really need to then do anything. When they announced the new target, they would move the discount rate, interest rates would simply move to that new target.

It wasn’t complicated, even though it’s often portrayed as such, to operate a corridor system. A lot of that was that a lot of the heavy lifting was done by what’s called open mouth, or what was called open mouth operations. You don’t do anything. In fact, I will say, as a very young economist, the first time I walked into the room where we were going to talk to the New York Fed about where to set reserve balances to achieve the new interest rate, on the day after the FOMC had changed that rate, I was so excited. I was like, “Now I’ll finally see it in action.” I was a little bit deflated to realize that they weren’t changing their provision of reserves at all. That wasn’t how—

Beckworth: Interesting.

Nelson: —the interest rates are moved. You move the discount rate, you announce a new target. Banks’ demand for reserve balances were very sticky and inelastic. People knew what the banks wanted, and they just continued to provide it, adjusting for small, special factors. Announcing a target helped. Whereas they moved to a band in 2008, initially because they didn’t really know where the rate was going to settle, and they didn’t want to have to. Also, later on, it was partly because they never did set, back then, the IORB rate to zero.

They might have eventually, but they weren’t targeting zero, and they didn’t want to really focus on that, because that was mostly to make sure that money funds could keep operating. That they had at least a little spread. It was technical reasons, but it’s not helpful to have a spread. Since we’re now well above the zero lower bound, the Fed funds rate is rock solid relative to the IORB rate. There is no mystery about where it’s going to be. They should go back to simply announcing a target. That will help when they engage in this transition.

Now, I also, as you mentioned, have long been a believer in the Fed targeting the federal funds rate and not the repo rate. That’s true even though the federal funds market now is just a sort of an odd arbitrage between branches of foreign banks and federal home loan banks. It’s not really a place where—because it will be, eventually. The important thing is that it was easy to control that market. It was a small market, the Fed had a very small imprint. They moved around the quantity of reserves by engaging in transactions in a completely different market, with a completely different set of counterparties that were themselves fairly small.

They would engage in open market operations with primary dealers who don’t hold reserves. They had great interest rate control. The repo market is huge. It’s vast. Just like central banks that try to control the exchange rate and have to battle and then lose, the Federal Reserve, if you give the staff the job of controlling the repo rate, they will go at it. You cannot say, “Control it, but don’t keep it too tight.” That’s their job. They will necessarily then have to take very large positions to control the repo rate, which is moved around by a lot of idiosyncratic reasons.

Beckworth: Yes. You’re saying it’s much easier to go into a smaller market, and they have direct connection access. The whole point of reserves is only banks have reserves at that part of the Fed’s balance sheet. Also, repo markets are larger. I’m just thinking of like, if I can call it, the jaws of the Fed’s operating system. At the top, you’ve got the upper jaw being like the discount window, the standard repo facility, right? I guess the repo facility for foreign institutions as well. At the bottom, you would have the interest on reserves. I guess overnight reverse repo would be the repo facility at the bottom.

Nelson: That’s right.

Beckworth: I guess those would become almost more important than anything else if you did go to a repo facility, because those would be what keeps the repo rate between the two. You’re saying that’s just a whole lot more work and engagement in a big market.

Nelson: Yes, that’s right.

Beckworth: Okay. Now, you also outline a corridor system explicitly, so the federal funds rate would be in the middle of it, and then you say that it would need to be 50 basis points on either side for that corridor. Is that right?

Nelson: That’s right.

Beckworth: Why 50 basis points?

Nelson: I was lucky enough to be very involved in the design of the discount window in 2003. That was when the discount window was moved from a below-market rate to an above-market rate. When it was at a below-market rate, there had to be a lot of rules about who could use it and when. We concluded that those rules, or we’ll call it administration sometimes, actually contributed to the severe stigma that’s associated with using the window. People don’t like to borrow from the discount window.

In that regime of rules, Betsy Duke, former governor of the Federal Reserve System and a banker, community banker, described going to the window as like borrowing money from your parents. You don’t want to do it, but you’ll do it if you have to. We went to an above-market rate and then got rid of all the rules. Basically, we wanted to let the spread itself make banks decide on their own to just use it as a backup or when the Fed funds rate was tight, getting rid of all the rules. You need to have some daylight between what you’re lending and market rates in order to operate it as a no-questions-asked facility.

Unless you want to become an ongoing source of funding, and the Fed decided back in the ’20s that they wanted people to be willing to use the window, but they didn’t want to be an ongoing source of funding. That remains, I think, the philosophy. You need a spread above it. We decided it needed to be 100 basis points. Having a high spread itself can contribute to stigma. It’s a tradeoff, right?

Beckworth: There’s a tension.

Nelson: Yes. Maybe 50 is a reasonable amount.

Beckworth: Okay.

Nelson: Now, on the other side, one of the great things about my post-Fed job here in the Bank Policy Institute is I get to go out and talk to a lot of bankers. I don’t have to theorize about some of these things. I can just say, “Well, how about it?” Talking to a treasurer of a G-SIB about the likelihood that the federal funds rate could recover in a world where the Fed allowed money market rates to rise above IORB, he pointed out that, “Well, there’s a lot of resistance now. People aren’t used to that. You’d be making a big, uncollateralized exposure to some other bank. Your credit committee would be concerned. People don’t run Fed funds desks anymore, and you’d have to get everyone comfortable with it.” 

I was like, “Well, so 10 basis points?” He’s like, “I don’t know.” I said, “Well, 50 basis points?” He says, “Well, 50 basis points, now you’re sort of talking.” That made me think that, at least at the outset, you’d need to have enough there to get that market to recover, to get institutions to come in and start lending to each other at the end of the day.

Beckworth: If the price is right—

Nelson: That’s right.

Beckworth: —we’ll regain our muscle memory on how to use the interbank market.

Nelson: That’s the great thing about markets. It’s hard to say, “Well, I don’t really know exactly how this is going to be solved.” You give people a profit motive, and—

Beckworth: That has been a concern, right? We’ve lost our muscle memory. We don’t know how to engage in this. What is this? Interbank market? Crazy. Everything has a price, so we can make it work.

Nelson: All those other central banks are indicating that this is a good thing to shoot for.

Beckworth: That’s what’s interesting, too, is the almost pervasive view that we need to resurrect, or if they have it already, make it even more robust, an overnight, unsecured lending interbank market, because it’s important that they, one, as you mentioned, go to each other first before they go to the central bank. Two, I think, is also price discovery. How healthy are banks? Should banks be lent to? Sometimes it’s easier for the banks themselves to police each other. Okay, great, great points.

You also mentioned—I’m going to have to quickly go through some of these—primary credit as liquidity backstop. I think we touched on this already, but go to the discount window, no questions asked, if that spread is high enough. I love what you mentioned here: “Discount window borrowing capacity is economically equivalent to reserve balance.” From a liquidity regulatory perspective, whether you have reserves, or you have capacity to discount, one of those should count the same. We’ve talked about this before. That’s actually part of conversations going on now with regulators. This will be part of the story.

You also want to bring back something like TAF, but a little bit different, weekly auctions of one-month and three-month primary loans. You said you would call this primary credit, you wouldn’t call it TAF. Why is the name important?

Nelson: The Term Auction Facility is a facility that was operated during the Global Financial Crisis in which the Federal Reserve auctioned term discount window loans. Same set of counterparties, same collateral, same legal authority as the discount window. They provided the funds through an auction. It was motivated by the very severe dislocation in term interbank funding at that time, as everyone pulled back. They called it the Term Auction Facility because one reason was they were trying to get around the stigma associated with primary credit.

My suggestion is actually coming from the exact opposite perspective, which is we want to flood the zone with primary credit. We want there to be primary credit everywhere, so it is not considered so odd and so alarming. In this case, we want the comfort with the TAF to make primary credit less stigma-laden. The important thing would be to call it primary credit. Now, when the data is released two years later, people could look and they could see. By then, that’s a lot later. If just having people showing regular primary credit, I think, in volume would be beneficial. I’d note that this can be done in a corridor system or a floor system. You don’t have to be in a corridor system. You can simply say, “We want a big chunk of our assets to be auctions of discount window loans, and we’re going to conduct those auctions and grant them.” Then that becomes part of your toolkit. 

I will say, and I know we’re short of time, but the other points about the discount window, and I get into voluntary reserve targeting, is to try to solve that triumvirate problem or tripartite problem, because while I’m suggesting you need to leave daylight between the deposit rate and money market rates, I’m also pointing out that discount window capacity, the capacity to borrow from the central bank, is economically exactly like having a deposit at the central bank.

If you provide that generously, and recognize it, you’re providing liquidity for free. Also, I suggest having voluntary reserves that pay just a little bit below market rates, so that banks could choose, if they want, for liquidity purposes, to have a lot of reserves. That they could do that, although giving them a little encouragement to get smaller.

Beckworth: Yes. That was the next thing I was going to mention. You touched on it already, but voluntary reserve targets. It wouldn’t be required reserves, it would be voluntary reserves. Therefore, that would be an important part of your story, also an important part of the transition to getting there, correct?

Nelson: That’s right. If, and in particular, let’s suppose banks felt like they wanted to have a large quantity of reserve balances, that was the safest means, they felt, to meet their liquidity needs. Maybe, perhaps more likely, the banking agencies were unsuccessful convincing examiners to change their mind. Then our voluntary target is going to be a big number. I suggest that the compensation be about five basis points below the target because I want banks to still have an incentive to economize so that the Fed does get smaller. It’s close to the cost of doing that is low. Banks, if they choose to, could use that as a source of liquidity for contingencies.

Since they’re voluntary, unlike required reserves, they really would be a source of contingency funding. You may not recall this, but under the LCR in the United States, required reserves did not count as liquid assets toward the requirement. Voluntary reserves do.

Beckworth: One other really interesting thing in your proposal is to make the TGA reinvest in repo. As soon as funds come in, you invest them out. Is that right?

Nelson: Yes. This is really in some sense just a little balance sheet optics, but I think it matters how you think about things. The Treasury maintains a deposit at the Fed. That deposit used to be like $50 billion, and it was very stable. It wasn’t really a factor for monetary policy except for once in a while, Tax Day, that kind of thing. Now they target $800 billion, it’s a huge number. It’s ostensibly there to help meet contingencies. Really, it’s there to give them a built-in buffer against debt ceiling debacles. That’s not based on inside information. It’s just common sense. You can read between the lines.

Beckworth: Public knowledge.

Nelson: That means that when you start getting close to that debacle, you run down your cash holdings. Since that liability is going down, another Fed liability, reserve balances have to go up. That causes big swings in reserve balances. That’s part of the reason people have given for why we now need a corridor system. First, the Treasury is not compensated for those balances. They don’t get any interest on it. It actually helped boost Fed profitability, in fact.

One way that this could all be done instead is the Fed could just engage in repo term or overnight repo to largely match the Fed’s deposit. That would create a new asset equal to that liability and leave reserve balances unchanged. That really matters for right now, for example, the TGA is—I forget exactly—it’s between $500 billion and $600 billion. They’re running it down because the debt limit is beginning to bind. That boosts reserve balances. The Fed, the desk manager, in two recent briefings of the FOMC that were reported into the minutes, discussed a problem with this. That’s as the TGA moves down and reserve balances go up, QT is continuing.

They’re concerned that when this situation reverses and the TGA goes back up and reserve balances fall, they’re now, by design, going to be falling to a somewhat lower level, and they’re worried that it will fall below the level now where they should have stopped. They’re concerned about that. At the most recent meeting, they actually suggested, well, maybe we should stop QT right now in order to avoid this problem. If they’d been reinvesting the TGA, this wouldn’t be a problem. As the TGA fell, they would allow that repo to roll off, and that would leave reserve balances fixed.

What I’ve suggested in a separate blog, is since they’re not reinvesting the TGA, as it goes down, rather than stopping QT—because we both know that if they stop, they won’t start it again—that they invest in reverse repos, in term and overnight to simply offset that decline and leave reserves where they are. And then they could unravel that when the TGA goes back up. The New York Fed, of course, has always been the biggest advocate of a large balance sheet approach. I would be careful following their advice on whether or not they should stop QT simply because they might end up in the future close to that balance sheet.

Beckworth: This idea seems very intuitive, very clear. It seems like it solves a big problem that everyone’s brought up, the autonomous factors, this exogenous movement in the balance sheet or the composition of the balance sheet, based on something completely outside the Fed’s control. Has no one else discussed this as an option? It seems so obvious now that I read it from you.

Nelson: I bet people have. It’s a big number. If you’re thinking about engaging in these repo transactions, that’s a big number. You’d have to be prepared to do it. Since we all now know that it’s just not the case that you can provide all these extra reserves and then they just sit idly there in banks’ deposit accounts at the Fed and can move up and down with the banks not caring. That just doesn’t work. In honesty, the Fed needs to offset those big swings under all circumstances.

Responding to Lorie’s Case for the Floor System

Beckworth: Well, let’s move on from your article. We provide a link to it in the transcript and also, I believe, at the show notes we’ll have it as well. Let’s move on to another article of an individual who’s incredibly smart, great communicator, and probably makes the best and most articulate case for the existing floor system. That is President Lorie Logan, also a previous guest of the show. In fact, I’m hoping she’ll come on again. Maybe after this conversation, she’ll be driven to do so even more. She had a really interesting speech on February 25 titled “Efficient and Effective Central Bank Balance Sheets.”

I just want to highlight a few claims that she makes and have you respond to her. The first one is one that she’s made before, that is, this efficiency claim. She says that we should be efficient with our operating system. She talks about allocative or Pareto efficiency. Ultimately, she ends up with the Friedman rule. She argues that a floor system is more consistent with the Friedman rule than, say, a corridor system. How do you respond to that?

Nelson: First of all, I want to agree with you in terms of Lorie as being a very articulate defender of the floor system. Although, I’ll note that even in her speech now, they don’t really seem to be talking about having a $200 billion to $300 billion buffer above where they need to be. They, too, seem to be taking the message of where they want to go. Chair Powell, in transcripts that have been released, has indicated strongly, he would like the Fed to be smaller. Just to be clear, the classic Friedman rule is, of course, the constant money growth rule. That’s not the Friedman rule being discussed here.

He also is associated with the idea that the rate of inflation should be set equal to the negative of the real interest rate so that people who hold currency, their opportunity cost of doing so is zero. Ed Nelson, who I think also has been a guest on your show, and a deep scholar of Milton Friedman, has indicated that that’s not really a rule that he associated himself with. He said that that’s logically correct, but he isn’t really a fan of it. I think that Milton Friedman would be rolling over in his grave to contemplate a Federal Reserve that was constantly 26% of the market and made up more than half of money fund assets. 

Beckworth: Justified by his Friedman rule.

Nelson: Yes. I think he would disassociate himself from that. Nevertheless, President Logan’s point is that the Fed can produce reserves for free. Social efficiency indicates that they should do so until the cost or the demand is set equal to that free level of price. By free here, it’s the interest rate in markets should be equal to the interest rate the Fed pays. I think that that’s mistaken for a number of reasons. One is, as we’ve been discussing and have discussed in previous shows, creating reserves is not free. It presents all these problems for the Federal Reserve with respect to its independence potentially.

There was some cognitive dissonance in my part, listening to a paper by a New York Fed economist being presented at the annual Columbia BPI Conference that took place just a couple days after Lorie’s speech, where the paper was discussing how the over $2 trillion move of money out of the banking system into the money market mutual fund system, when reserves were no longer excluded from the denominator of the leverage ratio, demonstrates powerfully that the balance sheet costs associated with making banks hold all these reserves, or really lend to the Fed in the amount of these reserves, is in fact quite costly to the banking system. It’s far from free.

It moved around, you could see it in the move, and this was a movement to a means of borrowing from money funds in which the Fed was paying even 10 basis less than borrowing through reserve balances. It was so costly to the banks to provide those reserve balances that the new equilibrium was one where all of that funding moved or a lot of the funding moved to money market mutual funds. There’s just all kinds of reasons to understand that it’s not free. When you’re really thinking about it, nobody really questions whether or not the Fed can have good interest rate control in a floor system or a corridor system. That’s not up for debate.

If you’re debating how should they be doing it, suddenly whether it’s costly to have such a huge balance sheet, that becomes a first-order problem that one needs to consider. It’s striking that there are such divergent views on the cost. I, of course, as we’ve discussed earlier in the call, put a very high premium also on having a robust interbank market, having the Fed not be the one that’s there to solve all of the market’s problems.

Beckworth: That would be another cost, in your view, that they’re overlooking in making that claim.

Nelson: Yes, absolutely.

Beckworth: You’ve written elsewhere that the Friedman rule, the logic of it, can also be satisfied in a corridor system.

Nelson: That’s right. Thank you for that point. It can be satisfied in a corridor system, and it is satisfied in two ways in the manner that I suggest. One is if the banks want to hold voluntary reserves, then they’re free to use them to meet liquidity purposes. It would effectively provide something close to that market rate as compensation, still enough to encourage an interbank market or to encourage the Fed to get much smaller. Also, the discount window is there and should be recognized as a means of liquidity. The Fed has made some good steps in that regard. Really, that’s a very robust way to provide that Friedman-like contingent liquidity on the part of the central bank, and that’s provided for free.

Beckworth: Okay. When it comes to the Friedman rule, they make a claim on their side, and you say, well, I have an answer. It satisfied both systems. We will take that off the table as a deciding factor for now. 

Let me read a paragraph that she has in response to your interest, as well as these other central banks who seem to be really interested in interbank health and robustness. Let me read here a paragraph or two. She says, “Some observers have objected to keeping money market rates close to interest on reserves, saying this suppresses trading in the interbank money market. In the face of unexpected payment shocks, an active interbank market helps the private sector redistribute funds across banks. By comparison, with an interbank market, the central bank may intermediate more of that redistribution through its deposit and lending facilities. Policy preferences can differ about the desirability of tradeoffs of smaller central bank footprints against the spread of market rates over interest on reserves and the resulting inefficiency that would be needed to incentivize active interbank trading. While that’s an interesting philosophical discussion, I”—Lorie Logan—“don’t see much of a practical argument for the US at least.”

“When the U.S. banking system has excess reserves, they tend to accumulate at the largest money center banks. The regulations applicable to such banks disincentivize unsecured lending to other banks. As a result, only very large spreads between money market rates and interest on reserves would likely suffice to revive the interbank market.” 

Her point is, given the complexity of the US banking system and maybe different regulations, the reserves, they’re hoarded in particular parts of the system, and to get them out, you’d have to have big, big spreads to make it work. How do you respond to that?

Nelson: In several ways. One is—and this is something that I discussed in my note about the Bank of England—I think that it gives us a poor intuition for how these things work if we describe reserves as if they were some special thing that banks lend to each other. Reserves are just a deposit in your account at the Fed. A bank doesn’t only borrow reserves when it’s borrowing from the Fed, or even when it’s borrowing in the interbank market. They’re borrowing reserves when they issue CP. They’re borrowing reserves when a household makes a deposit. Reserves are just a deposit in the account.

If you think about them, “Oh, well, they’re all being hoarded here and not there. You’re treating them as if they were sort of blue dollars versus red dollars.” The reality is any bank that wants to have reserves can have reserves. It isn’t something that you get in the interbank market. It’s just you choose to have a higher deposit. You can pay out less. Just like you and I. If I decide I need to have more money in my checking account, it’s not as if I need to take out a personal loan from the bank that provides me that checking account to add to my reserves. I just spend a little less. That, I think, it’s a sort of a subtle point, but it really confuses the issue frequently.

I have often heard this view that, there used to be all those resources going into this daily interbank market, and that’s really unnecessary now because we want to have all these extra reserves. That’s just money that’s not well spent, having a market, having that price discovery, as you just said. I think in the end, it comes down to, do you want markets to solve problems for us to allocate resources generally, or are you comfortable with the government doing so? I am very much a believer in the value of having markets there.

As you mentioned, I was involved in designing liquidity requirements, and one of the things that I think was not well understood by those designing it, because they were often lawyers or supervisors, was the importance of the interbank market. Banks don’t need to have enough liquidity, gold in their basement, to solve day-to-day liquidity needs. That’s what the interbank market is for. The magic of that is that it diversifies that liquidity risk across the system, and banks can devote much more of their balance sheet to lending to businesses and households rather than keeping money in big cash accounts at the Fed. That’s pro-growth and pro-stability and pro-markets.

Beckworth: Now later, when she talks about the assets side of the balance sheet—so I’m jumping ahead, but I think it’s important—she talks about an idea that you like, and we talked about it a few minutes ago, reintroducing TAF or some version of it. Wouldn’t TAF address this concern she has that reserves get clogged up in a certain part? Anyone could go to the TAF and get these reserves if they’re needed. Is this really such a fundamental issue that it’s one that can’t be solved through some clever use of other facilities, or maybe even change the regulations so the big banks don’t have to hoard these reserves?

Nelson: You’ve just done what I’ve said is a common source of confusion, which is to suggest that somehow a bank would need to borrow from the TAF in order to have more reserves if they want more reserves. Any bank anywhere can have more of the reserves whenever they want. They don’t need to borrow from the Fed. It’s just it’s a choice they make all the time and each bank chooses it. It is true that this could be done without moving to a corridor system. It’s compatible with a floor system. Floor system is about the quantity of assets that you auction.

I forget how many assets they have now, $6 trillion, they could have $5.5 billion and $500 million in TAF loans. It’s really the capacity to borrow from the Fed rather than having borrowed from the Fed that provides you liquidity. Liquidity is that dry powder that you have, the ability to meet a need. Once you’ve borrowed, you’ve actually used that capacity. Borrowing doesn’t create your reserves. You can do those in any case. Borrowing does use up your liquidity resources by using up your capacity to meet a contingent need by borrowing. That’s any form of borrowing, whether it be from your correspondent or from your federal home loan bank. It doesn’t have to be the Fed.

Beckworth: Another point she makes in favor of the existing system is that it saves on forecasting and operational effort. You don’t have to be trying to forecast reserve demand every day. I know we’ve talked about this repeatedly. One, you mentioned, well, look at the actual office that manages the balance sheets. It’s actually gotten a lot bigger, so there’s that. Also, yes, maybe you’re not estimating day-to-day reserve demand, but guess what? You do have to be estimating that kink in the demand curve.

I think 2019 is a great example of this where the Fed’s like, “Oopsie, we slipped back into scarce reserve system. We didn’t mean to, but taxes were flowing and we’re doing QT and who knew.” Things happen. To me, it’s not such a clear-cut case.

Nelson: That’s right. No, I think that that’s absolutely right. On all points, first of all, the number of human resources involved were relatively small. These simply should not be factoring into such an important decision about how the Federal Reserve interacts with the financial system. Secondly, as you’ve pointed out, there have been massive growth in both sides, in the New York Fed and in the board, in the units that did this job. I think that having a few more people working on that. Just the very fact that that came up gives you a window into how inside baseball plumbing this is being driven by as opposed to a broader conceptual how should the Fed interact with the financial system. Should markets be the first line?

As I mentioned, it wasn’t that hard. When they operated by announcing a target, rates just moved to that target. The whole idea that you talk about, the steep part of the demand curve, really gives a false impression. You don’t have to precisely set reserves in that place where actually you know how much banks generally need, you know what taxes are going to do probably, you know all of these other things, and you make minor adjustments. Then, if things are off on one day, that’s what the corridor’s for.

I agree with you that 2019 is an illustration of if the Fed really wants to be smaller, then it’s going to need to be closer to that edge. If it isn’t close to the edge, then it’s going to get ever bigger because the demand for reserves will just keep going out. They’re going to need to be closer. That means that they need to pay attention to these big swings, and they need to counteract them like the TGA right now. Hopefully, they won’t counteract them by stopping QT. Rather, they will take their responsibility to manage those swings.

Beckworth: Bill, in the time we have left, there are a lot of fun things in her speech. We’re responding to some things. I encourage listeners to go out and read her speech. We’ll have a link to it in the transcript. Several things she mentioned I really like. We already touched on TAF. We both champion that along with her.

I like her idea of a neutral portfolio that would match the Treasury’s average-weighted maturity of what it’s issued. Try to basically mimic what the Treasury is doing. She says, right now, it’s not practical. There’s a lot of things happening, QT, but in the long run, that would be her steady state solution. What do you think about it?

Nelson: I see both sides. Of course, the other option is to have a shorter-dated portfolio.

Beckworth: All bills.

Nelson: It needn’t be all bills, but it could be skewed toward bills. I think I see both sides of this. One thing I think I often note is that the very idea that just owning Treasuries is the least intrusive way for a central bank to interact, that’s a very US-centric view. The Bundesbank, for example, and their philosophy was inherited by the ECB, their view was just owning government securities was the worst thing that you could do, because you’re monetizing the debt. It’s going to lead to hyperinflation. They thought the least distortionary thing you could do would be to invest in banks and have a broad portfolio of broadly collateralized bank securities. They reached the exact opposite decision, in some case.

With respect to the maturity of the Treasury portfolio, one of the many projects I’ve had the pleasure of working on—and now that I’m old—was what maturity should the Treasury portfolio be, when I was back at the Fed. The traditional story for why you might want to skew toward reserve balances is, if there has to be a big discount window loan, you need to sterilize it. You’re making a big loan that’s adding a lot of reserves. If you want to continue with your corridor system, you need to sterilize it, and the way you sterilize it is by either letting repo or bills roll off.

There would be a similar situation that they will be in. It is true that if you want to maintain some daylight between money market rates and the interest on unreserved balance rate normally, then it’s good to have tools that allow you to sterilize big transactions. Even then, it’s not that big of a deal. If you’re in a crisis and you don’t sterilize it, all that happens is you move back into a floor system. If done for a temporary crisis, that’s not that big of a problem.

Beckworth: Well, I love this discussion. You mentioned the German view. I had Ulrich Bindseilon many years ago on this podcast. He’s an expert in this area like you, and it really broadened my thinking and maybe for the American monetary policy person, a little bit heretical, but I’m like, “Man, that’s a great idea.” Not just bank stocks, but in general, if the Fed wants to minimize its footprint on the financial system, then in theory, have whatever the average portfolio is out there, which would include stocks, corporate bonds, and I know that raises all kinds of questions. That would be more neutral than just having a certain segment of a class of assets.

Nelson: Yes. Yet another fun project I got to work on was back at the end of the 21st century, everybody anticipated that the federal debt would be paid down. Very sober, clear-thinking minded people were all projecting that it would be paid down, and the Fed needed to come up with a different asset to buy. There was a big project about, well, what should we buy instead. On the table, this would require a change in the law, would be a broad portfolio of corporate bonds that would be AAA, safe, riskless. Another option that got a lot of attention was the TAF.

This was back in 2000. It was called the Auction Credit Facility, but it was actually auction discount window loans. They just pulled that off the shelf when the crisis hit, renamed it to the TAF, and that’s how they used it. That’s also where I pull the principle that I articulated as one of the objectives that I think that they should maintain, which is to keep interest rate risk low. This idea of operating with very high interest rate risk has been a disaster. Their own principles say, we would like to keep interest rate risk low and should only do so, only take on a lot of interest rate risk after talking to Congress and Treasury.

Given all the money now that they’ve lost with having taken such a big bet, I think if you want to shorten the maturity of the debt in private hands to lower term rates because you’re at the zero lower bound, let the Treasury do that.

Beckworth: With that wisdom, our time is up. Our guest today has been Bill Nelson. Bill, thank you for coming on the program.

Nelson: It’s always a pleasure. Thank you for having me, David.

Beckworth: Macro Musings is produced by the Mercatus Center at George Mason University. Dive deeper into our research at mercatus.org/monetarypolicy. You can subscribe to the show on Apple Podcasts, Spotify, or your favorite podcast app. If you like this podcast, please consider giving us a rating and leaving a review. This helps other thoughtful people like you find the show. Find me on Twitter @DavidBeckworth and follow the show @Macro_Musings.