Bill Nelson on the Future of the Fed’s Balance Sheet

Bill Nelson is a chief research officer and chief economist at the Bank Policy Institute. In Bill’s 10^th appearance on the show he discusses his infamous email list, the ratchet effect from QE, his congressional testimony, the BPI’s Bank Treasurers Survey, how he thinks the Fed should shrink the balance sheet, whether the Fed is profitable, and much more.  

Subscribe to David's new Substack: Macroeconomic Policy Nexus

Read the full episode transcript:

This episode was recorded on March 3rd, 2026

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 research officer and chief economist at the Bank Policy Institute. He is also the former deputy director of the Division of Monetary Affairs at the Federal Reserve Board. Bill joins us today to discuss some recent research on the structural demand for reserves, as well as some updates on the Fed’s balance sheet and his testimony on that issue before Congress. Bill, welcome back to the program.

Bill Nelson: Thanks for having me, David. It’s always a pleasure.

Beckworth: Bill, this is appearance number 10.

Nelson: That’s crazy.

Beckworth: It is.

Nelson: That’s really surprising.

Beckworth: That means you get an official golden jacket, although we haven’t issued one yet. It will be the first. George also is entitled to one—George Selgin. Ten times—the reason it’s been so many times is because this is an issue that we’ve talked about before, the Fed’s balance sheet, the operating system, the structural demand for reserves. If this issue continues to be something that doesn’t get addressed, I’m sure you’ll be back on for 10 more.

Nelson: Maybe it doesn’t feel like 10 because I love talking about these things, and I love sitting here talking with you, David. You always make it very natural and easy.

Beckworth: Well, I have learned so much from you, and I appreciate you taking time out from your busy job to come over here and discuss it. One of the things I do is I have you on. Sometimes I go out to eat with you, but I also follow your newsletter. You have a really good newsletter that you send out on the email list. I was thinking maybe the listeners and the watchers may want to know about it. How can they get on that list?

Nelson: It’s simple. They just need to email me at [email protected] and ask to be added to the distribution list. It’s free, and I’ll be happy to add them.

Quantitative Easing Ratchet Effect

Beckworth: Okay. I encourage people to do that because it is very educational. I’ve learned a lot, and it’s helped shape some of my thinking, some of the issues that I follow. All right, so, Bill, last time you were on the show was, I believe, early last year in 2025. It’s been a while. A lot has happened. Probably the biggest balance sheet development since then is that we’ve come to the end of QT. We saw an end to it. We came from a high of $8.9 trillion, the size of the Fed’s balance sheet, down to $6.5 trillion when they stopped earlier this year.

Once again, we see a QE ratchet effect. Now, it’s important to note—and you know this, but for some listeners and watchers of the video—that even though the reserves did fall significantly, the ending level is still higher than the starting point. It may be a little confusing when I say there’s a ratchet effect. You’ve got to look at the level of reserves at the starting point, ending point. It’s not at the peaks and the downturns, right?

Nelson: That’s right. It’s a bit deceptive this time because a lot of the runoff in the assets actually corresponded to a reduction in the ON RRP facility, the facility that the Fed provides to money funds basically to invest in the Fed. That’s a source of funding, and it’s a liability. Most of the decline in their asset holdings during QT was actually matched by a rundown in the ON RRP facility. If you look at reserve balances, they’ve bounced around between $3.1 or $2 and $2.8 trillion. Where they are now, the ratchet is firmly entrenched.

Beckworth: Yes. The ratchet effect deals specifically with the level of reserves, whether it’s an absolute level or even relative to GDP. You had a nice table in a recent note. We’ll link to it in the transcript. Just maybe to put some numbers to this claim of a QE ratchet effect. Of course, April 2008, before we go to the ample reserve system, the Federal Reserve’s own estimates of the structural demand for reserves was $35 billion.

That was a long time ago—simpler times. November 2016, you note that they have it at $300 billion. November 2018, it jumps to $1 trillion. October 2019, before we get to QE under COVID, it was $1.4 trillion. Now, it’s around $3 trillion. That’s just the reserves. This doesn’t include the other liabilities. If you look at those numbers, you can draw a nice trend going up through those points. Even if you do it as a percent of bank assets or a percent of nominal GDP, it’s going up. It’s ratcheting.

Nelson: Yes, right. After talking to a lot of banks and understanding what was going on and observing how we were all so surprised back in September 2019, or even before, a year before that, how we were seeing tightness in markets before getting back down to the levels where we were, we started to talk to people about what’s causing the ratchet.

The general sense, and what I learned was, unlike the classic pool model, where you just push out all these extra reserves and they just sit there idly on banks’ balance sheets doing nothing ready to be subtracted back away, when you create a lot of reserves, typically that means that money market rates are going to move down below IORB by a little bit, the interest rate that the Fed pays on their deposits, the interest on reserve balance rate.

That means it’s quite cheap to hold reserves as a liquid asset, and banks adjust their balance sheets to make use of it. Bank examiners adjust their expectations about the quantity of reserve balances banks will have, and then bank liquidity regulations are written with this expectation that reserves are going to be this abundant, low-cost, practically free thing to use to satisfy liquidity needs.

One bank treasurer explained to me that back in 2016, when IORB was above the repo rate, they were like, “Well, we only have to hold two days’ worth of cash to meet our needs, but we’re going to hold five days’ worth.” Then, when the rate configuration changed, they said, “Well, should we go back to the way we were?” It’s like, “Well, we just really don’t want to have that difficult conversation with our examiner about why we’re holding less cash.” It’s harder. It’s not impossible. It’s just harder to push reserves down than to push them up.

Beckworth: Yes. One thing I’ve really learned from you, Bill, is the importance of the interaction between bank regulations and monetary policy. Sometimes they conflict. They’re at a crossroads. I think this is a good example where liquidity regulations, internal liquidity stress test, or the liquidity coverage ratio, all these things interact in ways that aren’t always optimal. You used to work on these at the Fed. I remember you telling me when you were writing some of these rules that no one had in mind the Fed’s balance sheet would be this big, reserves would be this big, and they’d become like a binding constraint with the leverage ratio, right?

Nelson: That’s right. No, if you go back to when the Fed adopted the ESLR rule, it was a leverage ratio requirement—there was an additional requirement for the largest institutions—you can watch the video. Everyone around the table, Dan Tarullo, Janet Yellen, Jeremy Stein, everybody expressed concern about having a binding leverage ratio. You don’t want a risk-insensitive capital ratio to be your binding requirement because that creates an incentive to make yourself riskier. They all said, “Well, but we’re not that worried about it.” In part because risk-based requirements were going up, but in part because that was at a post-Global Financial Crisis high for the reserve balances.

At that point, the projection was that they would return back down to $25 billion. You can just look at the Fed’s internal documents, which have been published. That just didn’t happen. Then, just quite recently, in the last several months, the banking agencies have finally adjusted the ESLR. That is no longer binding. No, the last several years of my time at the Fed—I left in 2016, I spent my career in monetary affairs working mostly on monetary policy and discount window stuff—I was seconded for part time to the supervisory side. I helped oversee the largest institutions, was on the steering committee for the horizontal liquidity stress. One of the last things that I did was, with my monetary policy hat, QT was going on, reserves were being run down. Then I started to learn, in my supervisory side, that the examiners were out telling banks they wanted them to hold more reserve balances. I made sure to try to get those two sides of the Fed talking to each other.

Beckworth: Yes. This is all so interesting. I really appreciate you really bringing to life the role that bank supervision plays in this process. Why can’t we do full QT? Because these expectations get baked into what banks should hold. I recently had Raghu Rajan on, and he brings another complementary story about how it affects the balance sheet structure of the banks. It’s hard to reverse QE. I’m someone, I think you are, too; we think QE is an important tool for the Fed to use when it’s needed. It’s just we need to be mindful, “Can we fully reverse it on the back end?” We’ve had, now, several rounds of QE, and apparently, the answer is no, with the current setup. Maybe we need to revisit the current setup. We’ll come back to that question in a minute. 

Today we’re recording on March the 3rd. You just came back from a roundtable on liquidity and lender of last resorts. My understanding was that vice chair of supervision, Governor Bowman, was there. Secretary Bessent was supposed to be there. What did we learn from this setting?

Nelson: It was a very interesting roundtable. It’s all under Chatham House, so I can’t list who all was there and what everyone said. The discussion over the course of the day was very focused on this issue of liquidity regulations and how they could be improved. Vice Chair Bowman opened the event with some remarks that you can find on the Fed’s website, generally indicating that there was certainly scope to reform the liquidity requirements that were designed after the crisis. With 10 years of experience, it was time to give them another look. She also spoke about the discount window and daylight credit, that banks were very reluctant to use it. That was an important part of the Fed’s toolkit and the resilience of banks in the financial system. 

Then, Under Secretary McKernan, delivering Secretary Bessent’s remarks, went a little further and endorsed the idea that banks’ capacity to borrow from the discount window should be reflected in their liquidity assessments. The basic components of banks’ liquidity requirement framework is the internal liquidity stress test that they are required to conduct at the overnight, 30-day, 90-day, one-year horizon.

Then there’s the LCR, which most people, if they’re familiar with any part of it, the LCR, it’s a 30-day projection of needs. Then you have to have liquid assets to meet them. Then there’s also liquidity requirements associated with resolution, that banks need to have plans to be resolved safely. Those plans include liquidity requirements. Secretary McKernan indicated that there was support, that the Treasury was supportive of folding the recognition of the discount window into all of those. In my view, it’s four wins. Win, win, win, win.

Beckworth: Yes, long overdue.

Nelson: Yes, partly because it makes those requirements more accurate. It makes those assessments of a bank’s liquidity more accurate. As we saw in September 2023, banks that are prepared to borrow from the discount window are more liquid and more resilient than ones that aren’t. It can potentially help reduce stigma by sending a message that the Fed supports such requirements. The perfect source of liquidity is reserve balances. That’s money in your account. A very close second is the discount window. If you have collateral pledged to the window, that’s like having a HELOC or a line of credit at your bank. They’re very similar things.

Those are the two things that can, until the end of the day, provide you cash if you’re run on or whatever. By allowing that substitution, it enables the Fed to get smaller. The Fed can’t get any smaller than the quantity of reserves needed to implement policy. Then the final win is that it’s good for growth. About 25% of banks’ balance sheets are now full of assets that they’re holding to meet liquidity needs. If they are instead allowed to make loans to businesses and households, pledge those to the discount window, and point to that capacity as a source of liquidity, then they can devote a lot more of their balance sheet to lending for productive purposes rather than just lending to the government.

I’d add, in case people aren’t familiar with this, that most of the collateral pledged to the discount window is, in fact, loans, not securities. It’s often actually consumer and business loans because the real estate loans get saved for the federal home loan banks. Right now, two-thirds of the collateral is. It used to be about 80% of the collateral. There’s a huge pool of liquidity to be tapped there against loans to businesses and households.

Beckworth: That is so encouraging to hear because we have talked about this before, that this would be one way to make the discount window business as usual, minimize the stigma, as well as reduce the structural demand for reserves. I think anyone who has a view on this, whether they’re fans of an ample reserve or a scarce reserve or a ceiling system, everyone agrees we need to make our ceiling facilities more business as usual—be able to access them. I think everyone can support this. That actually was a question I was going to ask you. Last year, there seemed to be momentum building for something like this. It sounds like it’s still there, like there’s still hope these reforms will go through.

Nelson: It does seem like a promising time. These things are a bit complicated. They’ll take some time to work out. You have to decide, well, how much should you count? How do you determine whether or not banks are, in fact, willing to use the window? You don’t want to give them credit if they won’t use it. There’s a fair amount of details to be worked out. Nevertheless, it does seem to be now a front-burner issue for all of the banking agencies.

Beckworth: I recently wrote an op-ed for Barron’s, and I put this in several Substacks—since we’re on this, I’m going to bring it up now—where I outlined four concrete steps to shrink the Fed’s balance sheet because there seems to be a lot more interest in this right now. The number one step was that make the Fed’s ceiling facilities more accessible, more business as usual, be able to approach it. I outlined those very ideas, which I give credit to you, Bill. That’s not original to me. Something else, though, that you mentioned I included in that essay is revive and bring back the term auction facility as well. That’s another way to make people comfortable going to the discount window.

Nelson: Yes. No, I thought it was a great piece. You also mentioned term deposits, right?

Beckworth: I do, yes. There’s several things I put in there. The first thing, which I think would be the easiest, because this is a committee, FOMC is a committee, would be make the ceiling facilities more accessible, more business as usual. It makes me very hopeful to hear that they’re talking about it. I’ll mention my other ones real quick. That’s number one. 

The other one was—well, three more—to neutralize swings in the Treasury general account, which again, I got that idea from you, as well as Annette Vissing-Jørgensen. You both have proposals, very similar in spirit, where you don’t want big swings in the TGA to make you nervous and worried, so you’ve got to have lots of reserves, ample reserves. If you can isolate, almost partition the Fed’s balance sheet, then you don’t need to worry about having all this excess cash sitting around. That was number two. 

Number three was to do a Fed asset Treasury swap. I would say institutionalize it after every QE. Make it automatic so that the Fed’s balance sheet, you don’t have operating losses, and it’s easier to roll out the assets because they’re shorter term. You’d swap the Fed’s bonds for bills from the Treasury, and presumably would not be a debt ceiling issue because you’re just swapping debt, no net increase. 

Finally, the term deposit facility, we have that. We just haven’t used it. Now, of course, that is probably the toughest step because it would really be a reordering of the operating system itself. I think that would take a lot longer conversation. 

Bill’s Congressional Testimony

Since we’re talking about this, Bill, let’s go to your testimony. We’ll come back to a fascinating survey that you did on these issues. Since we’re talking about revisiting the operating system, let’s talk about your recent testimony before Congress, because it was about shrinking the Fed’s balance sheet.

Ultimately, it’s in this broader conversation that I think we’re having or beginning to have about that very issue. Of course, changing the Fed’s balance sheet, I think, inevitably, has to be tied to the operating system. You can shrink one, I guess, and keep the other, but it seems hard to do one without the other in practice. Do you get a sense as well that there’s more appetite, more openness for having these questions considered?

Nelson: Absolutely. I think so. The fact that, as you noted, around the world, it seems like all the other major central banks have basically reconsidered and rejected this idea that they want to have a giant balance sheet with a floor system. We can talk about that as well. I testified before the Task Force on Monetary Policy, which is a subcommittee of the House Financial Services Committee. Their objective is to just educate the public and educate Congress about what the Fed does and is doing.

That’s so important because it’s very difficult, I think, for Congress to exercise their crucial oversight role over the Fed, in part because they’ve got a lot of things that they need to be expert on, and then monetary policy, especially the plumbing issues that we love nerding out on, are complicated. I’ve testified twice now with them, and I think it’s great work that they’re doing. Yes, this was a discussion on the balance sheet.

I was able to testify with my good friends and longtime colleagues, Jim Clouse and Bill English, and basically made my case for why the balance sheet should be smaller, why it’s better for the Fed to have a smaller footprint, and then described, like your recent editorial, how they could go about doing that, the steps that I outlined that they could take to do so. It was a very friendly discussion. There were Republican witnesses and Democratic witnesses, but it was pretty hard to tell the difference because they were all sort of ex-Fed, almost all.

Beckworth: I think even advocates of the ample reserve system also want to see a smaller balance sheet. They understand there is a cost to it. Maybe we can come back to that in a minute. It does seem like there’s this opening now that we haven’t had in a long time to have these conversations, as you just alluded to. Almost every other advanced economy central bank has had a conversation on this. By conversation, I mean a formal review, not just hallway conversation, but a serious review. I had your colleague on here, Laurie Bristow, talking about the RBA, Reserve Bank of Australia’s review of their framework.

Isabel Schnabel came on. She’s from the ECB. They had a formal review of their operating system. These other central banks are actively looking at the plumbing, the implementation system. It’d be great for us to have it. Even if we end up staying with what we have, let’s sit down and let’s review and think about it. Again, where I land on this is, and the term I will use now, Bill, is a state-contingent operating system. I want to be in a place where if we need to do QE, and if the Fed’s balance sheet temporarily gets large, let’s do it. Then let’s be sure we get on a journey where we return to a very small balance sheet. I think the term deposit approach that the Riksbank uses is so clever, so ingenious.

Of course, all of this requires many moving parts coming together. We’ve got to have robust ceiling facilities because I want banks to feel comfortable going to the Fed and getting liquidity. Chris Waller, who I consider a great governor—I love his talks. I’m impressed by his ability to power lift and shoot 50-caliber rifles, if you’ve seen it on social media. I’ll give him a shout-out here. He actually was the managing editor of an article I just got published in the International Journal for Central Banking. Shout out to you, Chris. Although he does not seem very excited about going away from an ample reserve system.

I’ve heard him several times talk about, “Why would you want to have banks look through the couch to find spare change in order to find liquidity?” In my mind, if you have robust ceiling facilities, then they don’t have to. It’s like having a credit card, right? They can draw from the central bank. Or you might say, “Maybe there’s someone else on the couch next to them who they can borrow from. Revive the interbank borrowing market. You don’t have to look into the couch.” Now, recently, he went even a bit bolder and said it was stupid to want to look elsewhere. Did you see that comment?

Nelson: I did.

Beckworth: What did you think about that?

Nelson: Well, Governor Waller’s logic hinges critically on this idea that it is costless for the Fed to produce reserve balances. To an economist, if something is costless to produce and it’s valued by people, well, then you should produce it until their marginal value of it is zero, basically. Applying that reasoning, the Fed should overproduce and create reserve balances until, effectively, the money market rates equal the IORB rate. The rate that people exchange them with each other is the same rate at which they deposit them at the Fed. I strongly disagree with the idea that it’s costless.

Of course, the costs are not financial in the sense that for all of those reserve balances, there are assets on the other side of the balance sheet that basically earn what they need, and that washes out. I think the political economy costs are significant. As I argued in my testimony, it kills the interbank market because everyone is flush with reserves and everyone is getting paid the market rate on their excess reserves, so no reason to lend them out. I like markets. I like to see markets working well, and before the government gets involved. I’m surprised that Chris doesn’t agree with that view because he seems like a free-markets guy. It adds to discount window stigma because nobody borrows.

It creates an attractive funding mechanism for other parts of the government to see that they will then, with the FDIC, of course, funded themselves with the Fed during the 2023 event. During COVID, Congress tasked the Fed with creating facilities to lend to middle-sized firms. Once the Fed’s balance sheet is unhinged or unrestricted, it adds that much more fuel and attractiveness to impinging on the Fed’s independence. For other reasons, I really think it’s costly. If it’s costly for the government to produce something, then it is not true that they should produce it until its marginal value is zero.

As you noted, if we’re stupid, then the RBA, the ECB, the Bank of England, the Riksbank, the Norges Bank, and the Bank of Canada are stupid too because they’ve all looked at this and said—and Andrew Bailey has been quite specific about this. We think that the costs of providing abundant, vast quantities of excess reserves exceed the benefit. They’ve made exactly the same calculation that Governor Waller seems to have in mind and reached a quite different conclusion. It may seem like an obvious instance of stupidity to Governor Waller, but not to the other major central banks.

Beckworth: Yes. Governor Waller, if you’re listening, you are more than welcome to come on the show and share your thoughts on this discussion. Again, I really like Governor Waller and his work that he’s doing there. This is a question, I think, of there’s more than just one margin. I think you can make that argument he’s making, which is really the Friedman view, the Friedman optimal amount of money view. Ceteris paribus, holding everything else constant. The point you’re making is there are other margins where costs do come in, and they do affect. A true holistic view of costs would look at all these other margins.

I just want to go back and summarize what you just said so it’s very clear to listeners, although you did a great job, Bill. Why do we care about a large balance sheet? Because that’s something else that’s happened. There’s been recent calls for a smaller balance sheet, and some commentators will say, “Well, what’s the big deal? We’re functioning, things are going well. Who cares?”

The first point, I guess, to summarize the one you outlined, is ultimately this matters to the Fed’s independence. Everything you just said, using the Fed’s balance sheet for political ends. Whether good ones, maybe the FDIC needed help in 2023. You could easily see calls for it to fund other, more politicized activity. I think there were calls for it before 2020 for it to be used to buy up student loans, or maybe they want to use it to build a wall.

Nelson: Fund the Bitcoin reserve.

Beckworth: There was a bill proposed for Strategic Bitcoin Reserve that would include the Fed as part of this process. These are things we want to be careful about. Actually, to go back to one of the key arguments for such a regime, an ample reserve regime, is that you have two levers. The Fed’s balance sheet is separate from its target rate, which is great. The Fed can set a policy rate, and then it can adjust the balance sheet if liquidity is needed in the system.

I acknowledge that point, but there’s a flip side to that, and that is, well, if that’s true, we can also use the balance sheet for other political purposes and still keep the Fed’s target rate on point. It’s a double-edged sword, and that’s the danger. I think the big point you’re making here is large balance sheets do raise questions about Fed independence, and we need to be very careful where we tread.

It’s not just as simple as a Friedman rule. It’s about the Fed independence, and that’s a cost. You also talked about, effectively, the Fed’s large footprint suffocating market and price discovery, the interbank overnight unsecured lending. The federal funds market is a shell of what it really used to be. We don’t have banks lending, price discovery. I think you mentioned this, that one of the reasons these other central banks are going to a smaller balance sheet approach or demand-driven approach is because they want to resurrect that market, right? Is that fair?

Nelson: That’s right. In fact, I wrote a note that was published—actually the Bank Policy Institute published a blog pointing out that it seems to be the Fed that is unique among the major central banks in terms of not putting value on reviving interbank markets. I included some quotes from other central bankers about how that’s one of their primary objectives.

When my friend, Claudio Borio, was on your show, he said something along the lines of, toward the end of his remarks, “Personally, I think it’s better if banks first go to each other to meet their liquidity needs and then only turn to the central bank as a backstop, rather than having the central bank always be the source of liquidity and replacing the whole interbank function.”

Beckworth: Yes. That’s another cost in terms of the cost calculation. Again, point one is political independence. Point two is a large footprint. We have missing markets. We could talk about the Fed’s involvement in repo markets is really large as well. You could be really cynical—and I’ll be careful here. This is not the objective of the Fed, but you could call it the largest fixed-income hedge fund in the world.

Nelson: Hedge funds hedge their risk. 

Beckworth: Okay, the largest fixed-income fund. Again, that’s not really fair because the Fed’s not striving to do that, but it’s ended up there. Again, political independence, large footprint. I’m going to add a third cautionary reason why we should be worried about a large balance sheet. This comes from my conversation with Raghu Rajan. After reading his articles, listening to him, I’m increasingly convinced that a large balance sheet is a symptom that the fragility of liquidity is much more than we acknowledge or recognize, because he makes this argument.

It’s similar to one you made earlier, that every time we inject the system with QE, reserves go in, banks respond behaviorally, not just mechanically. They see all those highly liquid assets on their balance sheet, and it incentivizes them to fund with short-term liabilities, demandable deposits. They do less term deposit funding. As a result, basically, the net aggregate liquidity is actually quite low because whatever liquidity the Fed injects is being reshuffled out to the banking system in terms of demandable deposits.

The banks are assuming they’re going to have reserves to back up the liquidity demands on them. When it comes time to do QT, okay, the Fed can do QT, but banks can’t quickly unwind their position. You actually end up in a more fragile state. Then what we end up having is it gets larger and larger. In a recent Substack, I call it a liquidity blob. It just keeps getting bigger. QE may actually make us more fragile over the long run.

Nelson: Several things that I want to say, observations on that point. One thing about returning back and unwinding QE, I’d emphasize that when the Fed first engaged in QE, the plan was to return back to doing things the way they were going to be doing. That was the plan for quite a while. There were people on the committee who were actually wanting to go to a floor system. They kept saying, “Well, we don’t really need to make this decision yet. We’re accumulating experience.” Then, eventually, that just became the default of operating in a floor system. Then it was like, “Well, why return to the other way? This way’s working.” Then it was difficult to go back.

When you read the transcripts, Chairman Powell emphasized the reason why it was important to reverse the size of the Fed’s balance sheet was to demonstrate to the public that QE was not a one-way street, that they could do it, and then they would come back, they would reverse it. Ben Bernanke argued before Congress when asked, were they monetizing the debt with this QE? “No, absolutely not.” He said, “We’re not monetizing the debt. If we were monetizing the debt, we would be planning on leaving our balance sheet at these new, higher levels. We’re planning on reversing our operations, so we’re not monetizing the debt. By that definition, the Fed’s left monetizing the debt in the rear-view mirror, basically. I really enjoyed your podcast with Raghu and your Substack pieces. Everyone should read David’s Substack.

Beckworth: Thank you, Bill, for the plug.

BPI’s Bank Treasurers Survey

Nelson: We did a recent survey of reserve bank treasurers about their reasons for their demand for reserves. We mainly did it because it’s very difficult when the Fed conducts these surveys to ask directly about the impact of regulations or examination. The people that are doing these surveys, they’re in the research function or in the markets group, and the supervision examination side don’t like them poaching on their turf. We basically asked, “What leads you to hold reserves? What are the reasons?” We provided a whole bunch of reasons, including complying with regulation and supervision.

To my surprise, and I was impressed by the frankness of our treasurers; the first reason that they gave was actually risk management, meaning holding reserves for liquidity risk management purposes. The second reason was complying with regulations. The third reason was avoiding ever having to borrow from the discount window. And some of them wrote in terms of liquidity risk management purposes, what they meant was avoiding ever having to go to the discount window. There’s definitely scope for fixing regulations and fixing the severe stigma associated with the discount window, and perhaps even daylight credit that contribute to banks.

If you think about it, I’m pretty lazy about keeping track of my finances. I keep a lot of extra cash in my checking account so that I don’t have to worry about overdrafting it. If I wanted to be able to reduce to absolutely zero the possibility that I would ever overdraft, I’d have to hold a huge quantity of cash in my account. That’s basically the calculation that the banks are doing.

Beckworth: That’s interesting. That was an interesting survey that you did. Now, remind me again, how many treasures did you ask? Are they the big banks, or how did it work?

Nelson: BPI’s membership is banks that were $100 billion and above. We include the larger regional banks, as well as the G-SIBs, as well as foreign banks with a large US presence. I don’t remember the actual breakdown. We have about 40 to 50 members, and I don’t remember how many responded, but a good respectable amount. It’s 25, maybe. We asked them, “What determined your demand for reserves? If you’ve reduced reserves, why? What changes would help convince you to hold a smaller quantity of reserves?” That last question was clearly very dear to you and I. The number one approach that they listed was recognizing discount window capacity and liquidity requirements, so definitely key. 

Beckworth: Just to reinforce what we said earlier, right?

Nelson: Yes.

Beckworth: That if we do make the discount window, business as usual, if we can take the collateral and park there and count it toward all kinds of liquidity regulations, it’d make a world of difference, according to your survey and according to the arguments we laid out earlier.

Nelson: Right. No, it definitely would. These things all come together. You can’t really reduce stigma unless borrowing is common. You can’t have borrowing be common unless the Fed returns to a leaner balance sheet. Everyone has the liquidity they need. There’s no clearing at the end of the day for monetary policy to work necessary, so there’s no monetary policy borrowing. After the 2003 revisions to the discount window, I used to fly around to the banks and try to convince them to use the window, and it was working.

That’s something that it’s worth remembering that between 2003 and early 2007, stigma was steadily being chipped away at. Then, when the crisis came, and then particularly after the crisis, when bank CEOs were hauled before Congress and yelled at for having used the discount window, and the Fed didn’t stand up and defend them, basically, then after that point, the stigma became catastrophically bad. We haven’t really ever recovered from that.

Beckworth: As you mentioned earlier, again, those two things, the ample reserve regime and the stigma, actually, they reinforce each other. If you have ample reserves, then you don’t really need to go to the discount window much. If someone does go, it really raises questions, right?

Nelson: Yes.

Beckworth: If you are in a system where there’s scarce reserves, and it’s business as usual, you don’t ask questions—it’s a dynamic system. It’s not clear that you can just fix one without the other. You need to do multiple steps, and hence, my proposal in the Barron’s piece. Again, I’m also realizing that this would take time. Let me ask you that, Bill. If you did something like I outlined or what you’ve outlined, how long of a horizon do you think we would realistically need? Let’s say everyone in the FOMC is on board with proposed changes. How quickly can we get this done?

Nelson: Well, there’s two sides to it. One is that the committee needs to change their mind about how they implement policy, and they may not do that. They haven’t expressed any discontent largely. There is an opportunity now, perhaps, under new leadership, to revisit these things. Typically, when the committee does something, they form a system study group that looks at it.

That group comes and reports back to the committee, and then, oftentimes, the committee will provide them feedback, and then maybe they’ll come back and talk to the committee again, having looked into the feedback, and maybe with a questionnaire this time. Then, everybody agrees on something, but you can’t really count on that. Then, there’s a third meeting where the final change is adopted.

From the monetary policy side of things, it would take a good six months to a year, I think, to turn that. Then, from the liquidity framework side of things, it’s a pretty complex issue, and there are decisions that would have to be worked out about how to calibrate it, what adjustments are appropriate. Those things would take time as well. Even if the banking agencies turn to this issue in the next few months, I think that it would take several months for some kind of change to be made.

Beckworth: Yes. We’re looking at least a couple of years probably to get momentum rolling on something like this.

Nelson: Perhaps. Maybe shorter than that.

Beckworth: Maybe shorter, okay.

How To Shrink the Balance Sheet

Nelson: One of the points that Under Secretary McKernan made was, “Well, let’s not let the best be the enemy of the good.” This particular change is something that we can probably do. The whole framework needs a review. We all recognize that, but this change seems to be something that we could do. Perhaps this year, you could see a change to the liquidity framework, at least along these lines. Then still, it takes time.

My personal set of things that need to happen in order to shrink the balance sheet—I just have three things on my list. There’s a few others that you could probably slip in. One of them is the Fed needs to get a little bit smaller, reduce reserves a little bit further, so that money market rates move up above IORB by a good five to 10 basis points. That gives banks an incentive to economize on reserve balances.

JP Morgan Chase had $400 billion in reserve balances at the end of 2023. At the end of 2025, they have $100 billion worth of reserve balances. They’ve at least found a way to reduce their investment in reserve balances by 75%. The second step is, once you get into that world of somewhat scarcity, then you’re going to see more volatility in response to big changes in reserve demand, or reserve supply. Those things are known well in advance.

It happens on quarter ends, or on tax days when money flows into the Treasury’s General Account, which reduces reserves, or on days when Treasury securities settle, and money again flows into the TGA, at the same time that there’s increased demand for repo. On those known days, known well in advance, you don’t need to get all the way back to daily open market operations. At this point, just avoiding those big potholes would reduce volatility enough to allow that continued pressure to continue.

Then, I think the final step is fixing liquidity requirements, so that examiners and the rules aren’t making banks hold reserves. Even once all of those changes are made, it would just take time for people to change their mindset and get used to it.

Beckworth: Let’s go to one of those points you raised, and that is dealing with the TGA. You’ve written about a reluctance to use temporary open market operations to deal with swings in the TGA. Maybe outline your proposal for what I call neutralizing, or partitioning off the TGA, so that it doesn’t really have a bearing on the rest of the Fed’s balance sheet. Maybe also speak to this reluctance to use temporary open market operations.

Nelson: Sure. I’d be happy to. There was recently an excellent paper written by Burcu Duygan-Bump and her co-author, R. Jay Kahn, about balance sheet issues. They described it as sort of a trilemma, getting between too big of a footprint if you’re giant, to encouraging discount window use, to avoiding daily operations.

I’m not completely sure why that’s something to be avoided once you’ve worried about the footprint. The footprint is actually much worse now. If you look at how the Fed used to do things, they engaged in small operations a few times a week with primary dealers, basically repo operations that changed the quantity of reserve balances. Primary dealers don’t have accounts at the Fed. They don’t have reserve balances.

You were engaging in these small operations with institutions over here. If you’re watching this by video, you can see I’m gesturing to my right. That affected the Fed funds market over here without actually intervening into the federal funds market. It was a very light touch intervention. I’ve been, as you know, as a good reader of my emails, and I appreciate it, that for the weeks running up to the end of QT, every time we came up on one of those pothole moments, known quarter ends, or Treasury coupon payments, I would write a little email saying, “Fed, in a week there’s going to be this moment, and you need to go out there and engage in a $200 billion fixed quantity variable auction determined rate repo in the morning before repo markets really get going to address this. Otherwise, there will be repo market volatility, and then you’re going to point to that repo market volatility as the reason why you have to stop with $3 trillion worth of reserves.”

I did this, and then I did this again. Of course, then there was volatility. They didn’t do it. There was volatility. Then, as you can see in the minutes to the December meeting, there was two reasons why they stopped QT. One of them was money market rates were moving up to the IORB rate, and the other reason was repo market volatility. They could have avoided that by doing these easily foreseen operations.

They don’t have to get them exactly right. These are rare. They can just add a lot of reserves, and just suppress that volatility on those days. That would have left them room to get a lot smaller conceivably, and push on that ratchet to try pushing it back in the other direction. After all, as you mentioned my table, in 2019 when the Fed concluded they needed $1.5 trillion worth of reserves, because this was after the September 2019 repo market volatility, all of the liquidity framework was in place. There was nothing new about it. It had been in place for four years at least, or five years.

Banks could certainly, at least, get back to that level. The percentage of bank assets was one-third lower than it is now. We know that with some incentive that that’s achievable without changing anything.

Beckworth: Yes. It is interesting, as you note, there are these potholes, but there are potholes we can see ahead of us. When we’re driving the car, we can see a pothole, we can veer to the side. You’re saying most potholes the Fed can see. Maybe there’s occasional surprises, but nine times out of 10, the Fed can veer around those potholes if it would do temporary open market operations?

Nelson: At least at this point, where they were still at the edge of abundant. If they were really to go all the way back to a corridor system, they would need to operate more frequently, but that really wasn’t that hard. There would be a handful of people in DC and a handful of people in New York who did the forecasting, and figured out these operations. Then, you talk to the banks, ask them what they were going to be doing today, and you just engaged in the operations.

I think that the reason why it’s perceived as so difficult to go back to that regime, partly, is a loss of institutional memory. It seems more difficult because people don’t do it that way anymore. It’s partly because I think it’s just unfamiliar, but also partly because when you talk to people about what framework the Fed should be using, you typically are oversampling. The people who actually work on the liquidity details of central banking, or the market participants who are right on the other side of it, they see this all as a big use of resources. But it’s really a finite number of people that are engaged in this.

Beckworth: To be clear, even if you went to something like scarce reserves, or what I would prefer to call a state-contingent operating system that can become anything depending on the circumstance, you would have robust ceiling and floor facilities, right? I mean, you would have discount windows that people would feel comfortable going to, so you would never have rates pop above them, and you’d have a robust floor, the interest on reserves, or if the overnight reverse repo was needed.

You could still control rates within that range, but you would have variability within that range such that there would be some price discovery, some market activity. That is the story we’re hearing from other central banks.

Nelson: Right.

Beckworth: Some of them have narrow ranges. I mean, the Riksbank, which I mentioned, they have 10 basis points spread between the target rate and the interest on reserves, and 10 basis points above, so 20 total. I think you would want a little bit larger, right, for the Fed?

Nelson: I would eventually. I mean, the Bank of England’s corridor was zero, right? That doesn’t mean it has to stay at zero. You might want to have a narrow corridor as you are getting everyone used to the new arrangement. I realized that I didn’t really respond to your question about the TGA proposal, and the Fed holding a portfolio of maybe repos, and that Vissing-Jørgensen has suggested T-bills, but I think repos are a little bit more flexible, in the same size as the Treasury’s deposit.

As the Treasury’s deposit goes up and down, if you hold assets constant, reserve balances go down and up by exactly the same amount because balance sheets must balance. But if you have a pool of assets that goes up and down with the TGA, then reserve balances don’t change. I’ve got to tell you something, maybe just between you and I, which is that, it’s really just a way to dress up temporary open market operations, so that perhaps they are more palatable to people.

I pointed to three possible reasons why people don’t like these—or maybe four. One of them is this misunderstanding that, that kind of daily open market operations were equivalent to the fine-tuning that, say, Milton Friedman was the first to object against. There’s just a certain sort of economist who objects to the idea of the Fed fine-tuning the economy. Milton Friedman thought actually that kind of fine-tuning was exactly how his constant money growth rule would be satisfied.

He points to open market operations, and frequent open market operations as exactly how the Fed should be implementing policy. That doesn’t really work. I forget the other reasons I speculated about. I think that one of them, probably the most important reason is to those who have been advocating a floor system, they basically portrayed it as this kind of set-it-and-forget-it monetary policy implementation regime.

Even engaging in an operation now and then to avoid those big potholes would be seen as admitting that they were wrong. It’s very difficult for a big institution to admit that it’s wrong. Calling it not an open market operation, but rather matching the TGA—

Beckworth: Oh, I see. 

Nelson: —actually, you’d be using open market operations to do exactly what you would be doing, because almost all of these variations in reserves are driven by, a lot of them, certainly the big ones, are driven by changes in the Treasury’s General Account. You’d really just be doing temporary open market operations. It’s a nice way to package it, so it doesn’t feel as bad.

The Fed’s Profits

Beckworth: Bill, one more thing I want to bring up about the Fed’s balance sheet and its operating system—earlier in the show, I outlined the size and how far we’ve come. Something else that happened recently is the Fed began to generate profit. Now, I’m not sure how steady or sustainable it is, because we saw it go from operating losses to operating income, and then back to operating losses. I haven’t looked recently, but it looks like we’ve bottomed out at least. The road ahead looks great for the Fed. They’ll be generating net income on their interest, net income versus the expenses they face. What do you think about that? What do you think about the Fed’s operating losses now coming to an end?

Nelson: Well, it looked like that they had returned to profitability in the fourth quarter. So far in the first quarter, it looks like they’ve returned to losses, but were probably bouncing around this point where they’re at the edge of returning to profitability. There’ll be some up quarters and some down quarters. We won’t really know until more data is released. 

I’ve written a lot on these losses. Of course, the Fed made about $200 billion in losses, but that doesn’t even take into account the fact that, since they have the royal franchise on creating currency, as well as having the Treasury’s general account, which is about $1 trillion, and doesn’t bear interest, that makes the Fed inherently profitable. They’ve actually had losses that have burned through all of that inherent profitability to get to the point of having operating losses.

When you add all that up, you come up with a number of more like $500 billion. They’ve lost a tremendous amount of money. I’ve always been very careful to say, it’s not necessarily appropriate to blame the Fed for this yet, because the point of QE is to take on interest rate risk. It’s not to create reserves. That’s actually an unfortunate side effect. The point of QE is to take on interest rate risk to push down term premia.

They took on interest rate risk, and they paid the piper when they had to raise interest rates very sharply. Wait until the 2020 transcripts have been released, because the real question is did they take that risk carefully? Did they understand the risk that they were taking? Did they evaluate the pros and cons, as they had done in 2010, 2012, 2016? 

Well, now the transcripts have been released. Unfortunately, it looks like they didn’t consider the interest rate risk, and the risk to their profitability at all. It didn’t enter into the discussion. Even the very strong forward guidance that they issued on their balance sheet, and on the Fed funds rate, guidance that was strangely lacking in the escape hatches, and that was present in the Global Financial Crisis forward guidance, you have to ask, “Well, why didn’t they put in those escape hatches that would have allowed them to start tightening sooner, and to the balance sheet accumulation sooner?” They weren’t even discussed.

When they were discussing the new forward guidance, there was very little to no reference back to how it had been done before, at least at the point of the staff. There were some governors that expressed some concerns when you look at the discussion about the rigidity of that guidance, but none really about the interest rate risk per se. That’s an additional reason why a large balance sheet, or a floor system approach is costly.

I think it encourages a certain sloppiness in thinking on the part of the central bank about using their balance sheet, as can be seen in their willingness to take on such a vast amount of interest rate risk. Admittedly, and very initially, very trying times, right? When COVID hit, we all thought we were going into the next Great Depression. It seemed inevitable, and nobody foresaw how quickly things would recover. Toward the end of 2020, the situation was returning, and the fact that the Fed was continuing on to accumulate assets at a gangbuster pace was hard to make sense of.

Beckworth: That is the key point I like to make, is I want to have QE as a tool that is available if we are on the cusp of another Great Depression. If we are at the zero lower bound, we need that tool, but that tool, so far, has only led to a ratchet effect. Can we find a better way to implement it? I think the answer is, yes, but we need to sit down and have an honest conversation. What have we seen over the past QE experiences—QE1, QE2, QE3, and I’ll call QE4 with the COVID experience? We should sit down and do a review like these other central banks. 

This is why, again, I really think something like a term deposit facility, which you would set at the target rates, at IOR below that, at a discount rate above it, would really create the right incentives, and I’ll encourage listeners to go read my article about this. We want a robust system, one that allows us to go to an ample reserve system when it’s prudent to and wise to, and then also one that returns to a scarce reserve when it’s time to do that as well.

Otherwise, we’re just going to get a larger and larger and larger balance sheet, and if there’s no cost to that, then let’s just blow it up to infinity. There’s got to be costs. There are costs. We’ve outlined them. I think it’s good that we have these conversations, for sure.

Nelson: The committee agrees that there’s costs, that when you go back and look at the minutes, or read the transcripts associated with their decision to adopt a floor system, which they made in January 2019, but it was really based on, as we’ve discussed, two meetings of discussions, and staff analysis at the last two meetings of 2018, they indicated that the cost of operating in a floor system as they saw it was having to be so big, and the challenge of determining how small that they could get back to. They didn’t seem to disagree with the proposition that having a big balance sheet is part of the cost, if only the perceived political cost.

Beckworth:  Well, Bill, those have been great observations and great words of wisdom for us to take to heart as we move forward. Thank you, once again, for joining us. Our guest today has been Bill Nelson. Bill, thank you for coming on.

Nelson: It’s a real pleasure. Thanks for having me.

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.