Bill Nelson on the Fed’s Discount Window Lending, the Overnight Reverse Repo Facility, and the Shifting Size of the Fed’s Balance Sheet

Bill Nelson is a chief economist and executive vice president of the Bank Policy Institute and was previously 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. He also worked closely with the BIS working groups on the design of liquidity regulations and is a previous guest of the podcast. Bill rejoins Macro Musings to talk about the Fed’s balance sheet, and in particular, the impact that the Fed’s response to the recent banking turmoil has had on its size, as well as the role being played by the Overnight Reverse Repo Facility. David and Bill also discuss the changes in collateral treatment brought about by the banking crisis, the invocation of 13(3) for the Bank Term Funding Program, the recent volume of discount window lending, and a lot more.

Read the full episode transcript:

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

David Beckworth: Bill, welcome back to the show.

Bill Nelson: Hi David. Thanks for having me back.

Beckworth: Well, it's great to have you on and I have been looking forward to this conversation because the discount window has lit up. It's amazing how big it's become and you are the expert I go to on discount window lending, and we'll come back to that in a bit. But I thought we would first just take stock of where we are in terms of the Fed's balance sheet, what progress had been made on quantitative tightening, and then of course in March things kind of came to a stop and reversed a little bit. So Bill, I'm going to provide just a few numbers here on the Fed's balance sheet, and if you want to add a comment as I go through it then I'll have some questions for you. But just, again, paint a picture of where we are in terms of the Fed's balance sheet. Let me begin by noting that at the peak, the Fed's balance sheet reached about 8.97 trillion in the spring of 2022, and then in March of this year it had fallen down to 8.3 trillion as a result of quantitative tightening.

Beckworth: So as you recall, the plan that the FOMC had announced, so starting in May 2020, they made this announcement, had been in the works for a few months, but May 2022, they began the reduction of the Fed's balance sheet. And it ultimately ended up as a $60 billion reduction per month in treasuries, 35 billion per month in agencies, debt and mortgage backed securities for a total of 95 trillion per month. The committee also had this to say, they said they, “intend to slow down and then stop the decline in the size of the balance sheet when reserve balances are somewhat above a level it judges to be consistent with ample reserves.” So kind of a side note, we're in this ample reserve system and listeners go back and check out past shows with Bill and myself, we've discussed this extensively. It's a floor system, Bill and I both prefer a corridor system which would imply a much smaller, leaner balance sheet.

Beckworth: And Bill, we've talked about the interaction of Fed policy and regulation and how that's made it more challenging for the Fed and for bank examiners, supervisors to get there. So we've had this plan in place, as I mentioned the balance sheet got down to 8.3 trillion in March of this year. Now we're back up to 8.7 trillion, so we've expanded a little bit, largely through discount window lending, which I want to get into with you. And also they've laid out some markers, some destinations, and I'm going to quote here Governor Chris Waller. Back in January, Bill, he was at a Council on Foreign Relations event, and he said the following, he would like to bring the reserves down to eight to 10% of GDP. Now he counted reserves and overnight reverse repos together because he thought they'd be fungible as they begin to drain things down. And we'll talk about this later because I know you have some criticism of the facility which makes maybe this assumption a little too simple. But his argument is, if you combine those two and divide them as a percentage of GDP, we want to get it down to eight to 10%, which is where it was, I guess, right before this whole balance sheet expansion began. The balance sheet is around 20% and it continues to grow as we see the Fed's facilities being tapped. Now, any comments, Bill, on the size and where we've been and progress we've made and what we've maybe lost over the past few weeks?

Updates on the Fed’s Balance Sheet and Quantitative Tightening

Nelson: No, I mean I think it's probably worth noting that... So maybe the answer's yes, I do have a comment.

Beckworth: Okay.

Nelson: So from the perspective of I guess a typical ordinary Fed economist, the tightening part of quantitative tightening is not the shrinking of the balance sheet, but rather the reduction in the Fed's holdings of longer term securities. So those have proceeded at pace. So even though it's true that the balance sheet itself has reversed course and increased because of these other expansionary activities, the lending in particular, I don't think that that's reverse quantitative tightening.

Beckworth: Okay. So QT has continued the pace, but on the side we have these facilities adding liabilities to the Fed's balance sheet. And in fact, Governor Chris Waller, at the same event I just mentioned, he noted that even if the Fed had to cut rates… in his view they should continue doing quantitative tightening because it's a signal and they've made a commitment to it. So he argued you could do both things at once, which is I know not maybe normal, but does that make sense? Should you continue QT even if the Fed had to cut interest rates because the economy is beginning to slow down?

Nelson: I think so. I mean, if you think of quantitative easing as being a policy that you pursue when you're at the zero lower bound and you want to ease further, you could certainly continue it even though you're cutting the federal funds rate as long as you were able to achieve the degree of monetary stimulus that you were aiming for by adjusting the federal funds rate itself. [inaudible] communication challenges.

Beckworth: And just to put some numbers on the size of the Fed's liabilities, the ones that we're going to be talking about today, the Fed's reserves stand just over 3 trillion, 3.4 trillion and the Overnight Reverse Repo Facility at 2.63. So between five and 6 trillion, we have these liabilities that pay interest. They're short term. They're overnight. And we'll come back to this in a minute when we talk about the design of the Overnight Reverse Repo Facility with Bill and some of his concerns surrounding that. Okay, so that's kind of the state of where we are. We're still at a very large balance sheet, and just recently the Fed released an audited report of its financial statement and it noted that it had experienced just over a trillion dollars in unrealized losses. And that was approximately 600 billion in treasuries, 400 billion in agency securities.

Beckworth: So in both of those it took a loss and also in the last quarter it experienced an operating loss in terms of income. Over the whole year it was still positive, but it was negative and of course this year, it's definitely negative income earnings. And this has been interesting to follow, Bill, you and Andy Levin had a nice little policy brief for us where you noted these losses are taking place and they are ultimately borne by the taxpayer. And even though it doesn't affect the conduct of monetary policy currently, it's still something that one would think about in terms of should taxpayers be aware that this is happening.

Nelson: Right, I agree. I think the first point, I mean it's important to emphasize that these losses really should not interfere with the ability of the Fed to conduct monetary policy. They're not going to lead to a hyperinflation or anything along those lines. On the other hand, they are really losses that are borne by taxpayers and in some sense it's not so much the dollar loss itself, they're losing about 2 billion a week, as it is the comparison between that loss and the amount that they should be making simply by the fact that they have the royal monopoly over the creation of currency, which is a wonderful liability that doesn't mature and pays no interest. So there's quite a bit of money that they would normally earn simply by being the issuer of currency. So the relevant comparison is how much they would make just by doing that and how much they're actually losing.

Nelson: But I'd add, in some sense the relevant cost to be considered when deciding to engage in QE or not to engage in QE is the risk. And these losses are simply a result of having taken that risk. And risk can go in the other direction too. The Fed engaged in QE, of course, in the mid 2010s and interest rates ended up much lower than anyone expected. And so they made money, but again, I mean the Fed itself in a very comprehensive review of how it should conduct policy and what assets it should buy that I think was put out in about 2000, well worth reading if that's the kind of thing that you're interested in, they recognize that interest rate risk was very similar to credit risk, and that if the Fed were to engage in that risk it should be cognizant of doing so and should consult with Congress before doing so.

Beckworth: So I've talked with Andy Levin on the podcast before about this policy brief, this issue. I've had your former colleague Bill English on the show as well, and he's also taking a different view than you and Andy do. And his argument is, look, yeah, there's some losses now, they don't meaningfully affect monetary policy. And yes, there could be some real resource cost to taxpayers in terms of foregone… your remittances to the Treasury, bigger deficits, maybe even a closer date X for the debt ceiling crisis, should it happen. But the flip side of all that is we were able to have an amazing recovery, we're able to bring the economy quicker to full employment, maybe debt to GDP is lower than it otherwise would be because we got the economy going. And I know you have responded to that comment, let me throw another critique out that I got on Twitter.

Beckworth: I got some pushback talking about your views with folks on Twitter, and one of the observations that was made is that, do the counterfactuals follow? Assume the Fed keeps its interest rate path as it did, so it keeps it low for a long, long time. It doesn't begin to tighten until last year, until early last year, but assume that it ends QE much sooner, say 2021, which is I think what you and Andy were hoping for, QE ends much sooner. Is that fair? And maybe you would've also advocated rate hikes sooner, but at a minimum you wanted QE to end much sooner, correct?

Rate Hikes, QE, and Balance Sheet Counterfactuals

Nelson: Yes, so both are correct. So in our analysis, Andy and I separated out the asset purchases that took place in the second half of March, and in the first half of April, precisely because I think that the case for those purchases is very strong in the teeth of the Treasury market itself collapsing. Certainly in real time, I was strongly trying to encourage the Fed not to... I mean they seemed to kind of go into autopilot, they established a certain pattern of buying securities, they changed their explanation from we're buying them for market functioning purposes to stimulating the economy purposes, and then they folded it into a flow-based asset purchase program. A program where they bought the same amount, they continued to buy the same amount and promised to do so until certain macroeconomic requirements were met.

Nelson: And I was all along trying to personally encourage them not to do that, partly because I think that the flow-based asset purchase program that the Fed conducted in 2013/14 with QE3 was a big reason how they ended up with such a giant balance sheet in the first place, how they ended up stuck in a floor system. And as you know and as you noted a year ago or a year and a half ago, I was pointing out that the Fed had a very long way to go to get to sufficiently tight monetary policy. There was a time when they were talking about going to two and a half when inflation was 8%, which is not very much in line with the way most macroeconomists would view it. So it always seemed like they had a very long way to go, and that's the first time that I actually did some back of the envelope projections of the amount of money that they would be losing on the view that they'd actually have to go up to about 6%. So a bit higher than they are currently. We'll see where things end up.

Beckworth: So you've definitely advocated both rate hikes sooner as well as ending QE sooner, and I'm sympathetic to that view. I agree for all the reasons you just mentioned, in addition to the bigger goal of shrinking the Fed's balance sheet. One day in the distant future returning to a corridor operating system, Bill, that'll be our never ending quest probably in our lifetime, but do a scenario with me, a counterfactual with me here where the Fed does end QE when you wanted it to. So in other words, we definitely allow the Fed to do its market back stopping in March and April, but then it begins to unwind and end the asset purchase, but then it keeps its rates low for the period that it actually did. So in this counterfactual world rates are low through early last year as they were, but QE ends much sooner, alright? And also I'll make the strong assumption I think you're making too, implicitly, is something like Wallace neutrality is applying here, that the QE purchases really didn't have much bang for the buck.

Beckworth: That they really weren't making that much difference for the economy. So if I put all of those markers down, here's an argument one could make, and I know you'll probably push back against it, but let me throw it out there anyways. All the losses that were now seen on mark to market on bonds, so treasuries, corporates, all the mortgage backed securities, they're all taking big hits. In fact, there was a study out by some group of academics that estimate there's close to $2 trillion of unrealized losses on bank balance sheets and all of that could have been far larger had the Fed not done QE all the way to the point that it did. So in other words, had the Fed ended QE sooner, there would be more bonds out there and some of them on the Fed's balance sheet, and therefore banks would be under even more strain. So in other words, the Fed took off a lot of duration risk, interest rate risk, it's bearing their risk right now on its losses. It could have been worse if it had been on bank balance sheets. What would you say to that counterfactual?

Nelson: Well, I think that's a very insightful observation. So first of all, it's not clear to me that the Treasury would've kept its behavior unchanged under those circumstances, but it might have. So what the Fed did was from the perspective of the consolidated debt of the government, it basically unraveled the treasuries having issued longer term debt and replaced that with overnight debt. That's the sense in which this is a real cost to taxpayers, even from the consolidated federal balance sheet. It's not so much a cost as it is a gain that taxpayers didn't receive because they aren't now benefiting from super low interest rates on the federal debt. Instead, the Fed has converted that into longer term debt, and so our consolidated balance sheet costs are higher. And so I guess one way of thinking about this is, well, if taxpayers weren't bearing those losses, it would be financial institutions and others that are bearing the losses. And then that gets into, well, who would be holding all of these longer term securities? And some of it might have been banks, it would've been other institutions, a lot of it is held abroad. So I agree with you that had QE ended sooner taxpayers would've been better off and other financial institutions would be bearing these costs but I'm just not sure where exactly it would've ended up to be honest.

Beckworth: So the distribution of who was holding these securities also, how big the effect would've been, maybe it wouldn't have been that important. So your question is whether maybe if the Fed had done this other path, maybe the mark to market losses would've been smaller?

Nelson: Well, that's another issue. I'm just wondering whether you set this up as losses for banks, and even if some of the losses would've been for banks, I'm pretty certain that there would've been losses made by all kinds of different institutions, individuals, foreign governments, foreign sovereign wealth funds.

Beckworth: That's a good point. So who would've been holding the bonds if not the Fed?

Nelson: That's right. And that's the question I don't know the answer to, but I'm pretty sure that only a relatively modest fraction of it would be banks.

Beckworth: Yeah, fair point. Okay, well let's move on to the current developments that have affected the banks' balance sheet. And again, as you noted earlier, QT continues, the Fed is continuing to roll off securities off its balance sheet, but because of this assistance in the banking turmoil that we've seen in March, it has expanded its balance sheet. And I mentioned it's back up to approximately 8.7 trillion from a low of 8.3 trillion, and we almost hit 9 trillion in the spring of 2020. So we're slowly working our way back up to that 9 trillion number and that number's big, I mean historical perspective, it seems big, and I know it keeps you up at night, Bill. You dream of small balance sheets and lately you've been dreaming of horrible big balance sheets at the Fed. So we're going to try to help you rest better by working through this issue.

Beckworth: So the Fed has done three things that has contributed to their growth of the balance sheet, and it's all under, I guess, discount window lending broadly speaking, but there's been three things that have happened. The primary credit or the discount window lending, and let me just throw the most recent numbers out. The primary credit was at, if I'm reading this right, 88 billion. The Bank Term Funding Program was at 64 billion. And then other credit extensions, so this would be lending to the FDIC, to banks, they're being sorted through in this process, 180 billion, so total around $340 billion through Fed lending. Now, one question I have Bill, and we're going to get into the particulars of each of those different programs, but have we removed the stigma of discount window lending? I mean, it was pretty shocking to see these numbers when they first came out, but maybe I shouldn't be surprised.

Beckworth: In fact, I went back and I looked at a chart during 2008. So at a peak, let me restate this. At a peak, primary credit hit 104 billion, a few weeks ago. In 2008 it was 111, so it was pretty similar in terms of the peak. So we tapped out, of course the economy's larger, so we probably needed to contextualize this as a percentage GDP or total bank assets or something, but definitely much larger than 2020. 2020 was only around 49 billion. So this was, at least it seems pretty large, the amount that's being tapped into through discount window lending. Is it large in historical context? And second, do you think it's made a dent in terms of the stigma attached to that type of lending?

Evaluating Discount Window Lending

Nelson: So it is large, I think it's approaching the lending that took place in response to COVID, if I'm remembering that correctly. I think you have to look at this and say, this suggests that the stigma was perhaps less bad than we'd feared. The issue is what lending would you see if there was low stigma? It's not obvious to me that these are the numbers that you'd see, maybe you'd see more. In some sense, basically, the discount window acts like a sort of safety valve to release pressure off of the financial system when the system is under pressure, be it end of day monetary policy pressure or the kind of pressure that we've seen in recent weeks. And you want it to be a safety valve that releases pressure at a relatively modest level so that you don't have to have extraordinary strain on the financial system before people are willing to go in and borrow. So the question then becomes, well, were people willing to borrow at a level that seemed reasonable or was there so much pressure out there that had actually brought them into borrowing? And as best I can tell, it's a very hard thing to know. It kind of looks like it worked, like it was supposed to work, which is a credit to I'm sure a tremendous amount of very active work on the part of the Federal Reserve system out there, in the moment, encouraging banks to tap the window and as well as banks' decision that it was the appropriate thing to do.

Beckworth: Now, prior to the banking turmoil, there was an uptick in discount window lending already. Isn't that right?

Nelson: That's right. It had grown up to a few billion.

Beckworth: What was the story behind that?

Nelson: So it wasn't completely clear, honestly. The reserve balances had been declining of course, and it could have been just a general reflection of some building tightness in markets, could have been strains on banks that we now saw in full bloom. But I don't think that the answer is really... We'll know in two years when the individual borrowing data is released, but at this point, I really don't know what that was other than, borrowing is normally something that is a normal activity. There are times when the discount window is an emergency kind of thing that you do. You're tapping your contingency line of credit, but most of the time, most borrowing is perfectly ordinary. It's a smaller bank, had a municipal deposit ended, it needed to get additional funds, it went in and borrowed for a little while. It's meant to be there for banks to use as a normal course of business activity when it's right for them to use it at an above market rate. And it could have just been those things, it's just not clear.

Beckworth: So you said there's a two-year window before we find out who actually used it and how much, is that right?

Nelson: That's right.

Beckworth: Okay.

Nelson: Well, for 13(3) lending, which would include the Fed's new Term Bank Funding Program, it's one year, but for regular 10B lending, which would include primary credit, it's two years.

Beckworth: Well, let's go to that, because that's a very interesting question. Why was 13(3), the emergency part of the Federal Reserve Act, invoked for this new facility, this Bank Term Funding Program? Why do that versus 10B? Because the way it reports on the Fed's balance sheet, it looks like it's all under the discount window lending umbrella, but it's being invoked through 13(3), which in my mind, I think of money market funds, liquidity facility, commercial paper facilities, things like that. So help me understand, you are a former director of this whole patchwork of liquidity facilities. So help me understand, Bill, what's going on here?

The Bank Term Funding Program and 13(3)

Nelson: I think alphabet soup was the term that you-

Beckworth: Alphabet soup, yes.

Nelson: So I don't recall exactly, but I think that under loans on the statistical release, all of those different types of loans appear. Normally, the Fed offers three types of lending, seasonal credit, which is funding for small banks that experience seasonal patterns to their deposits and loans. So it's basically designed so that they don't have to hold a lot of extra liquidity on their balance sheet to cover those swings. Instead, they can lend to their communities. So it's actually really a demonstration of what we've talked about in past episodes that if the Fed were to recognize the liquidity that comes from the discount window, then banks would be able to lend more to the real side. But anyway, that's seasonal credit.

Nelson: Primary credit, which is the program that saw a lot of borrowing, that's for depository institutions, as are all of these things, that are in financially sound condition. And it's pretty loose criteria. It's CAMELS 3 rated, which means you're okay, but you're not really fantastic or adequately capitalized, which is again, sort of we're not worried about you, but you're not in fantastic shape. But it's meant to be relatively accessible. It's extended at an above market rate on a no questions asked basis and it was designed to be something that banks judge that they could rely upon to use when they needed it in an effort to reduce stigma. Right now, the rate is relatively low compared to its history, so it's at 5% just at the top of the Fed's target range for the federal funds rate, so still slightly elevated, but not that elevated.

Nelson: If you don't qualify for primary credit, then the Fed can give you secondary credit, which is for troubled institutions. We haven't seen any secondary credit, I don't believe. Now in response to the SVB turmoil, the Fed opened two new programs. So one was the Bank Term Funding Program, and under that program they're offering one-year loans to banks at the fixed rate of the one-year OIS swap rate plus 10 basis points, and the loans are repayable without penalty. And what makes them very different is that they're lent against the full par value of OMO securities, so Treasury securities, agency securities. OMO is shorthand for open market operation. It's the securities that the Fed can buy. And so since some of those securities are actually selling at 80 cents on the dollar of par, the Fed, if it lent the full amount of the collateral pledged against those securities would be lending 80% secured and 20% unsecured.

Nelson: And for that extra amount, the Treasury has committed $25 billion of money from the Exchange Stabilization Fund, which is a fund that they often creatively use for this kind of purpose to provide the Fed credit protection. Now, as we were discussing, although the Fed's normal lending authority is under Section 10B of the Federal Reserve Act, and that would cover primary credit, seasonal credit, and secondary credit, this program is extended under Section 13(3) of the Federal Reserve Act, the section that became quite famous during the global financial crisis because it was authority the Fed hadn't really used since the Depression, but they used with abandon in the global financial crisis when I was involved and then again in response to the COVID strains. And that's generally the Fed's authority to lend to non-banks, and it has a lot of rules and conditions associated with it. It has to be broad-based, it has to be adequately secured to protect taxpayers. It can't be used to help out a failing institution. A lot of these rules were put on after the global financial crisis, given the backlash against lending to help out AIG and other institutions like that. So the interesting question is, well, why did they use 13(3) rather than 10B? [It's] at least interesting to the plumbing geeks like you and me.

Beckworth: And listeners of the show.

Nelson: …To make these loans? And my guess is that they did so because they are one-year loans and loans made under the regular authority can only extend for 90 days. So they couldn't have offered these loans under 10B. It could be that it has something to do with that credit protection, but I'm less sure about that.

Beckworth: Well, how would we find that out for sure? Would someone need to ask that question to Jay Powell at an FOMC press conference? Will they release some document in the future that says, this is why we did 13(3) versus 10B? How would we eventually know the truth?

Nelson: So I guess somebody would have to ask. We know it's 13(3) because there's a reporting requirement. The Fed, within a few days, wrote a letter to Congress explaining what it had done, although I don't believe the letter explains why it did it.

Beckworth: Justification, yeah.

Nelson: So to me, the more interesting question is, and these were all interesting questions, was under what legal authority did they lend to the bridge banks? So the FDIC closed down SVB and Silvergate and then reopened them as bridge banks, which contained the liabilities of the institution, many of the assets, I think. And then the Fed lent, as you noted, currently $180 billion-

Beckworth: Yeah, it's huge.

Nelson: …To bridge banks. So why they did it and what's going on there is an interesting question, but staying in the weeds for just a moment, there are legal sort of restrictions on the Fed's ability to lend under its normal authority to undercapitalized institutions. They were put in place under the FDIC Improvement Act in 1991 because people thought that the Fed had maybe raised cost to the FDIC by lending to troubled institutions. So there are these rules in place. It's not clear to me that a bridge bank is in fact adequately capitalized. I don't actually know if it has any capital and how to view that. So then it raises the question of, do these loans fall under those [inaudible] restrictions? That's unknown. We know that the loans weren't made under 13(3) because the Fed would've had to report it.

Nelson: So one possibility is that the loans were actually extended under Section 13(13), the little known Section 13(13), which is effectively an unrestricted ability of the Fed to make loans against OMO collateral, against Treasury and agency collateral. So if that was the collateral, and that would probably include sort of a promise by the FDIC to pay it back. It's a government agency. So could have been a lot of creative legal thinking here. As far as I know the Fed has not lent to a bridge bank before, but it's hard to prove a negative and I could definitely could be wrong about that. And it could be that they used this little known authority to make it. They haven't answered that question, and I don't know if they ever will.

Beckworth: Well, this is fascinating. We've covered 13(3), 13(13), 10B of the Federal Reserve Act. So this is really great stuff. Now I want to go back to the Bank Term Funding Program and I know that things get a little hyperbolic on Twitter, a little exaggerated on Twitter, but all of us were just losing our minds when we read that part about the collateral requirement where a bank could give collateral and receive its face value, its par value, and that's not normal. Typically there's a haircut, there's a discount. I mean, the history of central banking, this is very different. This goes against the grain all the way back to Badgett. I mean, this seems very different. And some people were saying, “this is a revolutionary change in how central banks will operate going forward.” Now it may be just a bump in the road and we forget about it in a few years, but maybe it does open the door to the central banks of the world and advanced economies being creative and occasionally taking collateral at face value even though it may be 80 cents on the dollar. So were you shocked by this and should we be shocked by this?

Changes in Collateral Rules and the Standing Repo Facility

Nelson: So I was a bit shocked by it. So I would lose my finance professor credentials if I believed that there was anything particularly magical about par value. What matters, of course, to economists like you and I, is market value. That's the value of the collateral after all. Normally, the Fed seeks to take collateral against its regular discount window loans. It seeks to take all bankable assets. It wants to be in the business of lending against books of loans and other illiquid securities so that its lending adds liquidity to the market. It doesn't generally lend that much against securities, but when it lends against loans and when it lends against securities, it does so against an estimate of the market value of those assets. And then it applies a haircut. And the haircut is meant to cover the risk that it would take a while for the Fed to liquidate the collateral and how prices might move against them so that it doesn't make a loss.

Nelson: And in fact, it's also worth noting that up until a few days after SVB, the Fed has always applied a haircut to government securities, to Treasury securities and agency MBS. The haircut is only 1% against a T-Bill, but it rises up, I think it was four or 5% against a 10-year note. But they set those haircuts to zero as well, but again, still just against the market value. But nevertheless, that's pretty bold. And then here you have the Bank Term Funding Program, which is lending against the full par amount. So basically they're lending unsecured. There's no particular reason why it had to be the par amount, it could have been, we're going to lend 25% more than the market value of the securities. And that is, I think, extraordinary and what I took from it was that the Fed really was concerned about institutions having enough collateral to be able to get the liquidity that they needed over those few days to meet outflows. What I saw was the Fed really straining to be able to provide the most credit possible to institutions in the event that there were further runs. And so that's what I brought from it. The extraordinary nature of it led me to conclude that what the Fed was seeing… really concerned there.

Beckworth: So when I look at this, it reminds me of the Standing Repo Facility and some of the issues they've been encountering and getting people to join on board, become a part of that facility. And because there's haircuts there as well, they're trying to get more participation. And the spirit of it is, this is a facility that you should be able to go to on a regular basis, not just in panics. And it seems to me that this Bank Term Funding Program is kind of in the same spirit. It's like, we'll take your treasuries and we'll give you the reserves you need. And they made it super easy. They made it at collateral at par value. And I’m just wondering, what does this mean for something like the Standing Repo Facility? Did we see any test of it during this moment or did the generous terms or the flexibility of the Bank Term Funding Program make it less relevant?

Nelson: So I don't believe we saw any use of the Standing Repo Facility. And that's probably not surprising, because viewed from the perspective of providing credit to depository institutions, commercial banks, the Standing Repo Facility doesn't actually add anything. So all of the securities that you can bring to the Standing Repo Facility to convert into reserve balances are securities that the discount window takes. And the rates on the two are the same, and you can borrow only overnight at the Standing Repo Facility, but you can borrow up to 90 days at the discount window. And this is just the ordinary discount window. And the Standing Repo Facility just provides credit once at 1:30 in the afternoon, whereas the discount window is open until after the close of Fedwire, in fact. So it's the perfect sort of late day provision of liquidity. So I mean, what the Standing Repo Facility did was, on the one hand, it was a standing lending facility for primary dealers, right? And the other was that it was meant to have a different sort of look and feel of the discount window in hopes that it would reduce stigma, but nobody's really signed up for it, or very few regional banks have signed up for it. One has, and so it hasn't really taken off much at all.

Beckworth: So you mentioned it's an overnight facility, the regular discount window lending is 90 days and bank term funding is a year. Is that right? This new facility?

Nelson: Yeah, that's right.

Beckworth: So I guess if I'm running the Standing Repo Facility, I'm saying, come on man, you make this other facility so much more attractive than mine. You get a year term on it. You take collateral at par value, anyone can come up to it, and at mine you got to become a member, you got to join, you got to be onboarded. I don't know, it just struck me as they accomplished something similar to what they were trying to do with a Standing Repo Facility, but maybe I'm misreading it.

Nelson: Well, again, I think that the Standing Repo Facility never offered any advantage. It never did anything that the discount window didn't already do. And so it was only about optics, in terms of providing credit to depository institutions. And I think that the Fed, in my view, they sort of dropped the ball on the optics part of it because they never went out there and went out and said, "We really want you to use this. This is something that you should sign up for. This is something that we would feel comfortable with you using and feel good about you using." As far as I understand it, banks weren't even allowed to say to their examiners when the examiners said, "Well, how are you going to convert these securities into cash?" If the banks’… OMO securities, treasuries, and agency MBS, if the bank had signed up for the Standing Repo Facility, my understanding is that they weren't allowed to say, we'll use the Standing Repo Facility.

Nelson: So if you're a regional bank, to sign up for the facility, you have to arrange to get access to the tri-party repo platform. If you don't normally use it, that's an expense. You generally wouldn't be using it because that's not the line of business that you're in. And you can't even tell your examiners, you can't even say, this is how I'm liquefying my assets, even though that's its purpose. So there really wasn't much done in the way of follow through on the part of the Fed to do what was necessary to reduce the stigma associated with it. But I think, again, all of those things could also be done and should be done with the discount window, which takes not only those securities as collateral, but also all kinds of other things as collateral and so is a much more effective tool.

Beckworth: And to be fair, I recognize the Bank Term Funding Program is a temporary emergency facility. It's not going to probably be around forever. But I think your bigger point is, well, all of these things really should be done through the discount window if it were set up properly, if the stigma wasn't there, and it could be a one stop shopping place if it were working the right way.

Nelson: That's right, except then you get into the issue of lending to primary dealers or other broker dealers and issues of Treasury market functioning. And I think that a lot of the folks that were advocating for the Standing Repo Facility were doing so because they were very concerned about September 2019, about the repo turmoil that took place there. And they wanted participants in the repo market in the morning to know that they had access to this facility at 1:30. If they needed it, they would have funding. And it ended up then sort of becoming a facility for commercial banks because within the Federal Reserve system, there are 11 districts that don't have a repo market and only one that does. So if you're going to offer a new lending program, you need to offer it to all of the depository institutions as well. That's how I interpreted the debate as it went on in the minutes.

Beckworth: Well, let's move from one repo facility to another, and I want to use your article that you wrote at BPI titled, *Why Is the Federal Reserve Abetting a Drain of Deposits from Banks?* And this deals with the Overnight Reverse Repo Facility. Let me read the first paragraph and then I'll turn it over to you, Bill, but you write, "Every day, the Federal Reserve borrows money from money market mutual funds, GSEs, and certain other non-banks at its overnight reverse repurchase agreement or ON RRP facility. The facility is subsidizing money market funds as an attractive alternative for uninsured bank depositors. Why is the Fed continuing to operate it at its current 2.2 trillion size?" So walk us through this concern and the role it is playing right now in banking and as it relates to the Fed's balance sheet.

The Overnight Reverse Repo Facility and the Fed’s Balance Sheet

Nelson: Sure. And I'd emphasize this was written with the indomitable Greg Baer, my colleague, and our CEO. So the Overnight RRP Facility, it's a tool that the Fed designed and opened back in, I think it was 2013, 2014. It was intended to be there to help the Fed lift the federal funds rate off of zero when it decided to do so, it did a year or so later. And it was meant to be sort of a backstop to raising the interest rate that the Fed pays on reserve balances, the IORB rate. And basically, it accomplished that objective by broadening out the set of counterparties that, in effect, the Fed was paying interest on reserves. So here's one piece of information that, I think, maybe the most valuable piece of information that I'm going to provide on this podcast. It took me a long time to realize this, perhaps because you keep things compartmentalized, but when the Fed refers to repos and reverse repos, they're referring to them from the perspective of their counterparties.

Nelson: So if the Fed says, I went out and did a reverse repo, what they really mean is that they went out and did a repo. So the Overnight RRP Facility is a place where money funds go in to do reverse repos, which is a loan, an overnight loan to the Fed. The Fed is engaging in a repo, it's borrowing money from the money funds. And it provides collateral, not that it needs to, in the form of Treasury securities. So in effect, money funds are providing overnight funds to the Fed that they can continue to do on an ongoing basis. So really it's just paying interest on reserves, but Congress, for some reason, limited the Fed's authority to pay interest on reserves to depository institutions. So not to be constrained, the Fed opened this facility to broaden that out to money funds, which are very important participants in money markets.

Nelson: So that when it raised rates, it had this broader footprint on money markets and it was successful in helping to raise rates, but then it declined back to zero and it stayed at zero basically. And the reason why it was zero was partly because the Fed's balance sheet declined, but also because the spread between the interest rate that the Fed pays on deposits, the on reserve balances, and the overnight RRP rate, was 25 basis points. And what that meant was that money market rates were generally a fair amount above the rate that the Fed was paying to borrow money at the Overnight RRP Facility. And so nobody was really interested in lending the money there. Now in 2021, the Fed's balance sheet was continuing to grow and the Fed ended a temporary exclusion of reserve balances from the leverage ratio. And that meant suddenly if banks were going to lend the Fed money, that's an asset for the banks, the leverage ratio. They would have to hold capital against that. That made it expensive for banks and it made more sense for money funds to lend the Fed money to continue funding QE4 and the Overnight RRP Facility grew up to where it is now, basically about $2 trillion. Now, it stayed there and it stayed there partly because the spread is now only 10 basis points. The Fed lowered the IORB rate and raised the ON RRP rate. They're only 10 basis points apart and that means that money market rates are generally equal to the Overnight RRP Facility. So that makes it a more attractive place for money funds to put their money, so it's continued at that level.

Nelson: But that also means that, and I'd note, that these were all concerns that were discussed by the FOMC. The FOMC was briefed on these concerns and papers were written on these concerns. The Overnight RRP Facility is sort of a magnet for flight to quality flows. It's perfectly safe. The price doesn't change when assets flow into it. So there was quite a bit of concern when it was being designed that it would increase financial instability by pulling money out of the banking system when there were flight to quality flows. And so that's sort of exactly what it's doing right now. And it went down because FHLB is... It's all a little bit confusing because lots of things move around to determine usage, but nevertheless, as deposits flowed out of the banking system and into the money markets, money markets invested in the Overnight RRP Facility and the facility's actually 40% of money market mutual funds’ assets. It's an extraordinary footprint of the Federal Reserve on these assets, and the rate doesn't go down when all of that money is flowing in. So money funds are able to continue to offer basically the same rates despite the inflows and that's abetted the outflows from the Federal Reserve. So exactly what the FOMC was worried about and briefed on has come to pass. I mean, there's a number of good reasons to reconsider how to handle the Overnight RRP Facility, but this is just now another one.

Beckworth: I'm glad you brought up the funding side to the Fed's balance sheet because oftentimes we think of the Fed's assets, kind of as an independent thing. The Fed goes up and buys these assets, but it has to be funded somehow, right? It was funded by banks. Banks put deposits into the Fed, the Fed buys the treasuries. I think we often think of it in the other direction. The bank buys treasuries, puts reserves in the banks and so forth. But it's very, I think, useful when talking about money market mutual funds. They're also depositing funds at the Fed, the Fed recycles and increases the assets. So the question then is, are money market mutual funds helping pay for some of the lending the Fed is doing in the banking turmoil? Or maybe this would've been already coming from the banks, if not from money market funds, just maybe the funds are changing the source, but it would've been the same no matter what.

Nelson: So that's a great question. As the Fed made those loans, their liabilities increased, they had to borrow the money, if the Overnight RRP Facility hadn't been there, it would've taken the form of reserve balances. You can't tie one asset to one liability on the balance sheet, it's always sort of a mix. The one thing we know it wasn't was currency, because currency's by demand, and of course that's what the Fed used to be funded with, to go back to our discussion at the very outset. So yeah, it is being funded by the Overnight RRP Facility and reserve balances. And I do find it is unorthodox to talk about these things.

Beckworth: But it's useful.

Nelson: But I think it's a useful device.

Beckworth: It is. And going back to Governor Chris Waller, I mentioned earlier, he considers these two things fungible reserves, overnight reverse repos, they can be substituted depending on what happens to the spreads, as you mentioned, between the two. So another question on this Bill, is you mentioned that the Overnight Reverse Repo Facility began in 2013, and part of the reason for its ongoing existence is that it helps plug a leaky floor system. And I like that framing of it as well.

Nelson: I think it's a squishy floor and a leaky ceiling.

Beckworth: Okay, it's a leaky ceiling, squishy floor. Question is, is this something that's legal under the Federal Reserve Act? I mean, what's the justification or the authorization that allows the Fed to do this?

Nelson: So this is done under Section 14, which is where the Fed gets the authority to engage in open market operations. And the Fed does engage in, has always engaged in, repos and reverse repos with primary dealers. And there is no limit on the counterparties with whom it can engage in repos as far as I know. So I think the issue is, in the spirit of the law, by having a standing reverse repo facility, you're basically replicating deposits and you're very consciously doing so to broaden that deposit taking and interest paying authority out. I mean, it was done to widen that out beyond the commercial banks and other depository institutions that the Fed's authorized to pay interest to. So it is legal in the sense that the Fed lawyers who are not only wise and, I think, sound and are all my friends, so I must say it is legal. But I do wonder sometimes about whether it's in the spirit of-.

Beckworth: So it follows the letter of the law, but maybe it's pushing the boundaries of the spirit of law.

Nelson: It seems to [inaudible].

Beckworth: Well, let me throw another objection to it and then I'll throw an objection to my objection. But my objection is this, is that it seems to me it's opening up the Fed's balance sheet to more and more counterparties. And one could call it the creeping nationalization of the Fed's balance sheet or the creeping opening access to the public. So initially it's just banks and credit unions, similar institutions, depository institutions. Now we have money market funds, we have the GSEs, we also have central clearing parties that serve as an important role. So we're slowly opening access to the Fed's balance sheet to more and more non-bank financial intermediaries. So someone who, for example, wants to advocate for Fed accounts or digital currency could say, "Hey, why not us? You let those guys in. Why not let me in as well?" So that's one potential danger I see from this.

Beckworth: But let me push back against that objection I just raised. And that is, the Fed is not like any other central bank in that it's dealing with the global reserve currency of the world. Dollar creation happens all around the world in many different institutions, shadow banking, if you want to call it that. The global currency system… the dollar is unmatched, unrivaled, and the Fed has to step in whenever there's a crisis. If it wants to maintain order and money markets and the liquidity perspective, it has to be creative and offer facilities for money market funds, for other financial intermediaries in order to preserve the stability of the dollar system. Now, I also think what this does is it reinforces the spread of the dollar. If global investors know the Fed will backup, it will open up, for example, dollar swap lines and other central banks when there's a panic.

Beckworth: And in fact, during the banking turmoil, it extended the hours of the dollar swap lines that central banks could tap into it. And that, in my mind, sends a strong signal that we're here, we've got this system backed up if push comes to shove. And I think investors around the world notice that, and on the margin increases the demand for dollar denominated assets wherever they may be. So the Fed is kind of boxed into a corner, and it has to be more flexible in how it deals with the fact that it manages this most important currency in the world. So there's a tension there in those two views I've just outlined. On one hand, I worry about the opening of the Fed's balance sheet to everyone, but at the same time, I think it's inevitable that it has to open somewhat more given the role of the dollar.

The Dangers vs. Necessity of Opening Up the Fed’s Balance Sheet

Nelson: So I agree with you, but one of the previous podcasts that we did together was talking about an op-ed I wrote titled, *I Don't Know Why She Swallowed a Fly,* which of course was-

Beckworth: Oh, yes. Excellent.

Nelson: The old lady that swallowed the spider to catch the fly and then swallowed the bird to catch the spider, and so on and so on with the point being the Fed keeps creating these problems for itself, which it solves by having a bigger and more involved Fed. And so I think that that applies here. So by having QE3 and ending up with a balance sheet that was so big that it couldn't go back to a corridor system, and it was so big that it sort of pushed beyond the limits of the banking system to hold, it needed to create the Overnight RRP Facility in order to lift off in 2013. And that wasn't a crisis that they had to respond to, at least it was a crisis of their own making. In 2021 we weren't in the midst of a crisis, by reinstating capital charges on what everybody agrees is a riskless asset, reserve balances, again, and simultaneously growing unnecessarily large, well beyond the point where it was necessary to keep buying assets.

Nelson: The Fed again increased the Overnight RRP Facility and expanded its footprint wide extraordinarily. I'm already beginning to think about the possibility of writing an addendum to that piece as we're learning more about, not entirely, but the extent to which the banking problems that we've seen recently were the result of examination problems. And so that seems like, again, the response might be to come in with the heavy hand, but that's yet to be written. I do want to go back to Governor Waller's observation about the two being basically replaceable one for the other, or substitutable one for the other. I don't agree with that, and I think it matters. I mean, it is true that as the Overnight RRP Facility shrinks, reserve balances will grow. But I think it doesn't matter for a couple of reasons.

Nelson: So when the Fed is conducting an ample reserve framework, they're not conducting an ample reserves plus overnight RRP framework. Reserve balances themselves have to be ample because banks use reserve balances to meet specific needs in terms of meeting their liquidity requirements, meeting the risk of outflows if there's a deposit run, clearing needs, avoiding daylight overdrafts given that those two are now looked badly upon. For all of these reasons, banks have to hold reserves. Overnight RRPs are in no way a substitute for that. So for those reserve balances to be ample, it's literally the reserve balances that have to be several hundred billion dollars above the level at which the market for reserves would get tight. Now, you might think that the Overnight RRP Facility would sort of follow a LIFO process, it grew at the end and then it would shrink first as the Fed shrinks, it hasn't done that.

Nelson: And the Fed's description of what its plan is, it doesn't fill me with confidence. So their plan is that as the balance sheet shrinks, money market conditions will tighten. That will lift money market rates up above the Overnight RRP Facility and therefore the Overnight RRP Facility will shrink. That will in turn create reserve balances, expanding reserve balances, allowing QT to continue until again money market conditions get tight and then they're at the end of QT, they'll have to stop because reserves are no longer ample. I mean, that plan requires that those two types of tightness come in that order. It's quite a delicate thing. So they want markets to not be tight, they want them to be loose, but they want them to be tight enough to shrink the Overnight RRP Facility. I don't see any particular reason to believe that they won't run out of ample reserves before they achieve the level of tightness needed to get people to move out of the Overnight RRP Facility, which is why I think that they should return the spread to where it was before and simply move the IORB rate up to the top of the range and the overnight RRP rate to the bottom of the range just as it was for most of its operations and provide a push to push that Overnight RRP Facility down. And partly because like you, I'd like the Fed to be as small as possible and I'd like therefore for QT to continue for as long as possible. And if the Overnight RRP Facility doesn't shrink, then they maybe only have eight, nine months to go of QT. Whereas if it does, they have another couple of several years of QT that they can do.

Beckworth: So Bill, in closing, walk us through that proposal of yours explicitly. How much would you raise the spread or the rates above each other, and would you do it gradually, over some horizon? What are your thoughts?

Bill’s Proposal

Nelson: Well, a lot of it has to do with sort of optics and communication. The Fed can't just lower the overnight RRP rate by itself because if it just lowered it, the fed funds rate would just fall and it would fall very close to or even possibly below the FOMC's target range for the federal funds rate. And it needs to sort of simultaneously raise the IORB rate and lower the overnight RRP rate. And I think that it might be uncomfortable raising the IORB rate at a time when it's losing $2 billion a week. That might be part of the problem here. And there's no magic formula for how they could do this, but my thought is that it would be quite reasonable to explain, we're simply returning these things to where they were before in the past, and normally they were at the top and the bottom of the range. And now that we're well off of zero and we're shrinking our balance sheet, we're simply going to return them to where they were before and just to do it in one step. You don't want this to want to get in the way of keeping rates changed or raising rates. You just want it to be sort of in the background, in the weeds, something that only David Beckworth at Mercatus would notice and raise the rates to just a weedy sort of, we're just technically returning these things to where they were and that would help. So that's my thought. Raise the IORB rate 10 basis points, lower the overnight RRP rate five basis points.

Beckworth: And I would note that if they did that, there would be some savings, right? You lowered the overnight reverse repo rate. If you're worried about higher interest expense on the IORB, there is some offset from lowering the other rate, right? So it's not like a complete increase. There's some bit of an offset there, so it may not be as hard as one might imagine. But with that, our time is up. Our guest today has been Bill Nelson. Bill, thank you for coming on the show again.

Nelson: Oh, it's been a pleasure. Thank you for all of your great work. I love your show.

Photo by Alex Wong via Getty Images