Bill Nelson on the Repo Market Stress, the Fed's Operating System, and the Prospects for a Standing Repo Facility

David Beckworth: Our guest today is Bill Nelson. Bill was a chief economist at the Bank Policy Institute. Bill previously was a deputy director of the Division of Monetary Affairs at the Federal Reserve Board, where his responsibilities included monetary policy analysis, discount window policy analysis, and financial institution supervision. Bill also worked closely with the BIS working groups in the design of liquidity regulations. Bill has written widely on the Fed's operating system, and joins us today to discuss it, and the recent turmoil in money markets. Bill, welcome to the show.

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

Bill Nelson: Thank you. Thanks for having me.

Beckworth: Oh, it's great to have you on. I've participated with you in an event at AEI. We were just talking about corridor systems, and George Selgin's book, but you've written a lot about the operating system, the money market turmoil, and we'll provide links at the webpage show to some of your recent pieces, but very informative. If you really want to know what's going on, Bill's the man to talk to here.

Beckworth: Before we get into all of that, I'm curious. How did you get into this field? How did you become an economist, number one, and how did you wind up being an expert on the Fed's operating system?

Nelson: As it turns out, I first decided I wanted to be an economist at the Fed back when I was 13 years old, or maybe 14 years old. I was in a middle school class with an excellent professor, Salvador Fabrizio…

Beckworth:  Amazing.

Nelson: ...teaching economics, and evidently... I don't really remember the details, but evidently he described the Fed, and what the Fed did, and that's when I decided that was what I wanted to do. And then when I got to university, and started to take economics, it remained something that was very... of great interest to me. Ended up going to the Fed as a research assistant after college, and then went to grad school. Didn't really intend, necessarily, to come back to the Fed, but those were the opportunities that opened up for me, and came back and joined the Fed then, and that was where I stayed up until a few years ago.

Beckworth: Yes, and you've worked extensively on liquidity regulations. Among other things. You've done other work, as well. But as it pertains to our show, you did a lot of work on liquidity regulations, which is kind of at the center of the controversy now about money markets, repo markets. So tell us, how did that unfold at the Fed? You're working at the Fed, you sat on some committees at the BIS. You chaired some. So tell us about that experience.

Nelson: Sure. I started out... I've always worked in the Division of Monetary Affairs at the Fed, and that's the division where, among other things, the discount window is housed. The responsibility for lending, which is normally just a piece of monetary policy implementation. I know it's associated with emergency lending and crisis response, but it's in fact really just a tool of monetary policy, primarily. 99 percent of the time. And it was sort of from that... I had had that responsibility starting in ‘98, but in 2004, building on that responsibility, I became the first chief of a group to worry about financial stability.

But that was part of the Monetary Affairs Division, and not to be confused with the Division of Financial Stability. That's there now. But because of that, and because of that role as being responsible for the discount window, when the financial crisis came around, I ended up having a lot of responsibility for the emergency lending facilities that the Fed put up. So in particular, I worked on the AIG loan. I worked on a number of... and in particular, I was [on] the board staff leading the post-Lehman, sort of broad-based envelope, alphabet soup of facilities that you hear about.

After the crisis, there was a decision made at the Fed, particularly by Dan Tarullo, then governor and sort of the overseer of supervision, that there needed to be a redesign of how supervision was done. And in particular, they wanted to bring people from the research side into the supervision and regulation. And I was tapped. By that time I was deputy director of Monetary Affairs, and I was tapped to take on that responsibility, as well, so I was a founding member of the LISIC, which is the committee that oversees the supervision of the largest institutions. And I was on the steering committee for the CLAR, which does the horizontal liquidity review.

Started to get very involved in assessing and supervision. Continued to be involved in financial stability issues, and lending issues. Part of the Basel III post-crisis regulatory framework, one of the pieces was a liquidity requirement called the LCR, which is a requirement that banks have 30 days of liquidity available to meet contingencies. Well, when that came out, the very first draft of that came out, I and a number of... several other economists looked at it and had concerns about the systemic implications of how it was written. Expressed those concerns. They ended up being read by the committees at the BIS in Basel that were really worried more about central banks and financial stability. And they asked me to run a workshop on it, and then lead a sort of a working group on it that wrote a report that fed into the revision of that regulation.

And from there, I just ended up being involved in a number of the Basel work groups that helped design not just the LCR, but the NSFR, which is sort of a longer-term liquidity metric that hasn't been adopted yet widely. And then last, and sort of not long before I left the end of the Fed, I was involved in a group with one of your other guests, Ulrich Bindseil and I led this group that worked on the impact of the Basel III regulations on monetary policy, which was a... And that's kind of where I spent a lot of my time. Sort of in that space between monetary policy and regulation and supervision.

Beckworth: Okay, so you really know the plumbing well. On the regulatory side, the operations side. You've been in the trenches, and I guess I didn't realize you were in the trenches also in 2009, when things were getting kind of hairy there at the Fed, and you had to deal with banks that were going under and stuff. So you probably didn't get a lot of sleep, I'm guessing, during that-

Nelson: No, that's true. There was a long period where I didn't get to see much of my family.

Beckworth: Oh, wow. Okay. Wow, so you've seen it all. You've seen the crisis mode. You've seen how to do regulation mode. How to avoid the next crisis. And now you're still working on that at the Bank Policy Institute.

Nelson: Yes.

Beckworth: And you've really been insightful and helpful for me to understand what's happened these past few weeks in the repo market, money market, and I want to get there, and I think that's kind of the main point of our show. But I think in order for our listeners to maybe get there, all of us to get there, it's probably useful to go back and maybe outline the history of the Fed's operating system. That's kind of a backdrop to some of these points we're going to talk about. So tell us, what was the Fed's operating system prior to the end of 2008?

Nelson: Sure. And as you say, this is an issue that I'm very interested in, sort of how the Fed implements monetary policy, how it achieves the interest rates that it seeks to target, basically. Pre-crisis, and then for decades before that, what the Fed did was basically it chose an interest rate that it wanted to target. The FOMC did. And then it adjusted its balance sheet so that deposits at the Fed, or what are called reserves, were equal to an amount that banks would be happy to hold at that interest rate.

But that interest rate was somewhere below the discount rate, and somewhere above zero, because the Fed didn't pay interest on reserves up until actually just after the crisis. But basically, it kept reserves in... To say it's in scarce supply kind of makes it sound a little different than it was. Basically the Fed went out and talked to the banks, and determined what was necessary. It provided that amount, and interest rates just tended to prevail where the FOMC announced that its target would be. One of the defining moments for a young economist at the Fed happens when... Every morning, the Monetary Affairs and the New York Fed would have a call where they jointly made a decision about where to leave reserves to satisfy the banking system's demand. Then for the first time, the FOMC would have changed the level of the interest rates, and you'd think, "Oh wow, now I'm really going to see monetary policy happen."

And you'd go into that call, and you'd realize that they're going to provide today exactly the same level of reserves that they provided yesterday, because pretty much interest rates just shift to the new level, and banks largely continued to demand the amount that they needed for their clearing needs. And they didn't want to hold more than meeting their reserve requirements, because they weren't getting any interest on it. So excess reserves, which will come up again, were largely about $1 to $2 billion. And so the Fed's job was to decide well, what were the things that were going to move demand and supply of reserves around a little bit today. Let's make sure we do the right repo operation to leave that supply out there, and interest rates would move there.

Fed was very good at that, and it became increasingly good at that, so that interest... There was extraordinary little volatility in the federal funds rate, and it wasn't really that hard, either. I mean, it was executed with great skill and precision, but sometimes people make it sound like it was this narrow balancing act that had to be done. But honestly, it was really more straightforward than that. And so that's the way it was done. That rate...

Importantly, the Fed engaged in small operations with the primary dealers, which are the large broker dealers with which it did business. It would adjust its repo book with those institutions. The Fed's repos weren't that big, so they weren't critical to anybody's funding, or anything like that. It was a small footprint on the financial system. Those operations changed the Fed's balance sheet, to change the level of reserve balances, and that in turn affected the federal funds market, which was a different market, with different counterparties. Commercial banks, primarily.

It didn't operate in that market, except for the very rare discount window loan that happened if things maybe got messed up. But banks would borrow and lend from each other in an unsecured basis. That's what a federal funds loan is, and that's what the federal funds rate is. It's a market rate, not an official rate, despite the word federal in it. And then those rates were well... That federal funds rate was very well transmitted into the rest of the financial system. So the Fed had a light imprint, and achieved good interest rate control in the pre-crisis framework.

Beckworth: Okay, so the Fed's footprint was minimal. There was healthy interbank lending, overnight interbank lending. Unsecured lending, as you mentioned. And that served a purpose too, right? It helped banks develop discipline. They signaled about each other. Counterparty risk. Is that right? Is that fair, or not?

Nelson: Well, and I shouldn't downplay the role of secured lending. Secured lending was of course a huge and important part of the interbank market, as well. It's just that the federal funds market was defined to be an interbank, unsecured transaction. Or a bank, or a branch of a foreign bank operating in the U.S. So the market itself, I'm not... Honestly, banks were either in the market, or they were not in the market. And if you were in the market, there was very little differentiation among banks, in terms of the rate that was paid. Even when... I remember times when they might have been concerned about Japanese banks or not, it was very difficult to find evidence that there was a premium paid. You basically paid the federal funds rate.

Beckworth: Let me ask this question, then. Have we lost anything? We'll get to the current system, but in the current system, there really isn't this unsecured... A healthy, meaningful unsecured overnight interbank lending market, right? Have we lost any useful information by being where we are today compared to a previous system?

Nelson: So, maybe I'll take a step back, and walk up to where... how we kind of got to where we were.

Beckworth: Yeah. Please do that.

Nelson: During the crisis, the Fed vastly expanded its balance sheet, through the LSAP. Large scale asset purchase programs, or what people tend to call QE, and those programs were designed to take longer terms... Once the Fed's short-term interest rate target got to zero, there wasn't really any way for the Fed to stimulate the economy using its traditional tools. So one of the things that it did was it purchased longer-term securities, government securities, just to take those securities out of the market, boosting the price of securities with hefty duration, with long duration, and lowering the yield.

It was really in some sense very much in the same spirit as regular monetary policy, but it was an effort to get longer-term interest rates down, rather than shorter-term interest rates. Those purchases came in waves. There was QE1, QE2, and then there was QE3, which was a flow-based program. Rather than announcing an amount, they announced an amount they would purchase per month, until they were sort of satisfied, as it were. And that program ended up being much bigger than at least the staff forecasts that were being written down at the time, or the discussions at the time were anticipated, and became so big that the Fed sort of changed its plans for how it was going to return back to its pre-crisis implementation framework, and indicated that it was going to normalize things, rather than selling securities, by just letting securities roll off.

In some sense, that was sort of the first step to how we got to where we are now. Another important decision point was made when the Fed took actions which allowed, say the treasury, to instead of keeping all its money in the banking system, and only a little bit at the Fed, it moved all of its money to the Fed. And that resulted in not only a bigger deposit, which sort of... A dollar increase in the treasury's deposit reduces banks' deposits by a dollar, so it needs to be taken into account when conducting policy, but a much more volatile deposit, as it were. And that results in big swings in reserves that would be difficult to manage in the old system. At least that was the Fed's thinking.

And then there was a similar growth in the foreign RP pool. So, the foreign RP pool is a pool of fund... The Fed invests for foreign central banks, and for foreign international institutions, their dollar reserves for them, in basically RPs that it does against the Fed's securities, and pays them an interest rate on that. That pool also grew sharply, and the controls that the Fed had placed on those accounts, to limit their volatility, were removed. And so in some sense, the Fed made a decision that left it with a big balance sheet, and then given a big balance sheet, it took actions that ended up making reserves much more volatile, which kind of in some sense locked it in to conducting policy with a big balance sheet.

At the same time on the other side, sort of supervisors and commercial banks were now able to hold reserves as a form of liquidity, because they paid interest, and largely became accustomed to using those as a tool for managing their liquidity risk. The Fed has now been open about the fact that supervisors have been expressing a preference for excess reserves, even though liquidity requirements themselves don't necessarily.

Beckworth: They could be holding treasuries, right?

Instead. Right. They could be holding treasuries, according to the public liquidity requirements, but there's been a recognition that the supervisors have been saying, "No, no. We really want you to hold excess reserves because those are cash immediately." And then also commercial banks, and branches of foreign banks have actually become... They've become vastly more reluctant to ever end up at the discount window, which means partly they need to hold enough cash at the Fed to meet any kind of contingency that could show up. Even late in the day, after interbank markets close, and that's boosted their holdings.

Now, I've been saying for several years now that when the Fed shrinks its balance sheet, it's going to have to expect to encounter periodic resistance. Not because they've reached the floor, but just because institutions have built in this use of excess reserves, and it's going to take a while for the market prices that move in response to the Fed's shrinking balance sheet to get institutions and everyone to kind of move down their holdings of excess reserves in response. And you have seen some substantial movements down over the past year for those reasons. But still, as was seen last week, and as such we'll discuss, there's still nevertheless, it's hard for the institutions to move quickly in response to these prices, and for... Those institutions include supervisors and other supervisors.

So in some sense, the Fed found itself with a big balance sheet that it adopted because it was very focused on stimulating the economy. We were in very dire straits. But then ended up kind of being locked in to this large balance sheet. Now, I am very much of the view that in so doing, it's left itself in some sense stepping in for the private sector, in terms of managing banks' liquidity risk. And I recently learned... This summer, I was giving a paper at the Norges Bank, and was very intrigued to learn that in 2010, they made the reverse decision of what the Fed had made.

They went from a floor system to a corridor system.

Beckworth: Interesting.

Nelson: And the reason that they did so was that they found that no matter what quantity of reserves they provided the bank systems, the banking system would just absorb it, and it was no longer using the interbank market to diversify its liquidity and manage its liquidity risks, but was instead just relying on the Norges Bank, and the Norges bank said, "We're not in that business, and we're going to switch to... of managing liquidity risk. We want banks to do it, and we're going to switch to a corridor system, where reserves are kept relatively scarce, and deposits are not remunerated at a very high level." And indeed, the interbank system revived, and banks did much more of their own management of their liquidity risk.

And that's what I would like. That's where I think the risks lie in the current arrangement, as well as political risks for the Fed. I think it just leaves them much more exposed when they are counterparty to all, and you see things like the Green New Deal. The last part of the New Green Deal said, "How are you going to fund it? Well, we're going to get loans from the Fed to fund it." Suddenly the Fed is out there in a way that I think puts them at a greater risk. And results in a financial system where everyone is sort of safer because they're all doing... Their counterparty on both sides of the transactions is the Fed.

Beckworth:Yeah, let me ask you about the Norges Bank, so the central bank of Norway. It's a interesting story you share about them going back, and you mentioned that banks were acting the same way there that they are here. They were very comfortable, maybe too comfortable getting the reserves that were sent their way. Was there any pushback from the banks when the central bank said, "Hey, we want to go to a corridor." It sounds like to me there could be some discomfort, lodging the banks out of their comfortable positions that they were in.

Nelson: Yeah. I don't know. I've actually shared about all that I do know about their decision, but I can let you know who might be a good person from the Norges Bank to bring in, because it'd be very topical.

Beckworth: Yeah, because I imagine there's some political economy considerations there. The other example that comes to mind, I always point to this, is the Bank of Canada. Bank of Canada had a corridor system, temporarily went to a floor system in the crisis, and then went back to a corridor system, and they seem to do a fairly good job of keeping interest rates within the bands, within their range. And I look to them, and I know they're not the same as us, they've got different banking system, things are a little bit different, but there are examples of corridor systems that work, that have gone from a floor back to a corridor system. So it's reassuring to know it is possible.

Nelson: Yes. Yes. That's a great point, and there's some nice research out of the Fed recently saying how the federal funds market has recovered in the past under these instances. Another area of concern that people have expressed is that the federal funds market is now a fairly peculiar market. It's no longer sort of a robust market, where banks go to each other at the end of the day to square their account needs. But that's what you would expect when there's $1.3 trillion in excess reserves out there. And so I have every expectation that the market would recover.

Now, maybe banks will need a much higher level of excess reserves, given the greater caution, and that could be very... That could be prudent, in fact. But there are ways that the Fed could provide those reserves, while at the same time operating with a corridor system. Things like a voluntary reserve regime, where banks tell the Fed, "This is the quantity I want to hold," and then they're remunerated on that amount, but not on an amount above that.

Beckworth: Everything. Okay.

Nelson: And you get back to a system, no doubt one with the Fed larger, because Fed deposits are the wheels on which the interbank payment system works, but nevertheless, one where at the margin, banks and financial institutions are interacting with each other, and not with the Fed.

Beckworth: So just to go back and recap the history, when the Fed first did QE, or large scale asset purchases, at least you said the first two rounds, QE1, QE2, it had all the intention in the world to go back to what it had pre-2008, right? It intended to shrink the balance sheet back down to as small a level as possible. It was effectively aiming to go back to a corridor system, but only when you got to QE3, they begin to have doubts? I know this year in January they made it official, but when did they begin to realize, "Maybe we should stick with what we got." Because my sense was this decision was almost kind of... I think what you were getting at, kind of a status quo, inertia. We're here, it works well enough, and by the way, it allegedly gives us great interest rate control. Am I wrong in that assessment? They kind of found themselves in this position, as opposed to thoughtfully saying, "Hey, let's go to a floor system."

Nelson:  I think that that conclusion is correct. The way I would describe the deliberation on it, and so this can all be found in the transcripts, and I've written on the... I've written on how this process went. So, the Fed first started to think about its operating regime in the new way when Congress told it it would get the authority to pay interest on reserves, so that was back in 2006, and they promised the authority would begin... I forget when. It was quite a bit later.

But then that authority was moved up in fall of 2008, so that the Fed would have that tool to respond, as part of its response to the financial crisis. But if you look at the FOMC memos, most of the memos, many of the memos that the staff wrote for the FOMC are also publicly released after five years, so if you go back to look at the memos that were written, I think this was April 2008, about different operating regimes that the Fed could use. The idea of using a floor system was discussed, but it was dismissed largely.

Beckworth: Oh, interesting.

Nelson: Alright, so... And indeed at that point, the estimated amount of excess reserves that would be needed was $35 billion, as opposed to the $1.3 trillion that apparently is needed now. And so, nevertheless, I think views were split, and what happened was as the Fed added more reserves, the committee deferred discussion of what they would do after the crisis, and focused instead on what to do in response to the crisis. And I think reached the conclusion that they didn't need to make that decision now. They would learn more about things with experience. And so they just deferred that discussion until later.

And really, it was really only seriously reengaged again when interest rates were... when the Fed was foreseeing the rise of interest rates in the discernible horizon.

Beckworth: Okay, fair enough. Well, let's move to these past few weeks, what's happened in the money markets, because that's... What we've just talked about is the backdrop, the context, and we'll refer to some of that. And a few weeks ago we all woke up and saw that there's problems in the repo market, and what's interesting about this, and I really recommend this piece to our listeners, is you actually wrote about this a couple weeks before it happened. So, are you a prophet, or are you just a good economist?

Nelson: Well, actually I just have smart friends who I listen to, I would have to say. There's just... We're just a few economists at the Bank Policy Institute. We focus mostly on regulatory and supervisory issues. I do this... I follow monetary policy and these issues a bit on the side, but there is a, as I've mentioned, a place where they come together. But you know, was aware of the fact that the people that do forecast the level of reserve balances were forecasting pretty sharp drops in those levels in September, and in mid-September in particular. And I think a number of the people that were very in the weeds kind of people, that really know... I'll mention in particular, say Lou Crandall, who's the chief economist at Wrightson ICAP, and writes the Money Market Observer, whose forecast I was looking at. People were aware that there was the potential for some problems.

And what we did was we sort of linked that up to the reasons why the demand for excess reserves was so high, including liquidity requirements, and then these supervisory expectations. And took that opportunity to suggest some of the ways that we've identified, where liquidity requirements could be recalibrated to improve them, but also as it would happen, would reduce the level of excess reserves that were demanded. Things like including required reserves as a liquid asset. So you think of required reserves as the amount-

Beckworth: Yeah, that was interesting to read that, that they didn't count it.

Nelson: They didn't count it. So, to be fair, I was the one who made the decision not to count it.

Beckworth: D’oh!

Nelson: No, I think it was the right decision, and it was the right decision based on the regulations that now exist, and what we were saying is they should change those regulations. And the reason it was excluded was because they really weren't written to be in any sense usable or optional. They were written to be required. But what we're proposing is that they should be usable. They should be usable in the same way that other liquidity reserves are usable in a crisis. They should be usable to meet contingency needs, and then built back up. And with that change, one would appropriately include them as a liquid asset.

Now, there's also a legal argument for doing it that I won't get into here. And then a few other changes we were proposing, but then as it happened, that did set us up well for the…

Beckworth: Yeah. No, you guys look great in retrospect. You said, "Hey, Fed. By the way, look out." I mean, there's some other people who were saying I think similar things, but you called it pretty... You were like two weeks before so I... Kudos to you.

Nelson: Thank you.

Beckworth: But kind of all this simple story, I want to get the simple, or the proximate story for what happened, and then there's a deeper story that's related to some puzzles surrounding this, but kind of the proximate story is a perfect storm that effectively caused the Fed to trip up and stumble back into a corridor system. There's that flat part of the horizontal demand curve. You're on a floor system. Abundant reserves. They've been shrinking, and I'll have to attach a graph to our webpage, so listeners can... Bill knows what I'm talking about, but with a floor system, the supply of reserves is out on the horizontal part of the demand curve, or the area where you reduce opportunity cost to holding reserves.

But in any event, and the Fed officials actually talk about... Lorie Logan, I think Simon Potter, have all talked about they want to avoid getting back to the upward-sloping part of it, because that's where the corridor system kicks in. And so they were shrinking that, their supply, supply, and they wanted to get to that sweet spot where they're right before they hit that point where they're back on the corridor, and I think the story is kind of a perfect storm. They got close, maybe too close, and all these reserves were drained out because of Treasury's actions, and oopsie daisy, we're back in a... We don't mean to be here, back in a system where there's overnight activity. Rates shot to 10 percent, and now the Fed's doing temporary open market operations to push them back out on the horizontal part of the curve.

But that's kind of the proximate... It's supply and demand. Simple supply and demand story. I've been saying it. But as the weeks have gone on, there's a little bit of a puzzle, or maybe a deeper mystery, and you've written about this. Because it kind of begs the question, "Well, why did it happen?" I mean, surely the Fed saw this coming. Why didn't the big banks lend... banks sitting on reserves earning two percent, why didn't they lend into a market where they could have earned 10 percent overnight rate? So there's a lot of mystery.

The superficial story, okay, but tell... Dig into the deeper story for us.

Nelson: Gladly, but first of all I want to admire... I admire your bravery for trying to provide a verbal description of the demand for the pool. Demand for excess reserves. And I agree with your characterization, although just to be clear, as I was discussing earlier, I don't really think that the steep part of the excess demand curve is fixed at $1.3 trillion, when the Fed used to think it was $35 billion.

Beckworth: Okay.

Nelson: Right? But rather I think that those are sort of temporary periods of steepness, that would then be gradually worked down. And if you read the FOMC's minutes, that's kind of what they had in mind. They were going to move down closer, and they anticipated some volatility, but they were going to continue to move down, and give banks an-

Beckworth: Maybe a short-run equilibrium, versus a long-run-

Nelson: Sort of a short-run equilibrium versus a long-run equilibrium.

Beckworth: Okay. Fair enough.

Nelson: Exactly. And I do think that being near that short-run equilibrium, and then suddenly being pushed below it was a factor, but for me, just... I kind of think about the proximate cause in a little bit of a different way, so let me set it up just a little bit differently.

Beckworth: Please do. Yes.

Nelson: I think one could think about... To some extent, you could think about this really as just a supply and demand dislocation in repo markets, and not even bring in reserves, or the Fed's balance sheet, as a place to start. So, on September 16th, there was a large settlement of treasury securities. I forget the figure. $50 billion, or something like that. Maybe... I forget the figure. Maybe 36. But in any case, securities hit the market, that the institutions that hold those securities need to finance them. They finance them in the repo market, so the demand for repo financing went up.

At the same time, corporate tax day was the 15th, was a Sunday, so 16th. Corporations shift large quantities of money to the Treasury out of their institutional-only money fund. Government money funds where they've been keeping their cash. Institutional-only money funds are big investors in Treasury repo, so you have this... and normally you would have seen maybe on those kinds of days about $100 billion drop in excess reserves, and maybe a $30 billion drop in holdings and money funds, just to go by what folks have told me that are the normal estimates. So you have this drop, and since institutional-only money funds invest in repo, but have much less money, of their customers' money to invest, so they pull back.

You have this simultaneous reduction in the supply of repo financing, and increase in the demand for repo financing, that manifested itself in repo rates moving from two percent up to... In some of the trades were up to 10 percent. And that persisted for Monday, and Tuesday, and by Wednesday the Fed stepped in, I think it was. Maybe Tuesday.

So, that then just sort of raises some questions, which is maybe where the excess reserves would come into it. I think of there as being sort of two crucial questions. Banks have these huge quantities of excess reserves, right? Vastly more than they had in the past. Why weren't those... Why weren't banks willing to say, "I'm going to just substitute from excess reserves to reverse repo, another super safe investment, receiving treasuries in exchange that count towards my LCR, at least, and perhaps towards other liquidity requirements that we can't see, like resolution requirements, or internal supervisor requirements." So on the face of it, you'd think that they would do that substitution, but they were very reluctant to do so.

And I think that there are really sort of two causes for that reluctance, that we discussed to some extent earlier-

Beckworth: Yeah, please do. No, no.

Nelson: But I'll just... So, one was, as we were discussing earlier, that there are liquidity requirements that are more opaque. One's related to resolution, one's related to supervisory examinations and what are called reg YY requirements, requirements that are placed on the largest institutions, that left banks feeling like they needed to hold high levels of excess reserves. And there were just sort of conveyed expectations from the supervisors, according to... So, remarks... I asked two years ago. I actually, at Hoover, one of these Hoover monetary policy conferences, I asked Vice Chairman Quarles then. I said, "So, you're interested in getting your balance sheet down, and you've talked to us about liquidity regulations. Are you concerned... We've heard that supervisors are doing this. Are you concerned about that?"

And even then, two years ago, he indicated that is something that we've heard that supervisors are doing, and we're going to revisit that, and I think that they're revisiting it with a vengeance right now.

Beckworth: Where do they... Let me ask, where did that come from? Where did that bias towards reserves on the supervisors' part come from originally, do you think?

Nelson: Well, so reserves are the ultimate liquid asset. So, if I have a treasury, even a treasury security, and something happens late in the day that I need money right now in large amounts, in some sense there's nothing other than a deposit at the Fed, except for a discount window loan, which basically puts money in your deposit at the Fed, that could meet that need. So from a sense of let's think about every conceivable, possible thing that could go wrong, and well, reserves were paying interest rates that were at or above, a little above market rates.

Beckworth: So why not?

Nelson: So why not, in some sense. And because they do offer that, that perfect liquidity. Now of course, only commercial banks, or U.S. branches of foreign banks, have accounts at the Fed, and so can have reserves. That's something that's often not understood. Broker dealers can't have reserves. Bank holding companies don't have reserves. Only commercial banks, thrifts-

Beckworth: So only those institutions that have access to the Fed's bank balance sheet had this preference imposed upon them, or maybe their own preference, because they could-

Nelson: Exactly.

Beckworth: Reserves were more liquid than treasuries for them. Now, outside of them, there's been discussions how treasuries are more liquid, because other firms can't use reserves like banks can, but you're focusing in on the banks, and the financial firms that have access.

Nelson: That's where that preference could become. You can see how it would build in. They cost the same. It's more liquid. Then it sort of gets embedded.

Beckworth: All right. Fair enough.

Nelson: And then the other side of it, as I was discussing, was banks are super cautious. Partly because of unfortunately, sort of the association of borrowing with a bailout, even though borrowing was absolutely what it was intended for during the crisis. The Fed... The lender of last resort is there for that purpose. And certainly for a monetary policy purpose, like interest rates are higher, and reserves are in short supply, that's why you have the discount window. But nevertheless, there's a deep association with public flogging about any possibility of borrowing.

One senior fellow at a foreign institution with branches here said that when he joined the institution, he was told that in the event that his institution did ever end up at the discount window, there'd be two phone calls. The New York Fed would call their CEO and ask what happened, and human resources would call him and tell him to clean out his desk and find a new job.

Beckworth: Wow. Pretty harsh.

Nelson: So you know, in that environment, you can get an awful lot of interest rate movements without anybody being willing to go to the discount window. And that actually comes... That actually raises the issue of the standing repo facility, which we have time, we should perhaps talk about-

Beckworth: Absolutely. Let's definitely talk about that.

Nelson: The other part of the mystery is, well, why didn't banks just substitute? Even if they didn't want to reduce their holdings of excess reserves, not substitute, but why don't they just get bigger? Why didn't they go out into wholesale markets, borrow, and re-lend into the corners of the repo market where there were constraints. And there, that has a lot to do with capital requirements, right? The post-crisis, there were tremendously valuable improvements that were made to the quality of capital, and to how risk was measured. But there was also an increased leverage ratio requirements. Requirements that require banks to hold capital against low, super-low, or no-risk assets like treasuries or reserves, and those requirements have ended up making banks much more careful in how they deploy capital across their parts of the institution.

And repo books used to be able to jump in and respond to price signals, and equilibrating supply and demand imbalances, but now they're very carefully managed and governed, because they have a certain amount of capital that the bank has allotted them. And really the whole episode has highlighted something that the banking system has pointed to, which is that while the post-crisis regulations have made the financial system in many ways safer, it also has reduced the resilience of that market, and the ability to... It in some sense increased... reduced liquidity in the market.

And we've now been able to see that in a relatively benign experiment. Not one where there were credit concerns, or other international problems, but it really highlights the fact that yeah, the liquidity is much less.

Beckworth: Now, you mentioned earlier the Norway example, where they kind of weaned the banks off this reliance on reserves, and moved back towards more normal funding arrangements. Is that possible in the U.S.? I mean, maybe this is where the standing repo facility comes in. But put that to the side. Is it possible, I mean maybe it's just it takes time, like you said, for them to relearn how to engage in this activity?

Nelson: Certainly, and you know, the FOMC was clear in their minutes in November and December 2018, and January 2019, that that was their plan. They intended to gradually reduce the size of their balance sheet. They indicated then that they expected that they would end up bigger than before, but much smaller than where they are. And sort of let those pressures build. But what they found is that what they will have to do to execute that plan, as they are doing now, effectively, is actively go into the market to manage shocks to reserves by taking countervailing actions with respect to their balance sheet that will reduce reserve volatility.

And as long as they've reduced reserve volatility, they will be able to continue to shrink the size of their balance sheet. Ultimately, interest rates will adjust to reconfigure things, so that market rates will be above the rate that the Fed pays on deposits. Banks will economize further on that rate, and begin to manage their liquidity risks by transacting with each other more than by leaving their money at the Fed.

Beckworth: So it's possible.

Nelson: It's certainly possible.

Beckworth: Let me ask this, the very basic question. Is this then partly a story about too many well-intentioned regulations at crossroads with each other? I mean, I've heard one person make the comment there's over 21 different regulations that tie in to bank capital. I mean, it's just so many new regs. One approach I like, I hear, and I'm going to acknowledge my ignorance on this issue, is why not reduce the number of regs, and just impose higher capital requirements? Have banks fund at higher capital... Remember the Choice Act? Two congresses ago?

Nelson: Yeah.

Beckworth: The principle was let's have I think 10 percent capital, and get rid of all the, or much of the Dodd-Frank regs. Would that be a potential solution, or is that just wishful thinking there?

Nelson: There's elements of what you described that I think would be part of a solution. And as I mentioned, I was... One of the groups that I chaired in Basel was about how have the changes in regulations impacted monetary policy, and what that group concluded was that tighter capital requirements, and particularly leverage requirements of the type that you're describing, have reduced the connectivity between financial markets, and reduced the liquidity of financial markets, in a way that will require central banks to operate with a broader set of counterparties to get around. And in some sense, not necessarily with a bigger balance sheet, but with a much more intrusive approach than before, in order to get around the reduced transmission that would occur within financial markets.

So certainly, there are aspects of regulations that this has highlighted, that I've described, that are part of the story here, that I think... and I applaud the efforts of the supervisory regulators to continue in their announced effort to look at... A lot of regulations were put on very quickly, and general direction was well needed. Everyone was pulling in the same direction. I was part of those pulling an oar in that direction. But it's inconceivable that all of those things would have been done, and done exactly right.

Beckworth:  Right. That's fair.

Nelson: And the idea that sort of looking at them to see where there are efficiencies that can be achieved, I think is a very appropriate thing to do, and included among that would be looking at the regulations that contributed to this. And you can see that again in the efforts to say, "Whoa, wait a minute. We don't need everybody to hold excess reserves necessarily." And then this gets back to the part of, I think your description, that I maybe wouldn't agree with, which is that those... The efforts that you described really focused on the leverage ratio. And the leverage ratio is a capital metric that treats all assets as equal. As equally risky.

So it says a treasury security is the same risk as a-

Beckworth: Equity.

Nelson: Well, banks can't hold equity, but let's say construction and land development loan, or other things that are riskier. They're all prudent investments, but nevertheless, a bank needs to hold different levels of capital to ensure that it will be well insulated from the risks of those. And when you have something like the leverage ratio, which treats safe assets, like treasury bills, or deposits at the Fed, as equally risky to loans, or other... Then what you do is you provide an incentive for banks to increase the risk of their portfolio. They're responding to the regulatory incentives that are created.

Beckworth: I see.

Nelson: Banks have an incentive to stop doing things like investing in super-safe reverse repos, where they're making an overnight loan to somebody and getting treasury securities in exchange. Those things become much more expensive for the bank to do-

Beckworth: If you're going to get the same requirements on that and on riskier assets, it creates an incentive to-

Nelson: To shift towards the riskier high-yielding assets.

Beckworth: That's a good point.

Nelson: If you're in a world where it is capital regulations, and not prudent risk management itself, which is determining. And unfortunately, we're kind of... I think the ultimate challenge about thinking about regulations, and bank regulations, is finding... There's got to be a theorem out there that says, "If a regulation isn't irrelevant, it's going to effectively put the government in the job of managing risk." So you need to have regulations that aren't irrelevant, that leave the financial institution in charge, in some sense overall, of making its risk management decisions and allocating credit.

Beckworth: Now that's a great discussion. I want to get to this standing repo facility in the time we have left. Before I get there, I do want to highlight the fact that the Bank Policy Institute did a survey of banks, which was interesting to read. And I want you to talk about that, but I want to even go back before that, because the Federal Reserve does its own survey of banks, and I think at the conference we were at, the AEA, or the little talk we had there, I cited some of the data from that, which suggested that if the... I think the term they used, the consolation of overnight rates were to go up a certain amount above the interest on excess reserves, would banks unwind their holdings of excess reserves if the opportunity cost got high enough?

And I totally bought into that. Wrote a paper kind of arguing that. But this recent... These past two weeks really raised some doubts in my mind, because you did see. That's 2 to 10 percent, that's a huge gap between the consolation overnight rates, and the IOER, and yeah, banks were not... Some banks were not willing to get rid of their reserves. Which then leads to the survey you provided, I think, that sheds lights on that. Tell us about your survey.

Nelson: I would emphasize again, you're looking at events of a couple days, so the fact that banks... There's this sort of the rigidity within a couple days is important to think about. But nevertheless, thank you for pointing to the survey. So, the Fed... People might be familiar with the Fed's quarterly senior loan officer survey. That's the one that tends to get the most attention.

Beckworth: Yeah. That's the one I looked at.

Nelson: It's been going on for a long period of time. Fed also has the authority to do quarterly Senior Financial Officer Surveys, but it just hasn't used that authority for a very long time. But you know, the current environment is a very appropriate one to ask questions about what's driving banks' high level of demand for excess reserves. And they used that survey to investigate that.

What we were struck by, though, is that in the menu of reasons that they provided, they didn't provide liquidity regulations, or supervisory expectations, or any of these things that we were hearing as potential reasons. And so, I don't want to speculate as to why the Fed wasn't able or willing to provide those explanations, but we thought that that was an important part of the story, so we actually didn't... We just went in and did our own survey of our member banks, in which we asked them about their demand for reserves. We basically replicated the questions that were on the Fed's survey, but we provided them reasons that had to do with supervisory expectations, liquidity expectations, other things.

Nelson: And what we found was that banks actually placed quite a high weight on those things as determining their holdings of excess reserves. But to be fair, we have also heard pretty consistently from many banks that that's just part of the story. Another part of the story, again, is the reluctance to ever potentially end up at the discount window again. Even for overnight when it's intended. And therefore need to be very cautious.

Beckworth: So just to be clear, you've brought this point up several times about the stigma, the reluctance to go to the discount window. Are you saying it's more pronounced now than even before 2008? I mean, is there even a greater aversion to going there?

Nelson: No. No, I think it's more pronounced now-

Beckworth: It's always been there.

Nelson: ...than it was. Certainly more pronounced, much more pronounced now than it was pre-crisis.

Beckworth: Okay.

Nelson: Much more. But it's been there for 50 years. As you know. Stigma is a problem that's been associated with the discount window for a long time. Probably because the Fed uses the same tool as a monetary policy tool, and as emergency crisis tool. But in any case, it's massively worse than it was before then, because of the experiences during and after the crisis.

Beckworth: Okay, so you want to... Maybe the farther we get away from 2008, some of that stigma might come down, you think? As people become less risk averse, banks become less risk averse?

Nelson: Well, one of the interesting questions on the Fed's Senior Financial Officer Survey was basically to ask the banks how likely it was that they would use the discount window as a means to address a very short-term need, like the discount window is intended to address. And I do want to repeat that just the Fed made zero losses on any loans, not just through the discount window, but on any program during the crisis. It's something to keep in mind, that people don't appreciate.

But nevertheless, the banks said that basically they attached zero probability of them ever using the discount window again. And that was just a few... That was just a couple years ago, or a year ago, so that makes a nice segue to the standing repo facility.

Beckworth: Yeah, so tell us about that. I know David Andolfatto, Janet Ihrig at the Board of Governors, they proposed this. I know there's been other ideas like it before. But tell us about the proposal. Tell us some of the challenges. And what would it take to get it up and running?

Nelson: Sure. Yes. Okay, so there's two potential reasons why, roughly, why the Fed might be interested in opening a standing repo facility. So, just to explain, a standing repo facility would be a facility that would offer counterparties effectively overnight loans that... The counterparties would provide the Fed government securities. Probably treasuries, but potentially other open market operation eligible securities, like agency MBS. Provide the Fed those securities, the Fed would provide them money, and then the next day that transaction would reverse. The institutions would pay back the money, and the Fed would provide the securities back.

So it looks very much like a collateralized loan. An overnight collateralized loan. But it's done in the form of a purchase and repurchase, which is why it's called a repurchase facility. One confusing thing about the Fed is that they describe these things from the perspective of their counterparties. So if they do a repo operation, they're actually doing a reverse repo operation. It's a repo operation for the other guy.

Beckworth: Interesting.

Nelson: If anyone's ever been confused about that.

Beckworth: It's okay to be confused.

Nelson: Yeah, it's okay to be confused. But anyway, the Fed would say, "We will do that on demand for this fixed rate." That's what makes it a standing repo facility, is that they're going to do it on demand, rather than enter the market to do it.

Beckworth: Okay.

Nelson: One reason to do it would be basically because it effectively is the discount window for commercial banks, but it might look and feel differently. It would be done under Section 14, which is the legal authority to conduct open market operations, instead of Section 10B, which is the authority to make loans under the discount window. Currently there isn't any stigma really associated with monetary policy operations, so there would just be a hope that because they would have a different look and feel, it would be... There wouldn't be stigma associated with it. And for that reason, institutions would be willing to hold smaller quantities of excess reserves. Armed with the knowledge that they could always convert their treasures to cash at the standing repo facility.

And it could be a potential more effective ceiling on a corridor system, which we haven't really discussed, and I don't think we have time to discuss. But that would be the advantage for banks. That presents a challenge, though, and I was very involved in the redesign of the discount window in 2003, where the same issue kind of came up, so I'm sure that they're grappling with similar issues now. Which is that all you really care about for that are the biggest banks, because they're the ones that move the market around. And you could satisfy that objective by having a rate that was just a small quantity above market rates.

But really you've just for political reason, not for legal reason, but really to realistically, if the Fed is going to offer a lending facility to big banks, then it has to offer it to every depository institution, including the very smallest banks. But it also doesn't want to be an ongoing lender, so it needs to offer a rate that's high enough that it makes it unattractive for the very smallest institutions that have higher alternative costs of funds, to not see it as an attractive source of ongoing borrowing. That would require a somewhat higher spread.

But if the spread is higher, then it's kind of less helpful, and it's less appealing as a way to make banks confident in their... So that's one thing. The other thing, though, is it could perhaps be a release, a safety valve for repo markets, themselves. Now, the Fed isn't targeting repo rates, and I would encourage them not to target repo rates, but they sure look like they're sort of targeting repo rates, and so if you're worried about repo rates going too high, and the Fed is offering repo at a fixed rate, then that might help control those rates in the event that there was another dislocation.

I say might because it kind of depends on how they design the facility. If the only counterparties were the big broker dealers that are primary dealers, those are the institutions... Those institutions are the ones with the balance sheet constraints. Ones that are going to be a real problem at year end, for example. Part of the concern. And so, lending to them doesn't actually help. You need to lend to the institutions that need the money, that aren't getting the financing that they need. Those tend to be smaller broker dealers. The hedge funds.

So, they would have to sign, just like they've done for the standing reverse repo facility, where they've included money funds, not just the primary dealers. And the GSEs. They would have to include a much broader set of counterparties, at least to be effective in alleviating these balance sheet constraints. And that takes time and effort, and will no doubt require some discussion.

I had been thinking that they would take... that they weren't going to be doing this soon. That it could be, if they did it, a year. Although the events of last few weeks may give them much incentive to do it much faster.

Beckworth: Maybe these past two weeks have been good for the Fed on several fronts. One, it's helped them to really focus on the challenges that we've been dealt, we've been talking about. But two, to also put front and central the standing repo facility. It's a good catalyst. Maybe the best thing that could have happened is to have experienced these past two weeks for the Fed. Without going in a full-blown crisis, just have a good little knock on the side that says, "Hey, wake up. Focus on these issues. And also think about a standing repo facility."

So maybe this could all turn out to be a blessing in disguise.

Nelson: Well, perhaps, and in particularly for the former. For the latter, it's maybe a good immediate solution, but it also highlights the fact that the way things are currently arranged, the Fed kind of has to be the counterparty to all. It has to be the market maker of first resort.

Beckworth: Right. I guess that's one of the drawbacks you've highlighted in your paper is that it... Ultimately, it becomes the backstop, and so depending on your views, and how much of a footprint intervention you care about, it could be an issue. And I'll also mention George Selgin's made this point. The standing repo facility could also be a step in the direction of a corridor system. It could be a transition point. We don't have time to get into that. We'll link to George Selgin's article.

But this has been a great discussion. Thanks so much for coming on. Our time is up, and our guest today has been Nelson. Bill, thank you for coming on the show.

Nelson: Thank you. It's been fun. Thanks for bringing me in.

Beckworth: Macro Musings is produced by the Mercatus Center at George Mason University. If you haven't already, please subscribe via iTunes or your favorite podcast app. And while you're there, please consider rating us and leaving a review. This helps other thoughtful people like you find the podcast. Thanks for listening.