George Selgin on Repo Market Stress, Fed Balance Sheet Volatility, and a Standing Repo Facility

George Selgin is the director of the Cato Institute’s Center for Monetary and Financial Alternatives and is a returning guest to the Macro Musings podcast. He joins the show today as part of a two week special on the Fed and repo markets, as he helps us take a look at recent repo market stress from the Fed’s perspective. Specifically, David and George discuss the basics of the Fed’s balance sheet, the problematic nature of the Treasury General Account and foreign repo pools, and how George would tweak standing repo facility proposals to more directly address balance sheet volatility

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: Hey Macro Musings listeners. This is your host David Beckworth. Today's show is the first of a two week special on the Fed and repo markets. As you may recall, repo market interest rates spiked in mid-September, reached nearly 10 percent, while the interest rate on reserves was around two percent. This caught many by surprise, including the Federal Reserve, and revealed there was a severe funding pressure in the repo market, that is, there weren't enough bank reserves to fund repo activity. As a result, the Fed injected liquidity into the financial system through overnight repos, term repos, and outright purchases of assets.

Beckworth: There is some concern however that at the end of 2019 there will once again be repo market stress as lending by the big banks into the repo market will be dialed back as these big banks tidy up their balance sheet for regulatory reasons. No one knows for sure, but there are good reasons to be worried, and today's show, part one of the series, we will take a look at this problem from the Fed's perspective with the help of George Selgin. Next week we'll be joined by Josh Galper who is deep in the trenches of the repo market, and he will help us examine this problem from the perspective of the repo market. George, welcome back to a show.

George Selgin: Great to be here again, David.

Beckworth: Yes, you've written several interesting pieces on this situation, this development. We'll link to those in the show notes. You've also outlined some proposals to fix this problem for the near term, some practical fixes that kind of sits to the side for the time being, what type of operating system the Fed has, but given where we are, what are some pragmatic next steps and we'll get to those probably near the end of the program. What I want to do though is work our way through what happened in September and then talk about what might happen in December, this month, but later in the month, and there are some concerns about that.

Beckworth: The last time you were on the show we talked about real-time payments and I encourage our listeners to go back and listen to that. George had some great suggestions there too for some fixes for helping those who struggle with the payment system, those lower income folks. But today we want to get into the repo market stress and the Fed's relationship to that, and this is the road map we're going to take with George today. First we're going to review what happened, as I mentioned in September, what could happen, and his suggested solutions. And there are some really interesting ones, a little provocative, George, I will say, I'm sure you've got some feedback on those proposals as well.

Beckworth: But to help us understand, help me understand before we get into all of this, let's talk about the Fed's balance sheet. The basics of the Fed's balance sheet. So like any body, any institution, the Fed's balance sheet has an asset side and a liability side. And on the asset side, I think that's pretty straight forward. The Fed buys treasuries, they bought agency bonds and debt, can issue loans. Where it gets tricky and where the story I think is really buried is on the liability side.

Selgin: That's right. Yeah.

Fed Balance Sheet Basics

Beckworth: So talk us to the liability side of the Fed's balance sheet.

Selgin: Right. So, people are familiar with one of the most important Fed liabilities, that's the currency that's circulating in the economy, Federal Reserve notes. Next to that, they may be familiar with the fact that banks keep deposits at the Fed, which deposits count along with currency they have in their tills and vaults and ATMs, as part of their reserves. So those are the two more familiar Federal Reserve liabilities. There are others, however, and those turn out to be very important in the repo market troubles that the Fed has been experiencing.

Selgin: And chief among them as far as that issue is concerned, are two liabilities. First, the so-called treasury general account, which is an account the Treasury keeps at the New York Fed, which is a consolidation of accounts it has at all of the Federal Reserve banks that they put the money all together at the end of every day. And the other one is the foreign repo pool, and that is liabilities consisting of deposits by foreign central banks and other foreign official institutions that can place money at the Fed on deposit as it were, and earn interest on it because the Fed turns around and does repo operations with those funds. So those are the liabilities that people are less familiar with and that I think we're going to be talking about a lot.

Beckworth: Yeah, absolutely. And part of the story for September, or THE story for September, was that there weren't enough reserves given the demand for them during this time, given some of the developments in these accounts, these sub-liability accounts. And so what I want to think through before we get into the details of the story is, how do you get changes in reserves on the Fed's balance sheet? So there's several ways I can think, I want to walk through them with you and help me understand them.

Selgin: Sure.

Beckworth: So the first way there could be fewer reserves would be for the Fed to shrink its balance sheet outright. So the Fed sells off assets and it has to pull in reserves. So when it sells a treasury bill back to the public, it's taken out the reserves that would correspond to that. And that's what happened with quantitative tightening. Is that right?

Selgin: Yes, that's right. One way to think about this is to first imagine that the Fed's liabilities are all bank deposits at the Fed, and they're all therefore reserves that banks have. And then start asking what happens, is those other kinds of liabilities become important. So the first case would be people are taking currency out of their banks. That's going to obviously reduce bank reserves. The second… and this is holding the size of the Fed's balance sheet constant. Second is a growth in the TGA where the Treasury puts more money in the Fed. But that means that there are less reserves in the banking system because they're taking away the liabilities of the banks and transferring them accounts from the banks to the Fed. And the last one is these foreign repo pool, which again whenever an institution keeps a larger balance at the Fed that's not a bank. It means the banks have to have less.

Selgin: Finally though, you have the case where the Fed is shrinking its balance sheet as it did during its unwind operation. And then of course its liabilities are going down along with the assets going down. So the whole balance sheet is-

Beckworth: Shrinking.

Selgin: ... getting smaller and there the decline in reserves doesn't have to take the form of a redistribution from reserves to other Fed liabilities. It's just that there are fewer liabilities in total.

Beckworth: Okay. So with quantitative tightening, the last case you just described where the Fed's reducing the absolute size of its balance sheet, my understanding is about $600 billion in reserves was eliminated.

Selgin: That's correct.

Beckworth: Disappeared out of thin air, or disappeared into thin air. So we saw about a $600 billion reduction in reserves during QT, quantitative tightening, from 2017 and 2019. Then on top of that… so that's the first part of the story. But on top of that, there's these other developments which I didn't really pay attention to, I think many people didn't pay attention to, and this is the shuffling of the categories on the liability side you just described. So walk us through why is it that when the Treasury increases its account at the Fed, there's fear of reserves? What's the actual steps?

Selgin: Well, to give a typical example, let's say the Treasury is collecting taxes. So people are paying their taxes to the Treasury, they're writing checks to the IRS, and of course those checks are checks against their banks, commercial banks. And the IRS in turn places these funds in the Treasury general account. Well, and then the checks clear. When the checks clear, the count to TGA is credited, finally credited. And at the same time, some commercial bank accounts at the Fed are debited. So the reserves, which are those commercial bank deposits at the Fed go down, and the TGA balance, which is not part of bank reserves, of course, because the Treasury isn’t a bank, goes up, and there you have it. So the amount by which the TGA increases is the amount by which the stock of bank reserves in the form of deposits at the Fed declines.

Beckworth: Okay. And that means there's fewer reserves for these banks to lend into the repo market when this happens. This is one of the reasons that there were fewer reserves available.

Selgin: That's right.

Beckworth: Let me ask about the history of this, because you wrote quite extensively in these two pieces that we're going to link to about the TGA account. It hasn't always been used. In fact, you have a great chart in the article that shows before 2008, before the switch to the floor system the Fed currently has, it was hardly used at all. Now if you look at it, it's used extensively and it's very volatile, big, big swings. So, what's the history of this account and why do we see this change today?

The Treasury General Account

Selgin: Well, first of all, it's important to realize that putting cash in the Fed is only one of several options that for some time now have been available for the Treasury, for its cash management. The Treasury is always going to have some cash resources on hand or some liquid resources on hand like any of us would, so that it is always prepared to cover its expenditures it keeps a balance. Now it keeps these resources in several forms. It can hold its own bonds, is can actually issue so-called cash management bonds, which are bonds it issues for the purpose of managing its own cash. It can put money in the Fed of course, it can also put money in commercial banks through the so called treasury tax and loan program, which in one form or another has been around for a long, long time, actually since before the Great Depression.

Selgin: Finally, there's another investment it can make, which is a longer term, basically term deposits. And that's called the Treasury Optional Investment Program. So you have a bunch of alternatives and depending on the circumstance, the Treasury, like any of us, is going to look at these alternatives and say, "Which is the most economical way for us to keep liquid funds. And as you said, for most of the time before the 2008 crisis, the Treasury found it economical to use resources other than the TGA for most of its cash balances or liquidity. In fact, for a long time before 2008, the Treasury had a policy that it would try to keep a balance in the TGA of only about $5 billion. That's a lot for you and me, but it's very small.

Selgin: Today for example, there's something like, I don't know, I haven't looked lately, but let's say $350 billion in the TGA. There may be more, maybe a little less, but it's around there. So that's a big change. So, of course that meant that in those days the Treasury was mostly using these other programs for its funds because even though it was smaller, it wasn't that much smaller. So it had a fair amount of money in the TT&L accounts, the Treasury tax and loan accounts. It also used the Treasury Optional Investment Program. And through a combination of that and also keeping its own cash management bonds on hand, it was able to keep that TGA balance low and stable. Occasionally it would, certain times of year it would go up more as much as maybe to seven billion, but it was stable and low, and that meant that the Fed's balance sheet could be that much smaller because the Fed didn't have to compensate for the fact that part of its resources were going towards providing the Treasury with cash or with a cash balance.

Beckworth: So the Treasury can either hold its cash outside the Fed and the private marketplace at banks, different storage facilities in terms of term deposits or overnight accounts, or it can hold them in the Fed and the Treasury general account.

Selgin: Yes.

Beckworth: And the decision to do so depends on the cost it generates to the Treasury. And part of the story, not the only story, but part of the story post 2008 is that it's been more cost effective to park the cash at the Fed than to store it outside the Fed and the private sector. Is that right?

Selgin: Yes, absolutely. The big change, the most important change, though there are other elements to the story of course, was interest on reserves, and it's a little bit strange why that matters. So let me explain. The Treasury's account at the Fed doesn't pay any interest. So you might think that the fact that the Fed introduced interest on reserves would mean that commercial banks now can pay more interest in theory on deposits and maybe that would be a better place for the Treasury to put its money. But it doesn't work that way because the interest that the Fed pays on the reserves of commercial banks is interest it could be paying to the Treasury. So when the Treasury puts money in TT&L accounts at commercial banks, it may get a little interest on those accounts, it does get some interest, but it's also losing the interest that is being paid on the bank's reserves, which would not be paid if it kept its money, which would not lose if it kept its money at the Fed.

The big change, the most important change, though there are other elements to the story of course, was interest on reserves, and it's a little bit strange why that matters.

Selgin: So, depending on conditions, but generally speaking, it is a money losing proposition with interest on reserves for the Fed to keep money in the commercial banking system instead of…

Beckworth: The Treasury.

Selgin: Pardon me, David, yes, for the Treasury to keep money in the TT&L accounts rather than just leave it at the Fed. So that's the big change.

Beckworth: So Treasury gets a bigger check from the Fed if it parks its funds at the Fed?

Selgin: Yeah. That's right.

Beckworth: I understand that appeal. That makes a lot of sense.

Selgin: The Treasury might get some interest from those banks, but then it gets so much smaller, a check from the Fed as it were, that it actually is worse off.

Beckworth: And there were a few other reasons as well. I mean, that's the main one, but it wanted to have a bigger account balance for emergencies, a lot of things going on post process.

Selgin: Yes, the Treasury at one point, 2015, it had already shifted a lot of its funds into the Fed, but in that year it decided that it wanted to maintain a balance of at least $150 billion at the Fed, up from the five billion we were talking about before. That's a huge change and its rationale for that was that in the event of another major crisis, it wanted to be able to have sufficient resources on hand at the Fed to cover all of its expenses for… I think it was for a week, without having to go to private markets under the assumption that anything it had outside of the Fed might not be available during a big crisis. That's a rather extreme assumption and there are ways I think you could deal with that concern.

Beckworth: Now, was there some role the Fed itself played in this? Did the Fed kind of maybe beforehand discourage the Treasury and then after 2008 say, "Well, whatever, we've got ample reserves." Because part of the story here is prior to 2008, the Fed had to actually actively engage and try to manage the amount of reserves based on what it thought would be the demand for reserves. But after 2008, if there's ample reserves in the system, why worry about it? Why not just let the Treasury keep as much as it wants at the Fed because as you mentioned, there's a consequence. If the Treasury puts more funds into its account, there's less reserves at banks, but no big deal if there's ample reserves. So the Fed kind of let its guard down. Is that right?

Selgin: Well, it did, but it thought that it knew what it was doing. So in the old system, which was a scarce reserve system, and of course you and I have talked about this, and you've written a lot about it on the corridors versus floors, in that corridor-like system, the Fed relied on very minimum bank reserves and fine tuning of the quantity of bank reserves to keep interest rates on target. But of course to do that, it was desirable that the Treasury should not have a large and volatile TGA balance, because that would interfere with the quantity of reserves causing it to also be volatile, unless the Fed very actively intervened in the market. It would make the job of fine tuning the reserve supply much harder for the Fed. So there was certainly an understanding between the Treasury and the Fed that the Treasury should try to manage its TGA balance in the interest of making life simple for the Fed and allowing it to operate that system without too much difficulty.

Selgin: When they switched to an abundant reserve system in 2008, it did become less desirable from the Treasury's point of view to keep money outside of the Fed as I explained, but it also became less important for it to do that from the Fed's point of view because under an abundant reserve system, the Fed doesn't worry about small changes in the reserve supply and doesn't rely on such changes to keep rates at their target level. Instead, it just sets the IOER rate, the interest on excess reserve rate, and that's how it controls other rates.

Selgin: So at that point, with that change, the Fed was happy to say to the Treasury, "Oh yeah, sure. Put all the money here you want. Do what you want, increase it, reduce it. It's no problem for us." For example, it might be, let's say a Treasury withdrawal of $140 billion from a TGA account that had that much and more in it originally under the old system. That would have been a huge shock to the quantity of bank reserves and the Fed would have had to come in there and make up for it by buying more securities. But circa 2015 let's say, that would have been a very small percentage of the outstanding amount of excess reserves. And so it wouldn't have mattered that much.

Beckworth: Yeah. And that's what the Fed thought and it was true for a long time up until recently when that complacency came back to bite the Fed in the rear. And we'll get to that because we'll see the TGA got so big relative to the demand for reserves it actually became an issue again. But before we get into that, we have a few more things we've got to work through, and I do want to highlight something else you bring out in your paper about the TGA account, and it was really interesting some of the history, and that the Fed and the Treasury had in similar issues they do today back in the '70s. So, the Treasury today is saving money by putting, as you mentioned, funds into the TGA, given interest and excess reserves, but in the '70s, it also resorted to that tactic because of double digit inflation and double digit interest rates. So tell that story because it's really interesting.

Selgin: Yes. So I think it was 1978 and basically the relationship between the interest rates being paid on the market. And I remember this was when interest rates that banks could pay, were still controlled by Regulation Q and related regulations and the rate that was being paid ... Sorry, those rates became sufficiently unattractive that the Treasury found it desirable to put money back in the TGA account and start using it heavily again, which it hadn't done for a long time. And in the end this had to do partly with the interest rates that the TT&L accounts were yielding.

Selgin: So, what ultimately happened was they worked out a new arrangement. They created a special kind of TT&L note account that bore more interest as a way to get the Treasury to once again use the TT&L program as it had done before. And the significance of this episode is that it was very similar to what happened recently in that the Fed found that because the Treasury was using the TGA account heavily again, and that account had become volatile, the Fed was having to intervene in the reserve market much more heavily to keep rates on target and was really having a hard time. So it had to work with the Treasury to come up with a way to fix that, which for a while it managed to do. So that's what it needs to be thinking of doing today. And for some odd reason, there's no talk that I'm aware of from the Fed or the Treasury about, "Hey, let's get together and-"

And the significance of this episode is that it was very similar to what happened recently in that the Fed found that because the Treasury was using the TGA account heavily again, and that account had become volatile, the Fed was having to intervene in the reserve market much more heavily to keep rates on target and was really having a hard time. So it had to work with the Treasury to come up with a way to fix that, which for a while it managed to do. So that's what it needs to be thinking of doing today.

Beckworth: Let's coordinate.

Selgin: "... figure out how to make these substitutes for TGA balances viable again."

Beckworth: Has there been a lot written on this episode from the '70s that's very similar to today?

Selgin: I don't know about a lot, but there's certainly a few sources out there on it. And I do link to one or two in my post.

Beckworth: Okay. So this would be a great area for research for some enterprising young scholar or current scholar, someone in the money markets group at the Fed to take a look at this period compared to the present. Okay. So the TGA, to summarize, is being more heavily used by the Treasury. At first, not a big deal given the ample amount of reserves, but we have seen it now is a big deal, like in the '70s, and it's causing problems for the Fed's management of its balance sheet. In fact, that's kind of a key theme. We're going to talk about the TGA and we're going to move on to the foreign repo. But in both cases, I mean, the striking point is the Fed has a sizeable portion of its balance sheet that is not under its control and is volatile.

So the TGA, to summarize, is being more heavily used by the Treasury. At first, not a big deal given the ample amount of reserves, but we have seen it now is a big deal, like in the '70s, and it's causing problems for the Fed's management of its balance sheet. In fact, that's kind of a key theme.

Beckworth: Now, currency has always been there. So currency has been beyond the Fed's control. It grows at a certain level. It's predictable.

Selgin: It's very predictable. Yes, very predictable.

And so part of the challenge we're trying to show here is that the Fed's having a hard time managing its balance sheet because there's these items that are beyond its control.

Beckworth: Very predictable. And it's relatively stable. Whereas these other two accounts have been explosive and volatile. And so you can just imagine your own balance sheet. I mean, imagine trying to plan your activities when there's a sizeable portion of your balance sheets, your liabilities are going to swing one direction to the other, be tough to make plans, be tough to implement what you want to do. And that's what the Fed's facing. And so part of the challenge we're trying to show here is that the Fed's having a hard time managing its balance sheet because there's these items that are beyond its control. So let's go to the second item and that's this foreign repo pool. So tell us what that is and what's happened to it recently.

The Foreign Repo Pool

Selgin: The foreign repo pool was created, I think it's close to 50 [ago] years now, but anyway, many several decades ago, and it's a facility, again, at the Fed designed for foreign official institutions. So that includes central banks, but also some of the international institutions that work with central banks. And it was designed to be a safe place for these foreign entities to place their funds, dollar funds. And it has its counterparts, it should be said, in like facilities provided by foreign central banks to the Fed for example. So, it's part of a reciprocal set of arrangements. However, of course, reciprocal in this case doesn't mean exactly symmetrical because the dollars, they're so much more important. So there's a lot more need for foreign central banks to have a place to park dollars than there is for the Fed to have a place to park most other kinds of… park its holdings of foreign exchange. So, that's the basic setup.

Selgin: Here once again though, the Fed's facility is a substitute or a supplement to private market alternatives, right? So the foreign central banks don't have to put dollars in the Fed. They can use private repo markets. They can, in principle, they could put funds in commercial banks, they can hold dollar securities, treasury securities-

Beckworth: Outside the Fed?

Selgin: All of these would be outside of the Fed.

Beckworth: Outside the Fed, yeah.

Selgin: All of these are alternatives for holding dollars or for depositing dollars or investing dollars that don't drain reserves from the US banking system. That don't mean an increase in the Fed's non-reserve liabilities. So here again, the question that foreign official entities [and] institutions face in this context is, "Okay. What is the best place for us to park dollars?" And that's going to depend on considerations of risk certainly, but also on what the return the foreign repo pool pays compared to other alternatives. And here I think the biggest problem is simply that the Fed, which sets the foreign repo pool rate, has made it too attractive. And it's the solution in this case, and I know we're going to get to solutions more detailed later, but it's relatively simple for the Fed to adjust that rate if it wants to and make it less attractive relative to alternatives so that the foreign repo pool doesn't get used to such a vast extent as has been the case lately.

And here I think the biggest problem is simply that the Fed, which sets the foreign repo pool rate, has made it too attractive. And it's the solution in this case to adjust that rate if it wants to and make it less attractive relative to alternatives so that the foreign repo pool doesn't get used to such a vast extent as has been the case lately.

Beckworth: And to be clear, its use, again, went up dramatically post 2008 compared to pre. So I'm sure part of the story here is also the Fed got a little relaxed again for the same reason they got relaxed about the TGA account, because ample reserves, not a big deal if these foreign central banks are depositing at the Fed, but it has created problems along the same dimension, same channels as the TGA and Zoltan Pozsar, a well-known analyst who looks at these issues, he's called it the black hole of the money market. So, why does he call it the black hole of the money markets?

The “Black Hole” of Money Markets and the “Accidental” Corridor System

Selgin: Well, the easiest way to think of that is you have a Federal Reserve that of course by expanding its balance sheet and buying securities adds to the supply of reserves. But to the extent that the foreign repo pool grows, and it has grown tremendously since 2008. It's like a black hole where those reserves instead of staying as bank reserves get-

Beckworth: Sucked in, okay. Yeah.

Selgin: ... sucked out of the banking system. But what that really means is that you have formerly, you have what were once Fed reserve liabilities being switched for non-reserve liabilities. So you can have this situation and Pozsar is very good about describing the circumstances in which the Fed's efforts to expand the supply of reserves by buying securities just end up pouring that many more reserves down the foreign repo pool drain. And it becomes a futile thing.

Selgin: So there are two different issues here, but they're related. One is that the absolute size of these non-reserve Fed liabilities, the foreign repo pool and the TGA, the bigger they are, the fewer bank reserves you have given the size of the Fed's balance sheet. Therefore, the bigger they are, the bigger the Fed's balance sheet has to be to create a given total reserve pool. But the other one is the volatility, and in this case the volatility is much worse for the Treasury general account. That thing really has bounced around dramatically. The foreign repo pool has ratcheted a lot, so you'd see a little bit of fluctuation all the time, but you also have seen several big steps.

So there are two different issues here, but they're related. One is that the absolute size of these non-reserve Fed liabilities, the foreign repo pool and the TGA, the bigger they are, the fewer bank reserves you have given the size of the Fed's balance sheet. Therefore, the bigger they are, the bigger the Fed's balance sheet has to be to create a given total reserve pool. But the other one is the volatility, and in this case the volatility is much worse for the Treasury general account.

Selgin: In any event, the volatility of these non-reserve liabilities, that particularly poses a problem as you get to a scarce reserve situation. So to tell the whole story, think about it this way. As these non-reserve liabilities grow, total reserves have been shrinking, other things equal. Add to that the Fed’s unwind, of course, they'll shrink even more.

Selgin: At some point they shrink to the point where reserves are no longer abundant. That is they're reaching a point where some banks are feeling shortages. Then the volatility means that you're going to have some interest rate action out there in the money market that's independent of the Fed's interest rate settings. Interest rates are going to take on a life of their own and the Fed is going to have to go in there with repos or whatever to try to stabilize those rates and keep them down as if we were back in a corridor system. But it's the worst kind of corridor system because it's an accidental corridor system.

Beckworth: No, no. That's the way I describe it in an article I wrote.

Selgin: Did you?

Beckworth: The Fed kind of stumbled or tripped back into a corridor system.

Selgin: It kind of, yes.

Beckworth: And now it's trying to get back out into a floor system.

Selgin: Yeah. It's not an argument, and I'm sure, I want to stress this point, both of us… it's not an argument against the corridor system. It's an argument against an accidental corridor system. And if you're going to have a corridor system, you can do it well or you can do it really badly. This is bad. This is very bad.

It's not an argument, and I'm sure, I want to stress this point, both of us… it's not an argument against the corridor system. It's an argument against an accidental corridor system. And if you're going to have a corridor system, you can do it well or you can do it really badly. This is bad. This is very bad.

Beckworth: Yeah. As I like to say, and I've mentioned this in the show many times, and people I talk to I've mentioned this, but Canada had a corridor system. Like the Fed it went to a floor system during the crisis, but then it successfully went back to a corridor system. So it is possible to have a smooth transition. Now they have some things that are different up there, so maybe it's not like a fair comparison, but there are ways, and we'll actually maybe get at the end, but we're really, again, working within the framework or the assumption that we have a floor system, how can we help the Fed manage its balance sheet in a practical, easier way?

Beckworth: Because again, the challenge of these last two categories, the TGA and then the foreign repo pool is, again, you've got a sizable portion, a meaningful portion of the Fed's balance sheet and a liability size that is exogenous, kind of beyond the control of the Fed, at least directly, and it creates real problems for the Fed in terms of managing reserves. And why does that matter? Because that then influences funding available to the repo market. And the repo market is hugely important to funding the financial system. It's how they fund activity on the longer end of the curve.

The challenge of these last two categories, the TGA and then the foreign repo pool is, again, you've got a sizable portion, a meaningful portion of the Fed's balance sheet and a liability size that is exogenous, kind of beyond the control of the Fed, at least directly, and it creates real problems for the Fed in terms of managing reserves. And why does that matter? Because that then influences funding available to the repo market. And the repo market is hugely important to funding the financial system. It's how they fund activity on the longer end of the curve.

Beckworth: Now, speaking of curves, one other question I have that's been brought up a lot, and that is that the inverted yield curve also made the foreign repo pool attractive relative to buying other treasuries, other long-term securities. So, having an inverted yield curve, this kind of tightened the screws of this problem, is that right?

Selgin: Yes. Well, it means that short-term funds are more attractive than long-term funds. And that means that as far as both the Treasury and the foreign official institutions are concerned, they're not thinking of longer term securities as a good place to vehicle for maintaining liquidity, and they're looking towards instead various kinds of deposits. And that's narrowing the field and it's narrowing the field in a way that makes it more likely that they're going to be holding funds, keeping funds at the fed.

Beckworth: And the Fed does have some influence over that.

Selgin: Yes, it did. Yes.

Beckworth: Okay. All right. We'll go beyond that. So here we are, we've discussed the Fed's balance sheet, the liability side. We've discussed these categories that are kind of beyond the Fed's control. I mean, really currency growth is beyond the Fed's control, but it's lived with that for a long time. It's just these other recent categories, the change, the volume, the volatility, and just to make this concrete, the Fed's shrinking of its balance sheet, quantitative tightening, reduced reserves by about $600 billion as we mentioned earlier, but the growth in the Treasury General Account and the foreign repo pool combined is a little over $600 billion too. So the magnitudes are very similar, just as big as the QT is, these other accounts have contributed as well. So this is a nontrivial development.

Selgin: Yes, that's right. It's nontrivial and your way of describing the sense in which it’s so is quite correct. The other way to think about it is that if you look at the amount of money parked in the TGA and foreign repo pool today combined, the Fed by reducing, taking steps with the Treasury's helps perhaps to reduce those accounts, those balances to small amounts, could create more reserves that way, than it's going to be creating in the next half a year or so by buying more securities again. And that means that if it wanted, in other words, if it wanted to create $600 billion in fresh reserves, it could do so without expanding its balance sheet or buying any more securities if you could just get those other liabilities back down again.

Beckworth: And just so we're clear on the numbers right now, the amount of reserves in the system are a little over one and a half trillion, and just to kind of do a back the envelope calculation, what if there had been no quantitative tightening? What if the TGA hadn't grown? What if it had maintained a kind of pre 2008 level? And what if there hadn't been any growth in the foreign repo pool? Then reserves would be up around 2.6 trillion. So, I mean that would have been presumably more than enough to maintain the ample reserve, the floor system that's in place. So again, the issue is these things have grown and the story now in September is closely tied to these and that there were big corporate tax receipts coming in, and the government was issuing a bunch of debt, which meant TGA was ballooning. The yield curve also contributed to the foreign repo pool growth.

Beckworth: So kind of the story there is reserves are shrinking both from quantitative tightening and from these other kind of exogenous developments beyond the Fed's control. So kind of the Fed was shifting its supply curve back, but it was trying to shift it back not too far, just far enough so it could reduce the balance sheet without becoming a corridor system. Again, that's accidental falling into it. But that coinciding with increased demand for reserves from regulatory reasons kind of pushed it in there.

Beckworth: Now looking forward to December, we know the Fed is earnestly trying to increase the amount of reserves in the system and it's quite hard to add up everything it's doing. But Jim Bianco, who does market research, he has an estimate out there that all combined, all these different measures, and again, to be clear, some of these are overnight, their loans will be liquidated, and in the future some of them are outright purchases. But combined, by the end of the year, there'll be about $300 billion plus injected. And the question is, will that be enough? And it's not clear we know the answer to that question, is it?

Selgin: No, it's not. And I know some very smart people who believe that a heightened reserve demand will, toward the end of the year, despite these additions by the Fed, will cause repo rates once again to climb up possibly above the Fed funds market upper limit, and that the effective Fed funds rate may also rise above its upper limit, which of course defines a failure of monetary control in our system.

And I know some very smart people who believe that a heightened reserve demand will, toward the end of the year, despite these additions by the Fed, will cause repo rates once again to climb up possibly above the Fed funds market upper limit, and that the effective Fed funds rate may also rise above its upper limit, which of course defines a failure of monetary control in our system.

Beckworth: Right. And in fact, one of the arguments for the floor system is we'll have much better interest rate control, and we've seen anything but that. And I sympathize with the Fed…

Selgin: It's been a long practice period too.

Beckworth: Yeah, I mean, again, compared to the Canadian corridor system, we're doing a pretty horrible job down here.

Selgin: Or our pre 2008 system, which had a much better record of the Fed despite the extra work involved, because in that system it was expected that the open market desk would be busy all the time. They had to predict, they had to do a lot of work. And one of the big arguments was we'd go to a floor system and everybody at the open market desk they'll have their feet up on the desk and they'll be telling jokes, and-

Beckworth: Be on Twitter.

Selgin: ... tossing paper airplanes. But in fact, they've been very busy, as we know, and still they have failed in a number of remarkable instances to get rates to behave the way they want them to. So that's not very good.

Beckworth: Yeah. So the issue then going forward, the end of this month, will be will there be enough reserves? And I guess the way I think about it is there's a potential outward shift in the demand curve for reserves, and it's going to arise because of these big banks having year end regulatory inspections. So they call them the global systematically important banks, the G-SIBs, and there's a capital surcharge that's applied to them. And my understanding is end of the year, they try to tidy up their balance sheets, make them smaller, leaner, so the charge they get will be less. But in doing so, these banks will lend less into the market. So effectively the demand for those reserves will have increased. So, even as the Fed is increasing supply of reserves, that demand curve might be shifting out going beyond where the Fed is.

Selgin: That's right. And it's important to remember that the demand for reserves of these large banks is, an increase in that demand, is equivalent to a reduction in the availability of reserves to be lent in the repo market by those banks, because those banks are holding most of the reserves. I don't have the recent percentage immediately on hand, but we know that they hold an outsized percentage of the total available stock of excess reserves. So if there's a reserve shortage anywhere in the system, and those banks are not in a lending mood, so to speak, or are not in a position to lend, then that's when you get these desperate bids by people trying to pay whatever to cover themselves because it's hard to find anybody who's got reserves who wants to part with them.

And it's important to remember that the demand for reserves of these large banks is, an increase in that demand, is equivalent to a reduction in the availability of reserves to be lent in the repo market by those banks, because those banks are holding most of the reserves.

Beckworth: Yeah. So these banks effectively will reduce the supply of funding available to the repo market. For them it'd be an increasing demand, but from the repo market perspective, reduction in the supply of reserves.

Selgin: That's right, and it can also affect the Fed funds market too, and that's why those rates could ...

Beckworth: Yeah. All the overnight markets.

Selgin: All the overnight rates.

Beckworth: But we see kind of a headlines coming from the repo market, because that is probably the most important overnight funding market.

Beckworth:  And Zoltan Pozsar makes this point that it's one thing to have maybe a temporary spike in the repo market, but if it continues to persist, then you're leading to actual credit problems, actual businesses going bust, funding drying up for longer term projects. So it becomes serious. I mean, this is a technical issue right now, but a technical issue can turn into a real credit issue if it's not fixed.

Selgin: Sure. Yeah. This is definitely serious. This is not just about wanting to make the Fed look better. This is about the wellbeing of many ordinary participants in those overnight markets.

Beckworth: Okay. So that's the problem. And the concern again is, we might accidentally fall back into a corridor system at the year end, rates will spike. All kinds of problems will emerge from that depending on how long it lasts. So the solution that's been presented by one of our previous guests, David Andolfatto and Jane Ihrig, who's at the Board of Governors, and it's received a lot of discussion and discussed elsewhere extensively as a standing repo facility, which would allow these big banks to let go of some of the cash reserves they're holding because they could take treasuries to the Fed and easily convert that into cash. So right now they could be holding treasuries instead of cash, but they think they may not be able to liquidate quickly. There might be losses if they try to do it quickly, but if the Fed were there, it would solve that problem.

Beckworth: But you bring up in your paper a way to improve the standing repo facility because what the standing repo facility fixes right now as proposed would only address these big banks. It doesn't address the problems we've discussed with the TGA or the foreign repo pool directly. And you have some additions to make it kind of an all-encompassing solution. So walk us through that.

How to Improve Proposals for a Standing Repo Facility

Selgin: Well first of all, David, the standing repo facility is conceived by Andolfatto and Ihrig, it is meant to be there as not as a substitute for these ad hoc repos that the Fed's been doing. But it's not a substitute for the Fed's security purchases, which are also going on now, which are meant to make permanent additions to the stock of reserves to get back to an abundant reserve situation. So the idea is you pursue those reserve creating steps by the Fed, and then you have the standing repo facility, which will help because it reduces the demand for reserves by allowing banks, big banks especially, to treat treasury securities as closer substitutes, as better perfect substitutes as it were, for actual excess reserves.

So the idea is you pursue those reserve creating steps by the Fed, and then you have the standing repo facility, which will help because it reduces the demand for reserves by allowing banks, big banks especially, to treat treasury securities as closer substitutes, as better perfect substitutes as it were, for actual excess reserves.

Selgin: Now, I should mention there that this only works if the supervisors and the regulators go along with it, right? They have to say, "Okay. You can treat treasuries as perfect substitutes for excess reserves," which Basel allows in its definition of high quality liquid assets, tier one, high quality liquid assets, but which I understand and several people understand is not necessarily satisfying the regulators, particularly with regard to the so-called living will liquidity requirements.

Selgin: So anyway, assuming that David and Jane's facility exists, it would both reduce the supply or the demand for reserves, and thereby make it possible for the Fed to increase the supply of reserves and be done with it and never have to do that anymore. That's the idea. And if it's working the way it's supposed to, it shouldn't actually be used very often. You know, there might be some rare instances where reserves are a little scarce at the margin and then it would kick in.

Beckworth: Okay. So it would stand available, it probably wouldn't be used much, but it would create this certainty so you wouldn't want to hold a bunch of precautionary reserves?

Selgin: Exactly, yeah.

Beckworth: So it's kind of a nice backstop that hopefully you don't have to ever use?

Selgin: Yeah.

Beckworth: And presumably it would overcome the stigma of the discount window. I mean, the discount window in some sense it's very similar what it would provide, but people don't want to use the discount window because of the stigma. So, this would be a work around to that solution as well. But again, it wouldn't solve this problem of the Fed's balance sheet being beyond its control and the TGA and the foreign repo. But you have a fix for that.

Selgin: Yeah. So, the thing is that what you want to do is to make the demand for reserves lower but also more stable, and make the supply of reserves more stable, and the standing repo facility helps. But to really, really do that and to achieve what has long been the Fed's supposed idea of a balance sheet as small as is consistent with carrying out its monetary operation successfully, steps should be taken in my opinion to rein in the TGA and to reign in the foreign repo pool. And so I propose a number of those.

Selgin: For the foreign repo pool, that's pretty simple because the Fed is completely in charge of that facility. It is a facility that it operates as a favor for foreign banks. It's not in the Federal Reserve Act. It's not obliged to operate it and it can certainly operate it any way it wants to, within reasonable limits. It should set that foreign repo pool rate lower than it does. I suggest in my paper that it take the SOFR rate, which it's probably now using as a guide to the foreign repo rate and set the foreign repo rate something like 10 basis points below and make it a little bit of a penalty rate relative to private market rates. SOFR rate, if anything, is a little bit above private market repo rates because it's an index that includes a broader set than the usual private market rates. So that's the one thing it could do.

I suggest in my paper that it take the SOFR rate, which it's probably now using as a guide to the foreign repo rate and set the foreign repo rate something like 10 basis points below and make it a little bit of a penalty rate relative to private market rates.

Selgin: It could also put caps on the foreign repo pool, it used to have caps. It could tell the foreign banks, "Look, you can only hold this much. Here's your maximum balance. Here's how much you can change it with any limited period of time. There's plenty of scope for the Fed to work on these agreements.

Beckworth: And it used to do that because it was a scarce reserve environment.

Selgin: That's right, yeah.

Beckworth: Okay.

Selgin: But the point is, and maybe this is a big takeaway from this, even if you have an abundant reserve environment, still there are reasons to not have more reserves than necessary to manage that system. You still want it to be efficient in that sense. And that has always been the Fed's principle. That's why it had the unwind in the first place. Otherwise, why didn't it just stay with the maximum amount of reserves it had way back then? So, if that's the principle, and I think it's a sound principle, well then these steps should be taken because they allow the Fed's balance sheet to be that much smaller within a floor system framework. And though they also serve as stepping stones for getting to an efficient corridor system, which is important, and maybe we'll get to that. But in any event, if the Fed is serious about keeping a small balance sheet as small as possible, consistent with its monetary operations being successful, this is a step towards doing it.

If the Fed is serious about keeping a small balance sheet as small as possible, consistent with its monetary operations being successful, this is a step towards doing it.

Selgin: And the other thing I recommend for the Treasury is, actually, what I recommended, one of the things I recommended is something it turns out the Treasury already is able to do. There was a program experimented with by the Treasury starting in 2006 called the Treasury Repo Program. It was a direct repo borrowing program so the Treasury could go into the repo market, and that was a good substitute for keeping balances at the Fed. It earned a higher return and it obviously earned the private repo rate return for the ... which is more than the TT&L accounts. And turns out that in fact, in an obscure law that had to do with adoptions in 2008, I didn't know about this when I wrote about it at first, that tentative program, that experimental program was in fact made permanent. But something like three weeks after that law was passed, the crisis hit, interest on reserves hit. And so-

Beckworth: People forgot about it, huh?

Selgin: ... they forgot about it because remember, for a long, long time, private repo rates were below the interest rate on reserve. So it was costly even with that to take advantage of that program for the Treasury. Now, that program should be-

Beckworth: It makes sense.

Selgin: Makes sense. Apart from, and this is the hitch, and it's also a bit of a hitch for the foreign official institutions, the hitch on all that I've said is you still have the contingency of a possible major financial crisis. And in those circumstances, these alternatives I've been proposing might not be fully attractive. So the Treasury might worry about using its direct repo purchases because there could be a shutdown and that's why it would still want to have, unless something else has done $150 billion in its TGA.

The hitch on all that I've said is you still have the contingency of a possible major financial crisis. And in those circumstances, these alternatives I've been proposing might not be fully attractive.

Beckworth: Okay. So the private sector may not be able to fund Treasury?

Selgin: Yeah, if the private sector completely breaks down of course, it's very hard to make up an alternative for the keeping money at the Fed that can compete. That's true also for the foreign repo pool. But, this is where I propose my very-

Beckworth: Radical.

Selgin: ... radical-

Beckworth: I like radical George, so go ahead.

Selgin: And some will say non-free market, non-libertarian or whatever solution, not that that kind of thing bothers me too much. And that is simply this, let the Treasury and the foreign official institutions that presently contribute to the foreign repo pool, let them be counterparties to the proposed standing repo facility. I understand that there are a lot of laws that have to be worked out and the details, let them be counterparties, impose penalties or haircuts on them. Haircuts are essentially penalties, so it's not too attractive for them to use these facilities, but the facilities are there and can reassure them so that they have no reason, so that the Treasury has no reason to have always $150 billion or more in its TGA account. In the worst possible scenario where it's been repo-ing, it can take advantage of the standing repo facility if everything dries up for the foreign official institutions.

Let the Treasury and the foreign official institutions that presently contribute to the foreign repo pool, let them be counterparties to the proposed standing repo facility. I understand that there are a lot of laws that have to be worked out and the details, let them be counterparties, impose penalties or haircuts on them.

Selgin: Now of course the complaint will be, "Why do we want to make last resort lending facilities of any sort available to foreign firms?" Is one complaint, or, "How can we possibly risk letting the Fed directly lend to the Treasury? We know how dangerous that is," et cetera, et cetera. And none of them are very compelling. Those arguments aren't very compelling because the penalty rates, these emergency facilities aren't going to be abused. And with respect to the Treasury, there's really no difference here between the Fed buying securities, which you'd have to do, or the Fed directly lending to the Treasury offering to do that.

Beckworth: So this would cut out the middleman in the middle of a crisis.

Selgin: In the middle of a crisis you cut out the middleman, but you're not changing the fiscal situation.

Beckworth: I think your point is in normal times, this facility wouldn't be used by banks, by the Treasury, by these foreign official entities.

Selgin: You could certainly structure it so that there's absolutely no way for them, you can make that part of the law even, "If there's not a crisis going on, forget it."

Beckworth: But in a crisis, the Treasury, banking system, foreign entities would come knock at the door, the standing repo facility, and have access to it. I mean, I think another argument against the critics of this would be if you want the dollar to maintain its status as a reserve currency of the world, then you have to open the door of the standing repo facility to the rest of the world. I mean, it's part of the price we pay for making the dollar so wide reaching and powerful.

Selgin: Perhaps that's right. But the point is there are right and wrong ways to manage a last resort-

Beckworth: Fair enough.

Selgin: ... lending facility in constitutions. We do need constitutional constraints. It's not clear that the right constraints are let's not have the Treasury under any circumstance ever be able to borrow directly from the Fed. And in fact, as I mentioned in my paper, until 1981 for many decades, the Treasury did have so called a direct draw authority, and it used it routinely. And it used it by the way to stabilize the TGA balance so it wouldn't be trouble for the Fed to manage monetary policy. It was exactly for that purpose, here it's a little bit different, but the same basic principle. And so it's been done before and it's a question of doing it again under a somewhat different operating system and with a different facility. But otherwise, there's nothing new here under the sun.

Beckworth: So the Treasury used to have access to the Fed through its security draw.

Selgin: Yes, or direct draw authority.

Beckworth: Direct draw authority. So it's not entirely new as you mentioned.

Selgin: Not unprecedented.

Beckworth: And again, going back to the point of this discussion, this is part one, part two will be next week. We're nearing the end of the show here. The whole point, or the big point here is that if these changes were implemented, the Fed could control its balance sheet a whole lot more effectively.

Let the Treasury and the foreign official institutions that presently contribute to the foreign repo pool, let them be counterparties to the proposed standing repo facility. I understand that there are a lot of laws that have to be worked out and the details, let them be counterparties, impose penalties or haircuts on them. And [it] could have a much smaller balance sheet.

Selgin: And it could have a much smaller balance-

Beckworth: And [it] could have a much smaller balance sheet. So yeah, for those who are concerned about the Fed's footprints, that actually would shrink it. But for right now, I guess the practical concern right now is the Fed is having a hard time controlling its balance sheet, a good portion of it, and that's the Treasury General Account and the foreign repo pool. And this would provide a way to kind of fix that problem. Very pragmatic, very practical, with other great benefits down the road of a smaller footprint as well. And I think that's a great, great point. So it may seem radical, but I think it's very pragmatic.

Possible Return to a Corridor System?

Beckworth: Yeah. So this is a great proposal, George. I'm glad you've written about it. I encourage our listeners to take a look at the pieces. It's a two part piece George has posted. So we've talked about how this will be useful. It's something that Fed officials should think about, because I know they're talking about the standing repo facility right now. So they should consider these other changes in concert with maybe consulting the Congress, but in addition to the pragmatic fixes for the here and now, this could also be a stepping stone down the road to a return to a corridor system. So this is more long thinking, future thinking. So walk us through that scenario.

Selgin: Well, it's actually pretty simple, David. Once you have all these reforms in place, if they're able to do that, then as I said, the Fed could operate a floor system with a much smaller balance sheet, but it also would be much easier for it to return to shrinking its balance sheet to the point where it's back in a scarce reserve situation because having tamed the TGA and having tamed the foreign repo pool, it could manage the needed open market operations to operate a corridor system again.

Selgin: If you read the statements of Fed officials about why they don't want to go back to a corridor system or don't think it's possible, they talk about the TGA, and the foreign repo pool, and these liabilities… would be so hard. So, solving that problem of how those liabilities behave, containing their extent and volatility would eliminate the main rationale given by the Fed for not going back to a corridor system… would make doing so much, much easier. But that's what I would ultimately like to see happen, of course.

So, solving that problem of how those liabilities behave, containing their extent and volatility would eliminate the main rationale given by the Fed for not going back to a corridor system… would make doing so much, much easier. But that's what I would ultimately like to see happen, of course.

Beckworth: And I share that view as well. Well, with that, our time is up. And again, I encourage our listeners to tune in next week when Josh Galper from Finadium, who's in the heart of the repo market, will talk to us on the same topic from the perspective of the repo market. But George, thank you for coming on and a happy holiday season to you.

Selgin: Thank you, David. You too. It's always a pleasure.