BONUS: George Selgin on the Fed Taper and Shrinking the Fed’s Balance Sheet

George Selgin is a senior fellow and director emeritus of the Center for Monetary and Financial Alternatives at the Cato Institute and is also a long-time returning guest of Macro Musings. In this bonus segment from the previous conversation, George rejoins the podcast to talk about the Fed’s near-term plans to shrink its balance sheet, the impact of the standing repo facility on demand for reserves, the potential benefits of returning to a corridor operating system, and more.

Read the full episode transcript:

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

Beckworth: This is the bonus segment for this week's show. We ran out of time, we want to honor your time and therefore we limited the first part of the program to the normal hour. But I wanted to keep talking with George about the Fed's balance sheet, because as all of you know, we are in the midst of a taper right now, the Fed is winding down its asset purchases. And it's also made very clear it's going to start shrinking its balance sheet maybe as soon as March. And we know the Fed's balance sheet has a lot of short duration, a lot of shorter term assets on it, which means if it just lets it run off, it could run off really quickly, much faster than last time. And so there's lots of questions surrounding this, will it be smooth, will it make a difference on the economy? And I want to hear George's take on this development. George, what say you?

On the Fed Shrinking its Balance Sheet

Selgin: Well, David, first of all, I'm in favor of seeing the balance sheet shrink again. First of all, it had always been the Fed's promise that large scale asset purchases would be temporary, that is, would involve temporary increases only in the size of its balance sheet, that they would eventually be undone, that is, balance sheet would be unwound. That's not inconsistent with maintaining an abundant reserve system because following episodes of large scale asset purchases, the Fed ends up typically with a lot more reserves in the system than are needed to achieve abundance, the necessary level of abundance required to run a floor system. Now I'd also like to see the floor system go away and have that much smaller balance sheet. But I think the Fed certainly should live up to its commitment to not have its balance sheet be any bigger than necessary.

Selgin: Now back in 2019, it was normalizing the balance sheet belatedly after the Great Financial Crisis. And it had promised to do that again because the idea was it was never supposed to keep this balance sheet blown up forever. And it ran into trouble, as I think almost every listener to your program will know. We started getting bad behavior of short-term interest rates in the Fall of 2019 with spikes here and there. Evidently shortages of reserves were popping up here and there in the financial system and the Fed got cold feet. And that was the end of the balance sheet unwind. And of course we know what ultimately happened. It settled down at about a $4 trillion and then the next crisis brought it all the way up to twice that amount. Now they want to bring it back down again, we don't know how much, but there's a new situation, a new factor in the story. And that's the standing repo facility.

I'm in favor of seeing the balance sheet shrink again. First of all, it had always been the Fed's promise that large scale asset purchases would be temporary, that is, would involve temporary increases only in the size of its balance sheet, that they would eventually be undone, that is, balance sheet would be unwound.

Selgin: Now the standing repo facility, which our friend David Andolfatto has had a lot to say about, was an early promoter of, that is designed to prevent the sort of thing that that happened back in September 2019 from happening again. So it's in place and if it works the way it's supposed to, the Fed should this time around be able to unwind for some time since we've got so many reserves in the system that before it has to stop. And so with luck, we should at least be able to see the whole thing, come back down to four trillion again or something like that, and probably could see it go further.

Beckworth: Yeah. I want to come back to that point about it might be able to go below four trillion. I think there's several reasons for that, but going back to the problems that it had in 2019, you mentioned the repo crisis of September of that year and the way I interpret this, and I think this is what you were alluding to, is the Fed accidentally, inadvertently fell back into a corridor system. It didn't realize the amount of reserves needed, which was larger than it thought to be in an ample reserve floor system and it was the perfect storm. They were doing QT, shrinking the balance sheet, the US government was pulling in tax revenues, the TGA was growing, so a lot of things came together, this perfect storm, they're like, whoa, we fell back, we need to change what we're doing. And a lot of that was due to regulations too, what banks had to hold. So going forward, the standing repo facility should take some of that pressure off, if I understand you correctly, and I think I agree with that also, it looks like the supplemental leverage ratio show is going to be tweaked too.

Selgin: Yes, yes. That's right.

Beckworth: And that should have a bearing as well. So my hope is that the regulatory hindrances that maybe caused the incident in September, 2019 won't be a binding constraint because of these new innovations, the standing repo facility and the tweaks to the supplemental leverage ratio going forward.

Selgin: Yes, that's right.

Beckworth: Okay. Here's another reason I think we can shrink the balance sheet down rapidly and to a large degree, maybe four trillion, like you said, that's a big number. And the one thing I am concerned about, I will say, is signaling, do the markets get worked out, but at least theoretically, let me lay out my case. We already mentioned, the standing repo facility and the tweak to the supplemental leverage ratio, but I want to pull out this whole notion that QE actually matters that much to the real economy. And I can go back, and we had some previous guests on, we discussed Wallace neutrality, which says basically outside of a crisis period and especially at a zero lower bound, these large scale asset purchases really don't make much difference, you're just changing the debt structure of the US economy. So I think between those two factors, Wallace neutrality and the tweaks to the structure, and repo facility, supplemental leverage ratio, I think we could have a relatively uneventful journey back down to a smaller balance sheet.

Selgin: I think so, too, David. I hate to be optimistic, it's always dangerous. But I think all of the factors you mentioned are important and should make a difference. I have insisted for some time that the balance sheet story should not be considered the main act when it comes to monetary tightening, it's a secondary importance at most. That, by the way, is both a reason for not worrying about over-tightening by tapering or an unwind, it's also though a reason for not relying on tapering or an unwind to tighten for control of inflation. That's why I favored a while back having the Fed start to gingerly increase its administered rates and target rate range last fall or winter. I'm glad it's planning to do so this year at last. I don't think that tapering or even modest unwinds are a substitute for raising rates. And let me just add to that.

Selgin: Our friend Manmohan Singh had a very good remark about that. He said, when you increase the short-term rates, it gets in all the cracks. And what that means is that when the Fed in comparison with tapering, certainly, but even with unwinding, those tend to tighten in some places more than others as it were, and not to have so uniform an effect on market rates at least at first. And so I think it's just much more reliable to treat the interest rate setting as the main instruments of monetary control and treat the ballot sheet as a separate, partly nonmonetary control issue.

Beckworth: Well, the FOMC itself said that much in the last minutes from December, they said their main lever is their overnight policy target rates. So they're going to use that lean on that more than balance sheet reduction. And this reminds me of attention that's always been there with the Fed's adoption of a floor operating system. The argument given for going down the path of a floor operating system or an ample reserve system is that you separate the size of the Fed's balance sheet from the stance of policy, right? You can increase, decrease the balance sheet, this extra degree of freedom or flexibility without affecting where you have your interest on reserves or your target rate.

I have insisted for some time that the balance sheet story should not be considered the main act when it comes to monetary tightening, it's a secondary importance at most. That, by the way, is both a reason for not worrying about over-tightening by tapering or an unwind, it's also though a reason for not relying on tapering or an unwind to tighten for control of inflation. That's why I favored a while back having the Fed start to gingerly increase its administered rates and target rate range last fall or winter. I'm glad it's planning to do so this year at last.

Beckworth: So, these two things can move in opposite direction, same direction, but they're delinked. Now if that's the case, and this is where the tension comes out, then large scale asset purchases shouldn't really have much effect, if large scale asset purchases do have an effect, they affect the, for example, say the growth of the economy that in turn should then affect interest rates and if it affects interest rates, but we're saying on the other hand that we can maintain the same stance of policy, you're creating some tension there between those two outcomes.

Selgin: I think that's right. But the tension is really in the theories, isn't it? When you get right down to it, right? They have to make up their mind. Does the balance sheet matter or doesn't it, and how? And of course we know that the theories of QE or LSAPs are controversial. And of course, and we have Bernanke's famous quip that QE works in practice, but not in theory.

Selgin: My own response to that is that I think the harder you look at the evidence, the more it looks like QE doesn't work in theory or in practice very well, especially when it comes to generating changes in the most important macroeconomic magnitudes, as opposed to just interest rates at the high end. But anyway, I do think there's a lot of tension there. And I do think there are some compelling reasons to want to get that balance sheet down, even if it's just to have an efficient floor system. And there's another reason that we haven't mentioned yet, but of course, right now, I just figure we're going to get to this is, what I call churning.

Selgin: Back in 1982, Milton Friedman had a, I think it was a Newsweek article where he complained about churning in the open market, by the Federal Reserve open market desk. And what he meant by that was that for every dollar of extra reserves, they'd engage in $200 of purchases and sales, most of which were offsetting. And he called this churning and he says, good for lining the pockets of the dealers, but not for much else. Anyway, what we've had going on since the crisis of 2020 is a different kind of churning where the Fed has of course engaged in large scale asset purchases, but it's put so much liquidity into the system that it's had trouble maintaining, the only way it's been able to maintain the lower limit of its target range, which is it as usual, isn't zero, it keeps it above.

Selgin: The Fed never really believed in the zero lower bound. It's never allowed it to become binding or it never wants it to become binding, partly because it doesn't think that money market funds can survive. So it's had to, in this case, buy oodles of liquidity back or borrow it back from the money market funds through the overnight repurchase facility. If you look at 2020, all the extra reserves that the liquidity of the Fed pumped in through LSAPs, almost all of it, it took back in the overnight reverse repo facility. And this is something that might not be necessary if the balance sheet started out at the beginning of a crisis, a lot smaller than ours did. So it's not as so much liquidity slushing around to begin with. At least I think that that's right.

Standing Repo Facility

Beckworth: Well, moving forward, I want to go back to the standing repo facility, and I think it's key that for it to really make a meaningful difference, two things need to be true. One, it remains stigma free, it doesn't get the stigma that you had on the discount window and banks get access to it as well. Right now it's just primary dealers. You want banks to be able to access it as-

Selgin: Oh, yes, absolutely.

Beckworth: In fact, it's key, right? If you don't want banks to be hoarding as many reserves as they currently are and have been in the past, you need to be in a situation where they know they have direct access to the standing repo facility and they can do it without stigma. So I think that's one of those devils in the detailed points that we need to make sure we get over that hurdle, if we're going to go down this journey. But assuming we do, do you think that the standing repo facility really changes the calculus of the actual long-term demand for reserves? I recently had Lorie Logan on the show and she's the head of the New York Fed's market desk, runs a big department there. And she made this point that the demand for reserves by banks is time varying. And I'm wondering how much it will vary because of the standing repo facility going forward.

Selgin: Well, if I understand the way that thing works, what the standing repo facility does is to make bank reserves and government securities, bills mainly, make them more perfect if not quite perfect substitutes. And so that makes it easier for banks to manage by holding treasuries and not worry about whether their mix of high-quality liquid assets is mostly treasuries or mostly reserves. They could be happy either way. And of course, that becomes very important if the quantity of reserves is being reduced as happened in 2019. But I do think [a standing repo facility] has much broader implications than merely smoothing the way towards a somewhat smaller balance sheet. I think it can also allow a return to a more scarce reserve system. And this was a point I made early on when this proposal was still being considered.

Selgin: I thought they were going to take it up at the Fed, by the way, a long time ago, I wrote something, I don't make predictions very often. And so I'm actually able to get them right most of the time, because I only make them when I'm cocksure. But this time at the big annual review, I said, oh, I'm pretty sure one of the things that's going to come out of this is the standing repo facility, and then they drag their feet, they drag their feet. But way back when I thought it was going to happen, I said, this isn't just a device that can help reduce the balance sheet in a floor system. It should and can allow the Fed to go back to a real corridor, not back to one, but back to have an orthodox corridor, which we never actually had. And that's because when you have a standing repo facility, you have a substitute for the old federal funds market.

I do think [a standing repo facility] has much broader implications than merely smoothing the way towards a somewhat smaller balance sheet. I think it can also allow a return to a more scarce reserve system...It should and can allow the Fed to go back to a real corridor, not back to one, but back to have an orthodox corridor, which we never actually had. And that's because when you have a standing repo facility, you have a substitute for the old federal funds market.

Selgin: The fed funds market doesn't do a thing. Now it's just arbitrage, it's an arbitrage arrangement. It's a place where GSCs and banks share interest on reserve. That's all it is. So banks don't go there anymore to make up for reserve shortages. That market is very relationship dependent. So by September 2019, all those relationships had been modern. So now you have the Fed doing the job the federal funds market would've done in the past. And that has all sorts of implications for how few reserves banks would be able to get by with. And particularly if you tweak the supplementary reserve ratio, that would make reserves actually more attractive than securities, as far as that goes. But otherwise banks would go for whatever asset gives them the best return and that says it should be.

Beckworth: So this journey to a smaller central bank balance sheet or Fed balance sheet, you would like to see it end up at a corridor system, a true symmetry corridor system. So prior to 2008, there was an asymmetric one, but you'd like to go to a symmetric, kind of like the bank of Canada has. But along the way, could we end up at a tiered operating system or if we got to a tiered system, would that be a final destination?

Selgin: Merely shrinking the balance sheets is never going to give you a tier system. That has to be done through a deliberate decision to grandfather a certain number of reserves and pay an interest rate, like the present interest rate on reserves, on those grandfathered reserves only that is above the target, that's the top, and the target rate, then the marginal reserves would pay less. So in a true corridor system, you have interest on reserves, but the interest rate is the lower bound of the corridor. And the top would be something like in this case, the standing repo facility rate, the targeted rate, which probably we should be talking about something like the SOFR now, because it's silly to talk about the Fed funds for reasons I just mentioned, that that targeted rate would bounce up and down between those limits.

Selgin: So in order to get to a corridor like that, you'd have to not only have a standing repo facility, but you'd have to have a lower limit defined by the interest rate on reserves, which would be a lower rate than the Fed is paying now. And you'd have to make reserve scarce, of course. Otherwise, the rate will tend to fall below that lower limit. So a tiered system would be a compromise where you say, well, we're not going to make reserves scarce at that lower limit, we're going to reduce the balance sheet to a point where we think we can't reduce it any further, and they're going to pay the old interest rate on reserves for those outstanding balances, but anything beyond that is going to get the new lower interest rate on reserves, which is defining the lower limit for the whole system. And they could pay that same rate on ONRPs or they could do all kinds of things to complicate things, but interest on reserves would no longer be at the upper end of the operating corridor, operating target range.

Beckworth: Yeah. I will mention that New Zealand has a tiered system and you've written on this. We'll provide a link to it. So this might be a compromised place to go, maybe part of the journey back to a symmetric corridor. But George, in the minutes we have remaining explain why we would want to go to a corridor operating system, what are the arguments for it, why is it better than what we have now?

In order to get to a corridor like that, you'd have to not only have a standing repo facility, but you'd have to have a lower limit defined by the interest rate on reserves, which would be a lower rate than the Fed is paying now. And you'd have to make reserve scarce, of course. Otherwise, the rate will tend to fall below that lower limit.

Arguments for a Corridor Operating System

Selgin: Well, I think that the main arguments, I make several arguments in my first book on this topic of the floor system called “Floored!”. But on my more recent book, I spell out the concerns that I think are my greatest concerns about a floor system and why we shouldn't have it. And that is that just as you were saying in a floor system, you can have QE infinity at any time. As long as you are able to control the lower end of interest rates and have that lower limit be positive, using interest on reserves and repos and such, then there are no times when you can't have quantitative easing, reserves will accumulate and because of interest on reserves and so you won't lose control in theory of the low end.

Selgin: Now in a corridor system, in contrast, because the interest rate on reserves itself is the low end and it tends to go to zero in a crisis, then it's only when interest rates are at zero that QE is practical because a corridor system turns into a floor system at that point. Otherwise, reserves pay less than the target overnight rate. So, at all other times, large scale asset purchases aren't feasible because they'll mean you can't maintain your positive target rate. So, the bottom line here is that the danger of abuse of QE for non-macroeconomic purposes, meaning mainly for getting out of a zero lower bound crisis, there's no ability to abuse those purchases in a corridor system, whereas there's a great risk of abuse of quantitative easing in a floor system. We could have the federal funds rate target be, say from two to three percentage points, and have somebody say, why doesn't the Fed buy oodles of this, that, and the other thing. And it could do so with relative impunity, technically, because we're in a floor system where it's setting interest rates in a way that would work even with large scale asset purchased. In a way, this is reversing the usual, the usual statement is we can keep interest rates where we want them without worrying about the size of the balance sheet. What I'm saying is it's good to have to worry about the size of the balance sheet sometimes because you don't want the balance sheet to become abused for nonmonetary, non-macroeconomic purposes.

Beckworth: So your first concern, your most serious concern with the floor system or ample reserve system is that it creates a temptation to be viewed as a piggy bank by different political interest groups to go after it, it looks like it doesn't cost a thing. We know that it does. It's tapping into fiscal space, but it looks on the surface like, hey, here's a freebie, let's just make the Fed go buy green bonds, let's make the Fed build the wall bonds. And so you want to avoid that temptation. That's that's the first thing you outline. You've also given some other reasons. I just wanted to go over briefly with you. One is that the floor system effectively kills off the unsecured overnight interbank market. Why should we care that that market is killed off, what does it bring us that we are missing now?

Selgin: Yeah, that's a good point. There was a literature back when, when we still have at our old system that talked about how the interbank market encouraged bank monitoring, interbank monitoring, because it is a relationship lending market, it's unsecured. So you need to keep tabs, the bankers would keep tabs on other bankers who in order to be able to trade in that market with them, they had to know whether they were safe, they had to know what kind of assets they held. If they thought that some weren't safe, they wouldn't deal with them. So you had all this interbank monitoring of risk going on. And that was very important because it meant that contagion effects would be limited because there'd be more knowledge.

Selgin: We know that contagions of panic and credit freezes are related to information asymmetry, which is a high tech way of saying, people don't know who's safe and who isn't. But of course the more monitoring that goes on, the less risk, the less asymmetry there is, the more banks know about. At least they know some other banks that are safe, that they can continue to trade with. And therefore that market for interbank lending would stay open. What that has done, by switching to a floor system, is it's killed all of that. It's killed the interbank fed funds market. It's killed the monitoring that went along. And now for that reason, let's say a bank needs emergency liquidity, it's short of liquidity, and it doesn't have the collateral needed to get it in the collateralized market, it's not going to have any banks that can turn to.

The bottom line here is that the danger of abuse of QE for non-macroeconomic purposes, meaning mainly for getting out of a zero lower bound crisis, there's no ability to abuse those purchases in a corridor system, whereas there's a great risk of abuse of quantitative easing in a floor system.

Selgin: So this increases just as we've created a system where the banks are dependent on having plenty of that can't fall short and quickly turn to another bank and just say, hey, can you lend me some, but they're also dependent on more dependent on the Fed in an emergency for support. So the Fed becomes less of a last resort source of liquidity for banks that are struggling for any reason to being the only source of liquidity. So there's a lot of damage that's done to the interbank market and to the whole ecology of the banking system. And this is typical of what regulators do because they're blind to the sources, the self-ordering sources that are good, I should say, self-ordering processes within the banking system and the way banks try to work with each other, to some extent, to stay safe and liquid.

Selgin: They've destroyed all that and put in place this monolithic Fed that is the only source of anything all the time. And then they say, well, just give them plenty of that stuff, and maybe everything will be okay. So there is no first resort anymore. There's no first resort lending, there's no first resort borrowing, the money market funds and the banks are all lending. Most of this stuff, they lend to the Fed. It's the first resort borrower, it's the first resort lender. This is not the way things should be working.

Beckworth: Along those lines, another concern that often comes up, and this is in particularly made by Peter Stella, but others as well, is that the Fed is becoming increasingly a public debt manager that it holds a large amount of the government liability. So if you count the reserves and you count also the overnight reverse repurchase facility and liabilities there, you're about five and a half trillion, well over five and a half trillion dollars and our debt that's held by the public is around 21. So that's a sizeable share of the total outstanding stock of debt. Should we be worried about the Fed managing that quarter of the public debt?

Selgin: Well, I think we should. What we really should be worried about is that we don't have fiscal monetary coordination going on. Of course, I am sympathetic with the arguments to the effect debt is debt and if we convert treasury debt to Fed debt or vice versa, this doesn't really change that much, but it does change some things, including the structure of the debt. And in general, I think it's not healthy to have a monetary system that plays such a role in shaping the structure of the debt by essentially turning so much of it into short term debt. When there's any reason why we should care about that structure and then, in that case it shouldn't be determined incidentally to monetary policy should be something we're thinking about more seriously and thinking about what the long run fiscal needs are, what's our best way of managing the debt for fiscal and monetary ends. So I think it's a bad system.

Selgin: I'm sympathetic with the modern monetary theorist who want to think in terms of a consolidated balance sheet, and imagine that we had one organization managing all of these things, coordinating all these things in a rational way. And I don't necessarily mean to say that the MMT way would be that rational way, but there is a rational way, and we probably aren't doing that. But of course, the best way to solve that problem, it seems to me the expeditious way is to have a small balance sheet at the Fed, a lean mean balance sheet, as they used to call it. In that way, at least the Treasury has the ability to manage most of the debt and the debt structure. And in principle to think about all these things, the Fed should only do what it needs to do for to have this minimum balance sheet necessary to conduct its monetary policy operations and achieve its goals.

In general, I think it's not healthy to have a monetary system that plays such a role in shaping the structure of the debt by essentially turning so much of it into short term debt.

Beckworth: I also think this last concern really compliments your first one. Your first one was that Fed's balance sheet becomes a tempting target for different political causes, different political interest groups. So it becomes politicized. Your first point is that Fed can become highly politicized. I think this does as well in the following sense. If the Fed starts paying out larger and larger interest payments to banks, to money market funds who are parked in the overnight repurchase facility, it's going to create the appearance that they're getting a subsidy. Now in practice, they probably aren't. If the rate is the same as the Treasury rate, then it'd be just the same as if they were holding treasuries. But the appearance at least creates bad optics and I just think it sets the up for criticism. Again, I can imagine Senator Warren going after the Fed, you just paid out this much money to the big banks that crashed in 2008. So I worry about that, do you see that as a concern?

Selgin: Yeah, it is a concern. There are a lot of bad optics that can happen with this arrangement. Another one is you have a Fed that needs to have a high interest rate to control inflation, but is sitting on a lot of debt acquired with that's earning low yields and suddenly it has to stop remitting any money to the Treasury. That can happen. The Fed can stay in business. But that's another thing that could be pointed to and say, what's going on here and who's getting all that money? All the banks are. So it's related to your point. So it's not just the banks are getting all this money, but look all the tax money that should be going to the Treasury, all the revenue.

Selgin: So there is a lot of that that can go on. And I think it's bad, but I think the worst thing is that the Fed has a much harder time than it would have in a corridor saying “no.” The easiest way to think about this is in the old days with a corridor system, imagine somebody said, why don't you guys print a trillion dollars and pay for a wall across the border, whatever, right? And the Fed would say, well, if we did that interest rates would fall below our target and inflation would go up and we'd lose control. End of story. There's no question, of course, of anyone asking for such a thing. Today the Fed can't do that. It would just be QE. They can keep their interest rate target.

Selgin: They just have to make sure the interest rate on reserves and the overnight repo rates are what they need to be to keep inflation under control, to offset whatever effects their balance sheet expansion might have. And so somebody's going to control inflation, but what is going to be is crowding out of private money, right? Because there's so much more purchasing of security going on at the Fed. Somebody has to be spending less, of course. So that's what I'm worried about. I want the Fed to be able to say no to QE when it isn't necessary for monetary purposes. And I believe the easiest way it can do that is if it can compellingly claim that if it didn't say no, it would not be able to achieve its macroeconomic objectives. And that is exactly how things worked in a corridor system.

Beckworth: So in short, a corridor system makes it easier for the Fed to stay independent?

Selgin: It ultimately comes back to Fed independence, and it makes it easier for Congress to retain the power of the purse because some people could see the Fed as a back way to do backdoor spending, could get the Fed to do these things. So you get a bill to create some new program and you have a rider in it that says the Fed's going to pay for all this. And the Appropriations Committee gets left out of the process. The next thing you know, you've got this problem where appropriations, which already are not fully controlled the way they were once supposed to be, but there's a lot more scope for backdoor spending.

Beckworth: Okay. With that, our bonus segment has come to an end. George, thank you for taking extra time out. Listeners, thank you for coming back for the bonus segments, and we'll see you next time.

Check out Conversations with Tyler, and subscribe to Conversations with Tyler on your favorite podcast app.

Photo by Pool via Getty Images