Manmohan Singh on the Role and Structure of Stablecoins and the Impact of Collateral in the Financial System

Manmohan Singh is a senior economist at the International Monetary Fund and works on rehypothecation, shadow banking, the plumbing of the monetary system, and more. Manmohan joins Macro Musings to talk about stablecoins, central bank balance sheets, central bank digital currencies, and their broader implication for central banks. David and Manmohan specifically discuss the role and structure of stablecoins, the impact of collateral within the financial system, how the Fed have looked to address plumbing issues within this 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].

David Beckworth: Manmohan, welcome to the show.

Manmohan Singh: Thank you, David. Thanks for the opportunity, and thank you Mercatus for having me. I just want to reiterate upfront that these views are my views, mostly coming from my book and some of the recent articles, and should not reflect anything on the IMF or its executive board.

Beckworth: Well, it's great to have you on. As you know, I've had your co-author and friend, Peter Stella, on the show several times. So you are long overdue, I should have had you on here a long time before now. And you are doing some interesting work on stablecoins, and as it turns out, that's a very timely topic. As our listeners will know, a lot of interesting developments, not such nice developments for stablecoins. And you've also done work, I think you're well-known for your work on collateral, and the velocity, and what does it mean when central banks are buying up and selling assets.

Beckworth: You're doing interesting work at the IMF, you give assistance to central banks in your case to central Asia, and you provide assistance on how the central banks should operate, their balance sheets. So you see this issue from other countries' perspectives and I think you have a lot to offer when it comes to the advanced economies as well. The Fed's balance sheet, the ECBs balance sheet, the bank of Japan's balance sheet. But we want to hold off on that discussion until a little bit later in the show. I want first talk about stablecoins because that's a really hot topic right now. As you know, there's been a lot of turbulence with stablecoins. I want to read an excerpt from a CNBC article that came out May 17th from a reporter named Ryan Browne.

Beckworth: I'm going to read this to kind of motivate our conversation, to kick it off because it speaks to some of the issues you cover in several of your recent pieces on stablecoins. So here we go, "Investors have withdrawn more than $7 billion from Tether since it briefly dropped below its dollar peg raising fresh concerns about the reserves underpinning the world's largest stablecoin. Tether's circulating supply has slipped from about $83 billion a week ago to less than $76 billion on Tuesday. The so-called stablecoin is meant to always be worth $1, but on Thursday its price slipped as low as $0.95 cents amid panic over the collapse of a rival token called Terra USD. Most stablecoins are backed by Fiat reserves."

Beckworth: "The idea being that they have enough collateral in case users decide to withdraw their funds. But a new breed of, ‘algorithmic stablecoins,’ like the Terra or UST, attempt to base their dollar peg on code. That's been put to the test lately, as investors have soured on cryptocurrency. Previously, Tether claimed all its tokens were backed one to one by dollars stored in the bank. However, a settlement with the New York attorney general revealed it relied on a range of other assets, including commercial paper, short term treasuries and unsecured debt issued by companies to support its token. The situation is once again, placed the subject of the reserves behind Tether under the spotlight. When tether last disclose its reserve breakdown, cash made up about $4.2 billion of its assets. The vast majority, $34.5 billion, consists of unidentified Treasury bills with a maturity less than three months and $24 billion of its holdings in commercial paper."

Beckworth: Okay, I'm going to stop right there. But the big story has been the algorithmic stablecoins collapsing. The article mentioned Terra, which our listeners will be aware of. But even Tether is having some challenges. And I checked the price before we came on the show, it's close to $1 again. So it fell briefly to $0.95, but it’s back to a dollar. So it's okay. More traditional stablecoin since it’s backed by reserves. But you've written a number of articles on these stablecoins and how we should think about them. Let me list two of the articles, and we'll try to provide links to these on the show notes. So you had an article on Risk.net titled, *How to Stop Stablecoins from Hoarding Precious Collateral.* You also had another article with Charlie Kahn that was titled, *Stablecoins Should be Backed by Reserves and be Interoperable.* So I want to go to the key points of these papers. They make similar arguments. The first one I would like you to comment on is that stablecoins, as currently set up, increase demand for high quality liquid assets. What does that mean? And why should we care about that?

Stablecoins and High Quality Liquid Assets

Singh: Okay, let me take a step back for the audience. There's about 180 billion of the so called stablecoins out there. Let's get this figure straight. So less than 200 billion relative to a three trillion crypto market. So we are talking about this smaller set of 180-200 billion. In this stablecoin market, which is not in the reg perimeter, the Tethers, the Terras, and the USDC with Circle is not in the reg perimeter. What professor Kahn and I have written, and then also another co-author is Caitlin Long in some of these articles. What we try to say is that in a digital world, if this is digital money, it's very unlike conventional assets of the banking system. The digital world settles in time T0, it's instantaneous practically speaking. In the digital world, it is incompatible with the intraday float, which happens with the banking system.

Singh: The banks actually are very comfortable because the intraday float of billions or hundreds of billion is quite attractive. Now in our articles, the point we're trying to make is if indeed you're looking at T0, then you need something instantaneous. Now some of the FinTech firms have come up with FinTech solutions faster than the Fed Wire. Fed Wire will become Fed Now. And we'll see the transition in one, two or three years. But the bottom line is if we need to keep our edge on technology, then there is demand for many not to leave the free float on the table, but to get instantaneous settlement. Having said that, these institutional clients would also like something stable. Now there is demand for Tether, but as you spelled out, Tether is siloing high quality liquid assets. Whether it's got $34 billion of Treasury or commercial paper, these are high quality liquid assets in the system.

Singh: Now at present, the US Treasury market is issuing, there's no real shortage of collateral. But if you just look across the Atlantic, Eurozone is still having [inaudible]. The German bonds are in short supply. In fact, if Eurozone raises rates, it's unclear whether the German bonds will also follow. Say, they go from -50 to zero. So collateral shortages is acute. Number two, the dealer intermediation capacity is also being constrained with all the bonds that will be issued. So even if you can make the collateral work, and we'll come back to collateral reuse, you need dealer balance sheets to make it work. So the reuse is higher so that the effective supplier of collateral from X becomes 2X you need dealer balance sheet space. Now having said that, even if there's no shortage of US treasuries, we have siloed a lot from QE. Over the past 10 years in various form, QE1, 2, 3 with COVID, etc. there's lots of treasuries outside, but the Fed balance sheet has not unwound. I'm talking about the Fed because ultimately more stablecoins are in dollars. So let's talk about dollars.

Singh: The unwind timetable is not very clear. Now in time, they will expedite the timetable. But still the balance sheet will remain from roughly… will probably go down from nine to six by the end of 2024. I was at a recent Atlanta Fed conference, I'm giving you a snapshot of the discussion. Now, should we silo more collateral given that the balance sheet is not unwinding? Should we see this market go from $200, $400, $500 billion? If a Tether is an icon for this market, I'm not going to talk about, say Terra, we will end up siloing. Because in this world, ring fencing US Treasury is key.

Singh: You don't want Tether to be reusing your US Treasury if indeed you are given a one to one guarantee. So in my view, if I have learned something from shadow banking, and I used to write about shadow banking quite a bit, being inside the reg perimeter gives the regulators a more clear optical lens to see what's going on. In that context, my article suggests that A, in the digital money space T0 versus instantaneous, there cannot be anything faster than central bank reserves. Because even if you look at US Treasury, there were bumps in September, 2019. There were bumps in March, 2020. Although we have many facilities right now, it takes time. So if you want to be inside the reg perimeter and you want to have something very, very liquid, central bank reserves would be the place.

Singh: I'm not talking about CBDC. You could still have your counterparty as a FinTech issuer. Now the FinTech issuer will lead a master account or access to payment rails. Now you may say, well that's like CBDC or you are getting the central bank on the hook. Because if something goes wrong with the counterparty, if there is a cyber attack or something, you still have a lender of last resort or LOLR being the central bank. I agree. That's why the president's working group had come up with a report that we need a FDIC type levy on these FinTechs. There may be a slightly different approach coming out from bank of England and say Bank of Lithuania. But let me focus on the Fed side. There's no applications which have been granted access to master accounts or the payment rails. You see the Fed likes to [inaudible] depository institutions.

Singh: The Fed did deliberate a lot before they came up with the RRP, which has about 160, 170 non-bank counterparties, the money market fund. But the Fed generally does not deal beyond the depository institutions. So if this technology… again technology needs to be brought forward. And if you want generally the word stable in stablecoins, I think there's nothing better than central bank reserves. The next best proxy would be having short term Treasury bills, one is for one. But for that, you will end up siloing collateral. And if this market grows, you don't want another half a trillion or even a trillion siloed, especially if the central bank unwind timetable is not forthcoming.

Technology needs to be brought forward. And if you want generally the word stable in stablecoins, I think there's nothing better than central bank reserves. The next best proxy would be having short term Treasury bills...But for that, you will end up siloing collateral. And if this market grows, you don't want another half a trillion or even a trillion siloed, especially if the central bank unwind timetable is not forthcoming.

Beckworth: Let me go to this point about locking up and taking out a circulation, this collateral that's important to the plumbing of the financial system. So your concern is if stablecoins rely on Treasury bills, for example, those are Treasury bills that won't be available to use as collateral in other settings. So this raises a couple questions for me. Number one, didn't we just increase the amount of treasuries by $5 trillion? I mean, why are we worried about collateral now? I mean, the amount of public debt has grown dramatically. So let me ask that question first, and I'll have another question to follow up.

Why the Concern About Collateral?

Singh: As I mentioned earlier in the context when we compare to the Eurozone, we do not have the Treasury market constraints today. But if we get this correct, if stablecoins generally come inside the reg perimeter and institutional clients generally believe that, "We are getting a better product. I don't want to leave free float with the banks." If indeed banks have hundreds of billions in free float, we trust that demand for stablecoins will go up. Now, the economics, the market will decide the economics, whether you want to take out your $100 billion from say, JP Morgan or Citi, give it to a FinTech issuer.

Singh: That basically means a swap between bank deposits and you deposit with the FinTech. So the liability side of the Fed has not changed, but basically the economics will tell you whether JP Morgan or Citi offers a better rate for the deposits, depending on the real economy, whether there's a demand for loan, etc. Or the client feels more at ease with the efficiency of having a free float instantaneously the money, rather than some hedge funds stuck with a float or margin between Friday ending and Monday. That market will decide, but you do open up the box if you bring stablecoins within the reg perimeter and to have them completely siloed. You may even have an option, and look, we'll give you reserves up to 50% and 50% treasuries. I may still be okay, but I would rather, in a digital world, why not go for the best?

Beckworth: Sure.

Singh: Why settle for the second best?

Beckworth: So the concern is, again, we're going to silo treasuries and restrict their ability to be used as collateral for other transactions taking place out there. So if I understand you correctly, what you're saying then is, even though we've had this increase in the amount of public debt, of the US national debt, the growth and demand for stablecoins could really take off. So we can't be comfortable with the fact that the national debt has grown 5 trillion. The growth of stablecoins, you believe will... Could be very strong and robust and just exceed that if it's allowed to use treasuries as the backstop.

Singh: If done correctly, and if it is a better model than [inaudible] when we know exactly that transparency, yes, there can be growth. Now let me make one more caveat over here. Right now, many of the so-called issuers are using a third party correspondent bank. So to my understanding Diem, the former... The name for Libra using Silvergate out of California. Kraken is using Silvergate out of California, because ultimately they want access to the master account. So they pay a cut to these correspondent banks, so that they can have access to the master account. They are still outside the reg perimeter. Silvergate would be inside the reg perimeter, but not theses [inaudible] who want a cut.

Singh: I would prefer rather to minimize shadow banking, if possible, to have it inside the reg perimeter. So you solve two issues. You don't silo collateral, which does have repercussions on the plumbing. And I'll tell you what, even if you have a lot of debt outside, if the reuse is constrained, partly because A, you ring-fence this. If you're doing it correctly, you're going to ring-fence this. You don't want these ring-fenced stablecoin collateral to be reused. And number two, the dealer intermediation capacity is getting constrained.

Singh: I was working recently on certain numbers on what is the overall pledged collateral market. And the collateral velocity is not going up. It has stuck roughly there in the last two, three years. Why? Even if you issue more debt, you need to use the balance sheet to reuse it. And I've looked at the raw number, which I've been looking at since 2007. Given the Basel III constraints, there's no more exemptions for reserves or treasuries. So if the financial engine, which are supposedly pulling 16 trucks, now you're telling it to pull 32 trucks or 64 trucks, that ain't going to happen. So that slowed down for good reasons. And I would rather admit that, if we are not going to beef up the reuse rate, and the economics favor ring-fencing, then is there something better than treasuries? Because the Treasury market has had bumps, I opt for reserves. Otherwise, that is the second best solution, use siloed collateral. And as you said, the US Treasury market doesn't show the signs as the Eurozone market has shown. So, I'll just leave it out there.

Beckworth: No, that's great. I agree with you. I'm just playing devil's advocate here with you. I'm very sympathetic to your argument that stablecoins should be backed by reserves or have access to a Fed master account. And we'll come to the specifics of how that might actually happen when we talk about the Fed's white paper they released on central bank digital currency and the potential for stablecoins playing a role there.

If we are not going to beef up the reuse rate, and the economics favor ring-fencing, then is there something better than treasuries? Because the Treasury market has had bumps, I opt for reserves. Otherwise, that is the second best solution, use siloed collateral.

Beckworth: So the issue again is that if you have a stablecoin that's backed by say, Treasury bills, you're going to silo that collateral, number one. And number two, there's also the interoperable concern. You want the payment system to operate seamlessly, and you can do that if you have bank reserves, more liquid. You also mentioned in your piece, treasuries, they take T+1 days to clear, where reserves can be instantaneous. So it's both a question of having the collateral available, but also making the payment system as quick and as efficient as possible. So let me ask a follow-up question. What would it mean for the Fed's balance sheet? Now you mentioned that liabilities would just change form, but if these stablecoins did get access to the Fed's balance sheet, they did grow, would that mean more rapid growth for the Fed's balance sheet, or would that growth have been there anyways?

Implications for the Fed’s Balance Sheet

Singh: So let me understand this correctly. The balance sheet as we see it today, is largely a function of the QE 1, 2s and 3s, and then the COVID era expansion, okay? The unwind timetable is, I think a welcome sign that they realized we need to shrink this. There's too much excess reserves. And if you're an old school, Peter Stella, Ulrich Bindseil type economist, a simple [inaudible] balance sheet without any crises has currency in circulation, has required reserves.

Singh: Everything after that needs to have explanation, quite frankly. And I think we have a lot of explanation here, where the paper with Peter Stella argues that you have a line item now, excess reserves, bigger than the currency circulation. These are abberations of recent times. So coming from that old school framework, I believe that shrinking the balance sheet is welcome. I've seen some recent discussions that it could also increase, maybe a 25 bps hike. But if you really dig down into the detail, then there was a paper at the recent conference by a Chicago Fed [economist], D'Amico, that 25 bps for 1 trillion of unwind.

Singh: And 1 trillion of unwind in the old unwind time table would've taken at least two years. In the June forthcoming time table, it will take one year. But you can imagine if you're talking about 25 bps from a model over two years or one year, that's a very marginal impact on the yield curve. Frankly, some of the big giants like PIMCO, BlackRock, et cetera, they move billions or transactions in tens of billions every day. So balance sheet unwind and the rate comparison, I think if generally there is a rate increase desire, you're better off doing 25 bps upfront rather than thinking that somewhere in the two year unwind timetable. So going from 9 trillion to 6 trillion, I don't think we'll have much.

Singh: Linking this to a stablecoin, if indeed there is demand, and I do not know how much of this roughly 180, 200 billion is genuine demand and what could be the nefarious demand. But even if you say hypothetically half of it is general, let it come inside the reg perimeter. And I think once it comes in, market will see whether generally if there's inertia for stablecoins. Because in this country and in most of the country, retail audience is quite happy with some of the Venmos and PayPals and Zooms, et cetera. It's the institutional clients who have deep pockets.

Singh: Now how deep are those pockets? Will this market grow fivefold, twofold, or will it not grow? But at least you will reduce the shadow banking footprint, which is happening. We just don't know much about the Terras and the Tethers. But one thing we learned from last week that much of the margins are related to the Bitcoins of the world, right? So very soon, the crypto gets entangled. And we don't know if this grows, whether with crypto, will some of the conventional balance sheet also get entangled? Do we really want to go there, or do you want to bifurcate some of the clean crypto who wants to come inside the reg perimeter, versus some of the nefarious crypto which may want to stay out to the reg perimeter. But at least you start bringing some of this work inside the reg perimeter, which I think would be welcomed.

Beckworth: Okay. So I should mention to the listeners that Manmohan, like me, is a fan of the corridor operating system. Is that right? Is it fair to say you also would like to see that? I think that's what you just alluded to, but you would like to see a smaller balance sheet return to a corridor operating system. Fair?

Singh: It is difficult to do so now, but yes. [inaudible]

Beckworth: Yes, whether we can do it is a different question, right, right. But in theory, at least you share that view with myself, Bill Nelson we've had on the show recently. And you mentioned the Fed's balance sheet is likely to reduce from 9 trillion to 6 trillion. Just normal QT reducing it while still maintaining the ample reserve system, the floor operating system. But moving forward, let's say 6 trillion is where we end up back when the Fed is done with quantitative tightening. Moving forward, the stablecoins, let's say they do get access to master accounts. So FinTechs that offer payments at the retail level, let's say they get access to a Fed master account. We get to that point where it's approved, it's legal. The Fed's handing them out. In other words, the Fed's handing out master accounts to non-bank financial firms, which some of them would provide stablecoins.

We just don't know much about the Terras and the Tethers. But one thing we learned from last week that much of the margins are related to the Bitcoins of the world, right? So very soon, the crypto gets entangled. And we don't know if this grows, whether with crypto, will some of the conventional balance sheet also get entangled? Do we really want to go there, or do you want to bifurcate some of the clean crypto who wants to come inside the reg perimeter, versus some of the nefarious crypto which may want to stay out to the reg perimeter.

Beckworth: Again my question is, do you see that as increasing the size of the Fed's balance sheet more than it would otherwise be the case? In other words, are we pulling in shadow banking, are we pulling in money creation that would've existed kind of outside the system and we're bringing it onto the Fed's balance sheet, or is this simply replacing for example, banks? In other words, I start using my stablecoin more than I use my checking account, is it a substitution or do you think it's a net gain, and therefore an expansion of the Fed's balance sheet?

Singh: I would say initially it is a substitution. A client which is very comfortable with instantaneous payment and no free float, he will see, say a hundred billion less in commercial bank deposits into a hundred billion move both on the liability side of the central bank. I don't see much change. How much inertia this has? What does the real economy demand in terms of loans? Then there will be a market tussle between those deposits, whether the deposits are gravitating to the real sector or whether there is enough profitability and economics with a stablecoin.

Singh: My take is there will always be demand. There will always be a price at which JP Morgan and Citi will lend to the real economy, Wells Fargo, et cetera. There's a lot of deposits, a lot of bank reserves after QE. So, I do not think this becomes a rare commodity in the near future. The fact that there's excess reserves, it is easier to think about this topic than if we were at zero excess reserves. So I think there is enough cushion out there that we could see this happen with a substituted angle and no change in the central bank balance sheet.

Beckworth: Yeah, I guess part of this also depends on what you think about the safe asset story. Is the demand for safe assets still going to be strong once we get to the other side of the pandemic, once we get to the other side... Let's assume and hope we get to the other side of the high inflation. We're back to a more normal setting, will those same pressures that led to low rates before the pandemic, so demographics, an aging planet, emerging markets, new regulations from Basel III, will all those forces still be there and still be driving growing demand for safe assets? And if so, stablecoins might fill part of that demand, satiate part of that demand. So I guess that's kind of a different question, but that might be a part of the story; what happens with stablecoins?

Beckworth: You have a quote in your paper I want to read real briefly here that's interesting. You say, "[The] non-bank stablecoin market is evolving like check clearing before the Fed's creation." So before the Fed's creation, they had a private clearing network similar to stablecoins. And later the Fed entered. So same thing for stablecoins. A good way to think about it is like check clearing before you had the Fed. So that's a very nice story to tell, but let's talk about that transition. So the Federal reserve recently had a paper come out that talked about its vision, its thoughts on a central bank digital currency, which could be implemented, or on the retail level to a stablecoin. They mentioned, at least at the retail level, having intermediaries provide access to that central bank digital currency. But you also see a lot of concern and discussion, especially recently after the collapse of Terra and other cryptocurrencies.

Beckworth: I want to read actually another excerpt from that CNBC article I started off with. Let me start here. It says, "The destabilization of tokens which have the sole purpose of maintaining a stable price has rattled regulators on either side of the Atlantic. Last week, US Treasury Secretary, Janet Yellen, warned of the risk posed to financial stability as stablecoins are left to grow unfettered by regulation, and urged lawmakers to approve regulation of the sector by the end of 2022. In Europe, Bank of France governor, Francois Villeroy de Galhau said, 'The turmoil in crypto markets recently should be taken as a wake up call for global regulators. Cryptocurrencies could disrupt the financial system if left unregulated.' Meanwhile, European Central Bank executive board member Fabio Panetta said stablecoins like Tether are vulnerable to runs, referring to bank runs, where clients flee financial institution en masse. The European Union is now planning to bring stablecoins under strict regulatory oversight with new rules known as the Markets in Crypto Asset Regulation, or MiCA for short."

There will always be demand. There will always be a price at which JP Morgan and Citi will lend to the real economy, Wells Fargo, et cetera. There's a lot of deposits, a lot of bank reserves after QE. So, I do not think this becomes a rare commodity in the near future.

Beckworth: So there's a lot of growing interest in stablecoins, how to regulate them, how to bring them in to the regulatory space properly. The Federal Reserve had this paper out I mentioned. What is your response to all this discussion? And particularly how would you envision stablecoins being set up, and what role would they play?

The Role and Structure of Stablecoins

Singh: Let me start from the big picture, and then I will converge to some of the more T-account type issues. You think last summer there were two very good speeches done a few hours apart. Randy Quarles gave a speech to the Utah Bankers Association, June 28, last year, 2021. And a few hours earlier, Mark Carney had given also a speech on a similar team at the BIS Annual Conference. And surprisingly, they were different speeches, but they did converge on stablecoins. They did find stablecoins to be amenable. Randy Quarles was very reserved on CBDC Mr. Carney wasn't.

Singh: But at the end of the day, if you read the speech very carefully, stablecoins could be a solution, partly because you see the central bank, as we discussed, the central bank balance sheet, traditionally has always been currency in circulation and required reserve. If you look at other countries also, and we, at the IMF, we look at... We just don't look at the QE countries, we have 190 countries. And I have a recent paper, I can bring some light from there. Many countries also depend on seigniorage, but they look at CIC, currency in circulation as... If there's ever a CBDC, it's the digital form of CIC. CIC gives you seigniorage. You print money in line with inflation and growth of money demand, you get seigniorage, which is given to the Ministry of Finance, to the budget. Now, I see a lot of discussion on CBDC, we use interest on CBDC, et cetera. But when you start giving interest on CBDC, you're paying money. It's a negative seigniorage. But these are some very new elements which are all on the horizon.

Singh: Coming back to the bigger picture on how I see stablecoins, initially, I find the central banks, if they are conservative, as they have been always, they may want some other FinTech or a narrow bank or a charter or banks to take the responsibility of AML, CFTC, counterparty risk, et cetera, et cetera. You see a central bank has a mandate to get the inflation to 2% or whatever mandate they have. And we have inundated them with green bonds and some other issues. I believe that if this market genuinely grows, do we really want the counterparty element to be with a third party, let's say a FinTech, or should central bank develop in-house techniques? Because CBDC means central bank is a counterparty.

Singh: Now there is something else going on over here; central banks in general find the FinTech applications to be something very new. They haven't really focused on this. And my take has been that the FinTechs have a very onerous process to apply for master accounts, et cetera, et cetera. But if you look at the stablecoin application, and my take has been that the FinTechs' counterparty risk are dealing with the FinTechs from a privacy provision. Most high net worth individuals, if it's retail or institution, would rather face a bank or a non-bank. The privacy provision is very strong, and if your audience is a bit academic, there's a very nice paper by Kahn and Roberds in 2005. Kahn, Roberds, and McAndrews. Very strong paper which says that even if information is free, there's something money has, physical money, which you will not get anywhere else. So privacy provision itself will tell me, I would rather deal with a Standard Chartered or a FinTech rather than the central bank. And I think for the central bank's angle, [inaudible]. Remember, privacy provision in the west is very, very deep rooted. I'm not talking about China, I'm just saying in general.

Singh: Most of the Western economy's privacy provision will remain. There is a demand for last currency bills in Switzerland, in Germany. There's a reason for that. So I think a privacy provision along with the fact that I don't see the old school central bank mindset changing. So for me currency in circulation should remain clean. CBDC should not contaminate in the sense that you start getting negative seigniorage. Required reserves are there, for a few countries excess reserves, time will tell how much is excess and how much will be drawn down or how much of that may move through stablecoins. Because I think if the economies grow, it is very possible that stablecoin demand for reducing the free float. Now, remember there are clients who still want to deal with the big banks and leave the money on the table. You see the hedge funds have a relationship. They understand that you may get from the payment system T0 settlement.

Singh: But if you're taking prime brokerage and you're taking funding, you may not want to deal with this fast settlement, you may not be okay with T plus one funding or the weekend is over because you have a relationship. So nobody can clearly say, "What is the net demand?" Right now, all I can say is there's 180 billion of stablecoins. Only the economics of the market, the relationship, the funding needs, and what do the client ultimately want? I don't think retail will make a big [inaudible], but institutional demand may bring this number forward, especially if it's inside the reg perimeter. So I will leave it at that, because after that it gets speculation, but it is better to see smaller balance sheets. And I think to the extent somebody wants to swap his deposits at a bank for a stablecoin is their choice. Maybe hedge funds won't because your relationship is deeper, maybe some of the other [inaudible] and others who want instantaneous settlement. But there is a market for this and I don't think everybody is so engrossed with instantaneous settlement.

Singh: That's my view. Payment settlement for a reason. This is a cynical view, but payment settlement title was never front and center, it was a back office function. In the last five to seven years, many of us have seen markets at zero rates. And believe me, many of my colleagues also at zero rate, there's not much to do whether it goes up or down so you start looking at a different topic. Almost all my colleagues have moved into this space because there was not much to write when you're sitting at zero. If markets change and operational issues come, policy issue comes, we may move into different topic, but we have spent a good five, seven years. And at the end of the day, except for a few marginal countries, CBDC is still a hot topic, but I don't see any advanced countries having yet implemented it. There's a lot of thought process in this. And I still believe the stablecoin, if done correctly, may not be a bad proxy for the market to solve an issue and not make that an explicit line item different than CIC on a central bank balance sheet.

Beckworth: Yeah, that's been my impression at least for the Fed, that they're not excited about having a retail central bank digital currency out there. Whether it's a token or whether it's setting up a Fed account where you directly have access to the Fed's balance sheet like you would at your local bank. The Fed seems very reluctant to go down that path. So the stablecoin would be the alternative path to go. And again, depending on how tied that stablecoin is to the Fed's balance sheet, it would be a close substitute or a proxy. I think that's what you're saying too, a stablecoin could fill that role, that niche, if there's a demand for something like that. So it would be done through retail. So that does seem to be the path we're heading down.

Singh: If I may correct you. You said retail, and I emphasize wholesale.

Beckworth: Wholesale. Okay.

Singh: Wholesale may be around more stablecoins and wholesale have a better fit. And retail, there's already enough competition out there. Fed should not be interested in my view. It's a different game. If I look at countries way beyond these four or five large advanced economies, it's a very, very different story. There could be legitimate reasons why countries with remittances, this can be a game changer for them. But the world outside the last five, six, seven countries we talked about is very different. And I've shown that in certain countries that digital money, and this is the paper with Peter Stella where the digital money, the demand for base money, if you look at the basic equation, the MV is equal to PT. The demand for base money in some countries may be going down because transactional velocity is gaining ground. There's something else going on over there.

CBDC is still a hot topic, but I don't see any advanced countries having yet implemented it. There's a lot of thought process in this. And I still believe the stablecoin, if done correctly, may not be a bad proxy for the market to solve an issue and not make that an explicit line item different than CIC on a central bank balance sheet.

Singh: So I cannot see the world only from the lens of the four, five advanced economies. There's a genuine discussion, which is very different. Seigniorage, demand for currency, base money falling down, remittances, et cetera, which is bread and butter. A country which was recently passed in the budget that they will implement CBDC 2022 or 2023 is India. It's a large economy, it's not at all like China, they have a privacy provision, but a lot of banking system deposit sits. So if it is disintermediated, it may actually be more optimal. So there is a case. There are also issues about counterfeit money floating around from neighboring countries et cetera, but the bottom line is that in some countries there is so much fat of deposits in the system that disintermediation may actually be healthier. And on top of that, if CBDC picks up this, then you get more seigniorage, which is very welcome in many poor or middle income countries.

Beckworth: So just to summarize, you're saying if there's going to be demand for stablecoins, it will likely be more the wholesale arena, demand for wholesale transactions, the wholesale level. Whereas retail's already been largely met by Venmo, PayPal, mobile operating networks. But even those operators, I imagine at some point would like to get direct access to a Fed master account, right? It'd make their life easier? If I'm PayPal, I got to use a bank currently to have access to the bank reserves, it'd be a lot easier to cut out that middleman and have a direct Fed master account myself.

Singh: I see the low income countries, which are not banked, some of the cases we saw in Kenya, some of the examples you see in central Asia, I think a lot more genuine retail demand from that side to reduce remittance costs, advance the economy, inertia. Let me be very honest, many of us in the advanced economies, you use your visa card, MasterCard, you get points. If you tell me today, I'll be very honest, from buying bread and butter to airline tickets, it's all on my card. Every three months, I can get an economy flight to Toronto free. If you give me that in lieu of retail CBDC with no points, I'll stick to my status quo. I'm very honest about it.

Beckworth: Yeah, no, no. There's reasons why we may stick with what we already have. So again, your point is the retail demand for CBDC probably isn't there in advanced economies. It may be fairly strong overseas. I think back to Zimbabwe when they had hyperinflation in 2008. People automatically moved into dollars, US dollars out of Zimbabwe dollars. And you can imagine a digital form of that may be being popular over there whenever they have unstable times. Let’s transition away from that discussion to what CBDCs and stablecoins might mean for the very issues you just touched on in terms of central banks like seigniorage. Also, what does it mean for the operations of monetary policy? So if in fact we do see more stablecoin usage, does it affect the effectiveness of central bank operations? Is it harder or is it easier for the central bank to tighten or ease monetary policy? Any thoughts?

Singh: Yes. By the way, let me make that disclaimer, which I may not have made, that these are my views, not of the IMF. Just to be very clear.

Beckworth: Fair enough. Fair enough.

How Do Stablecoins Affect Central Bank Operations?

Singh: So we'll have to divide the world out here into advanced economies and the not so advanced economies. I see stablecoin, euphoria, the thinking. And because the larger banks, the big G-SIB banks, all the FinTech advances in the advanced economy. We have discussed that in the last half an hour, how advanced economies may be able to create something where the counterparty raises the privacy provision element are more in favor of stablecoins versus CBDC direct. Now going to the other part of the world, the low income countries, the middle income countries, seigniorage is important. I don't see much discussion of stablecoins. So in that sense, they don't have the JP Morgans, Citis, they don't have some of the fancy FinTech names. Okay? So either they're going with some very rudimentary names, I've seen discussions how money comes from, let's say Russia to Armenia, et cetera, and those are very basic. But if you go through the banking system, and this is a very simple point, if you go through the banking system, like in Kenya, the float is kept in the banks.

Singh: They're required to put the float in the bank. However, in certain countries, you don't go through the bank because it involves a cut. If you have one side of the bank in Russia, another side of the bank, you have some cut and it can be up to 80, 90 basis points. And literally mobile transfers to phones, $300, $400 a month is nothing, it's literally free. So I see from a statistical and economic viewpoint, when you raise the issue of operations, are you picking up all the funding in your M2, the broad money aggregates, if their mobile payments or digital payments reach the banking system, it'll be picked up in M2. Here, a rough example would be Kenya. However, in certain countries where it's outside, where it's not in the banking system, it may not reach the banking system, there are ways to work with these non-bank providers or the mobile phone providers to see the float, and then you get a better handle on M2.

Singh: So let me now answer your question. For money aggregating countries, if money is still important, although by some rough measure, 50 to 60 were inflation targeters, but roughly 50 were also money targeters. So from a very low income country angle, money targeters should get a better handle on some of the money metrics which are changing. If base money demand declines because transaction velocity goes up, that's also an impact on M0. The leakage from say mobile transfers will be a leakage from M2. So this is what some of the money aggregate folks should bear in mind. For those who are inflation targeters, it's a different world, but understanding liquidity outside the banking system or liquidity from the digital money where the velocity is much higher is still a very relevant topic. So if this catches steam, I think everybody should understand how this would impact your monetary operation, because some basic pieces of money will be redefined or re-tweaked.

Singh: At the end of the day, my understanding is, seigniorage remains important in these countries, much of the technical assisted work we do. Even with the last two, three years of COVID, if a typical central bank would release 70% of its profits to the ministry of finance, in the last two or three years, they were releasing a hundred percent because there was genuine need for this extra money. So never trivializing, which may not be a topic in advanced economies. But from the lens, if you are euphoric about digital money and base money demand goes down, or there are implications for seigniorage, you need to juggle those. It all looks very nice, but at the end of the day, let's say it has an adverse impact on your seigniorage, let's say you start paying interest on CBDC. These are questions you will have to face from your ministry of finance. So the world is very wide and diverse, and what shoe fits one folk may not fit the other folk at all.

Beckworth: Sure. So you're saying seigniorage is still an important form of public finance in many emerging developing economies. And so this could be a big deal for them, especially if money is outside, and the central bank's sphere of influence begins to effectively take seigniorage away from the central bank. It could have a bearing on public finance. Where in the advanced economies, it's less of an issue. Is that what you're saying?

Singh: I think advanced economy seigniorage is important, but-

Beckworth: It's small compared…

Singh: Small, relatively small, and the Treasury can work its way around. That's not really needed for your budget. It's not a very integral part of the budget.

Beckworth: All right, let's move to another area where you have written extensively, one that I have followed. In fact, it’s the area where I first came across your work, and that is the issue of the collateral. And we touched on it earlier, but collateral and how it's used to facilitate transactions in the financial system, and how this is all related to the advent of large scale asset purchases or QE. Now, Japan did the first QE in recent history in 2001, 2006, but then the Fed brings it back in full force, starting with the great recession, 2008.

Understanding liquidity outside the banking system or liquidity from the digital money where the velocity is much higher is still a very relevant topic. So if this catches steam, I think everybody should understand how this would impact your monitoring operation, because some basic pieces of money will be redefined or re-tweaked.

Beckworth: And ever since then, we have QE1, QE2, QE3, the Eurozone adopted its own version of this. Bank of Japan, Abenomics. So we've seen a lot of large scale asset purchases taking place, it's become the norm. And of course it's easy to understand, as the world has gone closer and closer to 0%, and many advanced companies were stuck there, central banks felt compelled to resort to balance sheet policies as an alternative tool. But it does have a consequence for the assets that it purchases. So walk us through the concerns you have raised about this. When the Fed, for example, buys a bunch of Treasury securities, why does that matter for the ability of the financial system to do transactions and facilitate economic activity?

The Impact of Collateral on the Financial System

Singh: Okay, let's go back to basics. And for those interested in this, I would argue that some of the very interesting and intriguing work, even by top academics like Eric Leeper, et cetera, even struggled with this concept called reuse of collateral. And I think that is my main argument. When I tripped over this topic, I was just looking at the annual reports of the top 20, 25 major banks who have a footprint in moving collateral globally. Not everybody. You could be a large bank, but if you don't have a global footprint, that's the Standard Chartered, it's not a big deal. JP Morgan is, HSBC is, but some of them are not so big clears.

Singh: So when I looked at this and I look at them every year around this time, April, May the numbers were not trivial. The debit and credits in the market do not just settle with money. So if you are settling, say 25 million with me, you will push cheapest to deliver. So settling accounts, you could post cash, you could post treasuries, you could post any HQLA, which is amenable and acceptable on a mark to market basis. So just to give you some statistics, when I first looked at these numbers around Lehman's time, the amount of collateral used or traded in these big markets or via the big bank was about 10 trillion. Trillion with a T.

Beckworth: Wow.

Singh: The M2 in 2007, '08, was in the US about seven to eight trillion, in the Eurozone about seven to eight trillion, in the UK about one to two trillion. So all of M2 does not make it to Wall Street. A lot of M2 is in safe deposit, it doesn't leave the banks. So relatively speaking, the amount of debit, credits to the pledge collateral market is at par or maybe even bigger than money. So when I look at lubrication, you know we learned about M0's, M1's, M2's?

Singh: I, in my own mind, created this C0, C1, C2, the collateral aspect. I'll tell you why that's important, because you raise the issue of LSAPs. There's no theory in LSAPs, but what you're essentially doing is when you do QE, you print money, that's a $100 bucks, and you take a $100 bucks worth of good collateral; typically US Treasury or Bunds or JGBs. Now there's an assumption that you have released money, but there's no assumption that what you pulled out from the market from it, is also has moneyness.

Beckworth: Yeah.

Singh: US Treasury especially, well, it is not only short term US Treasury, even a 20 year bond or a 17 year outstanding bond, has the reuse element. You see, and this is the question often comes up. Why are people buying bonds at zero percent? Or why are Treasury yields lower when you can get IOER higher than the yield curve. This was the residue which Professor Leeper was also struggling with when you interviewed him. And the point is that the moneyness from the Treasury may be equal, may be more, may be less than the money you put into the market domain. As you and I have seen a lot of bank reserves, a lot of money put out there, is excess reserves. Okay, we can control it. We have IOER, et cetera, et cetera, but the bottom line is, it's just not net money. You have to subtract the moneyness, and when you subtract that US Treasury moneyness, because US treasuries are reused, we do not know the net impact, if the lubrication has been positive for the plumbing or negative for the plumbing.

Singh: So this is where, when you look at simple ISLM, I'm not an academic guy, but in ISLM, the LM may not always shift in parallel. It may pivot. I have some charts in my book where if LM pivots, your QE may be net, not lubricating in the moneyness and not just the money sense. I don't want to be too abstract, but I think this is an issue where the numbers are not trivial. LM can pivot, and that's why too much QE at the margin… and I have supervised a couple of dissertations to young folks, Muley from MIT and Wang from Yale, and he's back in PBOC right now, where they have shown… and I think professor Woodford also had that too much of QE actually gets into the issue where LM starts pivoting. And when you start pivoting, the net impact of moneyness can be very, very different. I'll leave it at that because I'm getting academic again, and I don't want to get too academic.

Beckworth: Well, let me try to explain it the way that I understand this. And this might be a very simple interpretation of your work. So, apologizing in advance if I get this wrong, if I don't provide the nuance that it deserves, but what I think you're arguing is, we need to think of money more broadly than just say M1, M2 traditional money assets. Treasuries themselves, there's a large literature on this, treasuries themselves have a convenience field or a liquidity service they provide. So if you want to think about a proper measure of money, you've got to include all assets, provide liquidity or transaction services. As you mentioned, even 20 year Treasury has some liquidity in it, maybe less so than a three month Treasury bill, but it has something to it. So you've got to think about that. And if you accept that premise, number one, that it's a much broader measure of money than just standard textbook treatments of money, then it gets interesting when you think about QE or large scale asset purchases, because what you're doing is you're injecting reserves, which one form of money, but you're pulling another form of money out of circulation. In fact, what you're pulling out of circulation, treasuries, everyone has access or everyone can use. But reserves are only limited to the banking system. So-

Singh: Exactly, exactly.

Beckworth: Yes, you're stuffing the bank balance sheets with more reserves, but these liquid assets that everyone else can use, you're actually pulling them. So your LM point, or I think the point you're making is, it could be a wash in terms of if you properly measure money, QE could be a wash on the total effect of the money supply. Is that a fair interpretation?

Singh: I think it's a very fair interpretation. I'm glad you mentioned that treasuries are used in all the pipes; non-banks, banks, everybody uses Treasury for whatever regulated reasons, et cetera, collaterally, the prime brokerage and sec lending, deliverability margins, repos. They are all plumbing. In fact, all these rates, repos, sec lending, prime brokerage, all these plumbing rates. It’s nothing else but a ratio of money and collateral in the market. And once you start making pipes different, for example, you talked about the banks only, with IOER, that's just one pipe, but if you hypothetically, and we have had a lot of direct pipe, the RRP is a direct pipe with the Fed. Banks have IOER as a direct pipe, CCP's also have a direct pipe. Think of this just in an abstract sense.

Singh: If many, many pipes are dealing with the Fed balance sheet directly, you have less money in the market relative to collateral. And if collateral money is a ratio, repo rate or a set lending rate, you start changing that ratio. This is what happens if hypothetically all the pipes go to the Fed, that's it. How will you price collateral? Where will you get a repo rate? Where will you get a sec lending rate? I think what many have called now the footprint, and back to the central bank balance sheet, the lesson of the footprint, the cleaner is the collateral money ratio, or the cleaner is the repo rate, the cleaner sec lending rate, et cetera, et cetera.

Beckworth: Well, this goes back to the earlier discussion about central bank digital currency. That's really more fundamentally a question about who has access to the Fed's balance sheet. So if everyone had pipes into the Fed's balance sheet, then we would be dealing with these issues. And for example, we have seen growth in the number of parties that have access to the Fed's balance sheet. So, traditionally it was depository institutions, primary dealers, but then in 2008, we opened up the overnight reverse repo facility to money market funds and to other financial intermediaries. So the balance sheet of the Fed has grown and there's more access to it. I mean, maybe let me frame it this way. The point I made earlier that only some entities have access and can use reserves, you could fix that by making the Fed's balance sheet open to everybody. But then you create a whole host of other problems that you've touched on, including the extent of the Fed's footprint.

Beckworth: Let me step back from all this though and take this insight of yours and apply it to the two crises, the great recession or great financial crisis as some call it, and then the pandemic. So if we go back and look at 2008, 2009, and look at attempts to measure this broader view of money, so I'm going to pull up here The Center for Financial Stability. They measure these Divisia measures and they have like an M4, which includes treasuries, commercial papers, repo. And if you look at this broad measure of money, it actually collapses during the great financial crisis. Much of the safe assets disappear. What we thought were safe assets, they disappear, and you actually see an outright contraction in the level of the money supply when you properly measure it according to this argument you're making.

Beckworth: Now, if you come to the present or the past few years, the pandemic, and you look at these measures, they actually have exploded. So let me go back to 2008 in the great financial crisis, the broad measure of money falls. But if you look at M2, it's relatively stable, because it's backstop, it's protected. But you actually see the amount of money collapsing during the crisis, whereas in the more recent crisis, the pandemic, you see it actually growing rapidly. How do you explain those two differences? Why did the effective money supply contract 2008, 2009, and then it grew rapidly? What are your thoughts, Manmohan?

Explaining Recessionary Differences in the Money Supply

Singh: Okay. It doesn't have a right answer, but I will look at it from the broader lens. When I looked at the 2008 crisis, there was also a collapse in the pledge collateral market from $10 trillion to roughly five and a half trillion. That was lubrication which disappeared, partly because A, the price collapsed of all these safe assets. And number two, the quantity also collapsed counterparties, as I said, the reuse was done from one bank to the other. When the Citi bank price is below one buck, or you have Santander in trouble, you don't reuse, right? You have less and less counterparties, so less and less collateral flows globally. So there were two elements: both price and quantity, which collapsed that. So part of the explanation about why your money or Divisia… now remember the Fed doesn't publish 2005 anything on M3.

Singh: So what you mentioned is, if I start putting in my numbers, plus collateral number, the M0 plus C0, or M2 plus C2, I also see a collapse. And that's why there was severe tightening. There was lack of lubrication from your side money, my side collateral. However, in the more recent phase, you already had the build up of bank reserves. The liquidity was already there. COVID gave some more liquidity. And frankly speaking, most of the banks had digested the Basel III elements. You also had a year where [inaudible] were exempt. Remember from 2020 to '21, the reserves were exempt for a year from some of the calculations. I think reserves and treasuries were both exempt. So the point I'm making is, the dealer [inaudible] space right now is tight. I don't think they're going to be able to push up the reuse rate, but I found the reuse rate along with money bouncing back.

Singh: So the reuse rate in the 2018, '19, '20, '21, also went up from about 1.8 to 2.5. The dealers understood that there are ways to get around Basel III. There was some exemption for a year. So that would probably be in sync with my thinking that M0 and the money and collateral collapsed around the Lehman time. But the euphoria printing did not stop. COVID brought in some more money. And the dealer balance sheet was still robust enough to push the reuse rate. And that's why you saw the overall lubrication on the up trend in the last two, three, four years. I don't know if that would carry on because as I repeated earlier, the dealer balance sheet constraint will not get any better. That's why you have an issue now. Can we clear treasuries? Can we alleviate the dealer balance sheet, et cetera, et cetera. These are new topics because they are admitting the engine cannot pull more trucks in.

Beckworth: Okay, let's end on that point by going to an issue that's now being considered by the Fed. And that is whether they should permanently change or tweak the supplemental leverage ratio, which would remove reserves from that ratio and provide some increased space on the balance sheets of these big banks, these big institutions. There's also this new standing repo facility, which in theory should make it easier to swap treasuries for reserves. Do any of those developments, the supplemental leverage ratio is a prospective development, but let's say that Fed does go ahead and drop reserves from that, and that the standing repo facility is used. Does that change anything in your view? Does it make it easier for financial firms and the plumbing of the system to operate efficiently?

Would the Fed’s Policy Developments Increase Financial System Efficiency?

Singh: Okay, let me break it down into the two, three proposals which I see coming. The US Treasury to be centrally cleared or more centrally cleared, to be made more safer, if you really get into the nitty gritties of CME and FICC, there are two very different machineries. CME is into [inaudible] future, the other one is cash repo. One's under Chicago Fed, one's under New York Fed. I don't think you'll get much more out of the clearing being done. Also one has a segregated model. One enjoys the float, the FICC. Your next issue was whether to exempt either reserves or Treasury, or both to get some space. Yes, I think you will get some space. If you need that Treasury pipeline, if you had a large package of COVID, well, Mr. Manchin stopped that, but I mean at the end of the day, if you had larger and larger Treasury release, do not expect the banks to come in and make markets.

Singh: This was 2019. It's not that they couldn't do it, but if a large bank, let's say JP Morgan and Jamie Dimon, if you are better served by prime brokerage returns, and the repo consumes more balance sheet, as a fiduciary duty to my shareholder, I'll go for prime brokerage. The fact that he could make some money off the table for repo is less relevant on the repo markets, touching double digits, when I can use a balance sheet more profitability for prime brokerage. So if you give balance sheet constraints, don't expect them to do market work for you because you are not allowing them to do the market work. So yes, I think if the pipeline increases and there is need for you to expect banks to do all sorts of repos so that the money collateral ratio remains in sync, and you don't get spikes, then all these initiatives are very positive.

So yes, I think if the pipeline increases and there is need for you to expect banks to do all sorts of repos so that the money collateral ratio remains in sync, and you don't get spikes, then all these initiatives are very positive.

Singh: The standing repo facility, then they have FIMA [Inaudible]. All the after runs were sort of not desired. So all these elements help. Half a trillion here, half a trillion there. Now if things go haywire, you may exempt it overnight or for a year if there's a crisis, but you don't want the Treasury market to have bumps. Fed coming once in 20 years is one thing. But you come in in 2019, come in in 2020, tells you something is going on out there. The banks, or the financial engine is constrained, but everything else is just evolving, that is being issued, et cetera, et cetera. And don't expect them to pull more and more trucks if the engine can only, say, pull 16 trucks. It'll stop. It'll not pull more trucks if you add it. Something has to give. And when something gives, and the Fed comes in, it just looks like, why? Why is there a repeat? So that doesn't send a good signal to anyone.

Beckworth: Yep. Well, it'll be interesting to see how these proposed reforms play out as we move forward. Well, our time is up. Our guest today has been Manmohan Singh. Manmohan, thank you so much for coming on the show.

Singh: Thank you, David. Again as I said, I enjoyed picking up these themes from my book, and some of the recent papers. I must again reiterate my views. The fund is a very different place, but we do get a chance to speak. I just wanted to reiterate these are my views, and maybe a bit quasi academic, but this is a theme which I have enjoyed and relished. I see a lot of inertia. I've seen papers now from the Fed, from Reserve Bank of Australia on the reuse concept. I've seen BIS have many working groups trying to think this better. So if nothing else, I think I did push the envelope. I'm glad if others can build on this. And thanks again for this opportunity. Thank you very much.

Beckworth: Thank you.

Take the Macro Musings listener survey here.

Photo by Justin Tallis via Getty Images