 Good morning, everyone, or good afternoon. So we are happy today to have our CB and DC webinar series. And today, we'll be hosted by David Mew of the Federal Reserve Board of the Governor. And Rod will talk about the interest bearing CB DC and the heterogeneous band. And Monica will discuss after this. So now I will go to David. Thank you. Thanks, everyone, and welcome. So just a few housekeeping items before we get started. Rod will have 25 minutes to do the presentation. Monica will then be discussing for about 10 minutes. And we'll have about 20 minutes for Q&A discussion. A couple of rules on the Q&A. The panelists are able to unmute themselves and ask questions, make comments. Other attendees, please use the Q&A tab to ask questions. I will then use the Q&A box to sort of channel some of those questions that we have from the attendees to the presenter and to the discussant and go from there. Please note that this conference is being recorded and the video will be posted on our website. So with that, I'd like to turn it back over to Rod to kick us off on interest bearing CB digital currency with Peter Junis Bank. Rod, over to you. Great, thank you. That's great. Okay, how does that view look for everyone? Perfect. Great, thank you. So it's my pleasure to present this work here. This is honestly the right place to be for this paper on central bank digital currencies. And let me just point out that this work is joint with Haji and Zhu. And so let me just jump right in, of course, and start off by just talking about the objectives of central bank digital currencies. And one obvious way to do that is to look at the recent survey by the BIS and they asked member banks what type of work they were doing on CBDCs and how important they thought various things were like payment safety, payment efficiency, financial inclusion, monetary policy and so on. And you can see there's a lot of interest in improving payments, but all of the objectives matter. And there's a very active discussion now in what the possibilities are for central bank digital currencies. And it's, I think the interesting if we look back, way back in history now to 2016, this quote, which seems like such a long time ago, this quote by Ben Broadbent where he states that if all the CBDC did was substitute for cash, if it bore no interest and came without any of the extra services we get from bank accounts, people would probably still want to keep their money mostly in commercial bank accounts. And if you remember, this was, he said this, when we were really thinking about CBDCs as this thing that might disintermediate banks and could be a financial stability risk due to flight to quality and so on. And so in a way, people were talking about ways to minimize the impact of CBDC. And that's kind of surprising in retrospect because how can something be effective without being in some ways disruptive? And so now I think people are actively looking more toward ways to make CBDC better than cash. And what I'm gonna do in this talk is sort of explain the work that we've been doing along these lines and in particular related to two functions of money which is the store of value and medium of exchange aspects. So one thing I also wanted to add is that this desire to sort of create the central bank digital currency that meets these various needs has also been reflected in a set of legislation that's been coming through the US Senate and there's actually a couple of other bills that have also been put forth. But what I find interesting about this banking for all act which was put forth by Senator Sherrod Brown is this idea that it talks about these different features. So it talks about the interest bearing, it talks about the services that this CBDC should provide. Now one might argue that this act is overly prescriptive. I mean, I'm sure David would have something to say about that they seem to have it all figured out. But like I said, I just wanna emphasize the idea that they're thinking on these multiple dimensions of what a CBDC should do. Okay, so what do we do in this paper? We provide a tractable model of deposit and lending with large and small banks. The large bank has a convenience value and hence it ends up having four market power in the deposit market. The models tailored to the US financial system and in particular we think it's well suited to looking at a system with large reserves, a system where monetary policy is largely driven by the choice of the interest on reserves rate. I'll say interest on reserves throughout the talk, but of course I also need interest on excess reserve. And banks are not constrained by reserve requirements. We're gonna, I'm gonna spend a fair bit of time out of my short amount of time evaluating the model and just showing that the model works because it's pretty easy to draw the conclusions about CBDCs from our model once you see the model. But I wanna explain the model to you and I wanna convince you that it's a good model by showing you that the model can do a reasonable job of explaining deposit markets at least before we even talk about introducing a CBDC. And then I wanna of course introduce an interest bearing CBDC. First I'm gonna do this by assuming it's not only issued through the commercial banks, we'll assume that throughout but that it adopts the convenience of the host bank. So it becomes another deposit account at each bank and it's gonna adopt the convenience value of that bank which differ across the small and the large banks. And then we're gonna contrast that to what happens when the CBD has its own convenience value. And what we're really thinking about in this context is a CBDC that's still issued through the commercial banks for a variety of reasons but really has its own convenience. You could think about this as being a CBDC that's maybe programmable and gives you access to these alternative likely permission platforms. But this is the same for everyone this access regardless of which bank they happen to access it through. And then I'm gonna talk a little bit about trading off interest and convenience because these are really these two alternative design features, right? So going back to what I said at the beginning we can think of interest as being something that enhances the store value feature whereas convenience is something that's related to the performance of a CBDC as a medium of exchange. Excuse me. So a preview of the results. So when we think about just an interest bearing CBDC that's interest rate is gonna put a lower bound on deposit interest rates. We're gonna show that that ends up improving monetary policy transmission. And what we mean by that is the transmission of IOER rates into deposit markets into deposit market rates but that it further reduces the market share of small banks. So it increases the inequality in the banking system. And then we're gonna talk about this idea of a convenience CBDC sort of improving on that other dimension that be our exchange dimension. And we're gonna show that this level is the playing field by increasing the market share of the small banks. So it takes away some of this advantage of the large banks, which may be a good thing. And in fact, if the convenience value is high enough we're gonna show that this also improves monetary policy transmission. So to get started, let me just say that our work builds on a previous literature that has modeled deposit and lending markets in this current regime of large excess reserves. So the papers that I would say we draw primarily from in terms of the baseline model as Martin, McAndrews and Ski and David Endolfato's paper in these frameworks, reserves are abundant. Lendings determined by the opportunity cost of funds and banks have monopoly power in the lending market. We have all these features and a loan is made if the return exceeds the marginal opportunity cost of reserves. So the difference in our model is that we're gonna model how deposits are generated when loans are made and most importantly that these deposits have to be retained within the banking system. So I've made this point before that this really is something that Todd Keester emphasized. He called it the fallacy of composition when he made his testimony to the Monetary Policy Subcommittee. This idea that banks don't get to choose how many reserves are in the system. So if banks made a loan, those deposits have to go somewhere and they may even stay at the original bank. What this ends up doing, since that's gonna depend on the size of the bank, the market share, it means that the opportunity cost of lending is gonna be related to the deposit market share which is, and so bank, different banks are gonna have different opportunity costs of lending and that's why we're gonna get these differential effects. Okay. This is a hot topic. I won't go into the literature review. There's a large one which is very exciting. Most of these papers have been presented in this seminar. I imagine the ones that haven't will be soon. So maybe that's all I'll say about that for now. Again, I think the real difference in our paper is this idea of looking at the heterogeneous banks which means that we get to sort of focus on these. We still look at total effects but we also get to focus on these compositional effects. Okay. All right, an outline. So we're gonna look at, I'm gonna, first of all, talk a little bit about the model and the equilibrium. I'm gonna look at the impact of that. Well, I'm gonna defend that model a little bit as I mentioned. Then I'm gonna look at the impact of a CBDC interest rate, the impact of a convenient CBDC and then I'm just gonna have one slide. If I get to it, it just talks about how we're gonna start thinking about this trade off because right now the paper primarily looks at if you have a CBDC and it has interest rate S, what happens? If you have a convenient CBDC and the convenience is V, what happens? But ultimately we wanna think about design, right? How should we, what should SV, what should we be? And so I'm gonna just introduce a little bit of a framework about how we're thinking about that. Okay. So the model, large bank and a small bank, each, as I've mentioned, each consumer has a convenience value for the large bank and that varies across individuals. So those are draws from it, just independent draws from a distribution. Banks have reserves, those are exogenous and large and their liabilities are their existing deposits and then the central bank based interest on reserves held by the banks and that's denoted by F. So that's a little bit of notation that they'll come up a few times. So we model a CBDC as a central bank liability, but we imagine that these are gonna be offered through the commercial banks. So in the first version of the model, which I'm presenting, the CBDC is offered through the central bank or through a commercial bank and it adopts the same convenience value as the host bank. So zero, if it's offered through the small bank and Delta, if it's offered through the large bank, the interest rate on CBDC is set exogiously. As I said, we're not quite at the idea of picking that yet. And because the CBDC in this version of the model is a substitute, a perfect substitute for the deposit, as the same convenience, this interest rate S ends up being a lower bound. The commercial bank has to offer an interest rate at least as good as the CBDC rate or else nobody would hold the commercial bank deposit. Okay, agents in the model, they play multiple roles. So there's an entrepreneur role. So each agent is endowed with a project. The projects have qualities, success probabilities QI, which are drawn from the distribution. The project requires a dollar and agents don't have a dollar. So they have to go get it from a bank. If the project is successful, it pays A, greater than one, if it fails, it pays zero. And with the monopoly power idea comes in. So agents have to borrow from the bank where they have their deposits. So this idea of relationship lending. If they get a project funded, they're gonna go ahead and they're gonna hire a worker. So they just pick an agent who's suddenly gonna play the role as worker and the worker gets paid a dollar to complete the project. And then that worker is now a depositor who has this dollar and she can deposit it in the bank of her choice. And that choice is gonna depend on what the interest rates are at the two banks. So in particular the differential and what that particular worker's convenience parameter is. So in terms of timing, banks at the rates, central bank sets IOER rate, CBDC if it exists as a rate S, agents start with at some bank. The agents are endowed with a project, go through this process I just mentioned. They see if they can get a loan, if they get a loan, they hire a worker or if they hire a worker, that worker has to find a place to deposit the money. Okay, so as I mentioned this idea, this fallacy of composition idea. And so this is being reflected just in the following little sort of balance sheet exercise. I just, it's important to illustrate this because it's key to the analysis of why the opportunity cost of lending differs across the banks. And it's important for the marginal profit conditions that determines bank lending. And so just think of the simple scenario where we start off with, I guess, a narrow bank. So reserves equal deposits. And think about that bank, if it wants to make a $1 loan, what it does is it creates a new deposit, okay? But the person who's borrowing that dollar isn't just gonna leave it deposited at that bank, it's gonna go hire the worker. It's gonna pay, buy supplies or build whatever it needs to build in order to carry out the project. And then whoever it buys that from has to deposit that dollar somewhere. And that's the key, that dollar doesn't disappear from the system. And so we assume that that dollar goes back to the bank with the same probability that the fraction of the deposits that that bank holds, which is endogiously determined. And so that person chooses where to deposit and basically in the equilibrium, alpha S of those deposits will go to the small bank and alpha L will be retained by the large bank. So we get this new balance sheet position at the bottom. And so what that means is that when we look at the marginal profit of lending, the marginal profit of a loan is determined not only by the net profit on the loan, so the standard net present value condition. There's also this profit on retained deposit. So if alpha L of that newly created deposit ends up staying or coming back to the bank, then it's gonna earn alpha L times the spread between IOER and the interest rate that that bank pays on that deposit. And so this is gonna differ across banks and that's the idea why the opportunity cost of lending or the decision of lending is going to differ across banks, okay. Okay, so I can quickly describe the equilibrium. So we start off by looking at the end so we can look at the deposit market decision of the worker. And so that's fairly easy. As I mentioned, the worker cares about interest and they care about convenience. So what they're going to do is they're going to, if the worker's convenience is less than this interest rate spread RS minus RL, so RS is the rate of the small bank, that's gonna be larger than the rate at the large bank because they don't have convenience. So they need to do that to attract depositors. So if you get a low enough convenience level, you're gonna be attracted to that spread and you'll go to the small bank and then the rest of the people who get higher draws of convenience will go to the large bank. So that's pretty simple, the choice of the workers. What's the lending decision of the firms? Well, the lenders are, first of all, they have a captive market. They're the only, your bank is the only bank you can go to so they can steal all the surplus. So first of all, the interest rate they're gonna charge you is they're just gonna take A. So they're gonna get QA minus one minus half. This is just that profit condition. And so if it's positive, if Q is high enough that this is positive, they'll make the loan, otherwise they won't make the loan and the small bank will have similarly have its own criteria, which is based on its marginal profit condition and related to its market share. I noticed that this doesn't have anything to do with existing deposit levels that's solely the net present value criteria adjusted by the retention of deposits. You don't have to go out and attract deposits to make a loan because reserves are abundant. So can I ask a question about the convenience yield? Sure. So what do you have in mind? Why is it different between different sized banks? Is it like debit cards issued by big banks and we're accept that what? Because that doesn't seem to be the case, but what do you have in mind? Yeah, so let me just give a quick answer and then we can always come back to this, but strictly speaking, what we assume is there is a convenient bank and there's a not convenient bank and then that convenient bank becomes large. We have in mind the idea that large banks because of their size generate convenience, but we don't formally model that. So I don't have to think about how acceptable the claim of the bank is. It could just be has more branches near my house or something like that. Exactly. Yeah, like one way to think about convenience is just this idea that more ATMs, that would be like the simplest story. Yeah, thanks. Of course, with electronic, the ATM example is not that great, right? No, but I think you can have better services, better rates on other services or more services due to scale, economies, various things like that. Yeah, thank you. Okay, so then ultimately, they have to choose the actual deposit rates. That's the most complicated problem because that also has to factor in, not only retain deposits, but also the fact that depending on the rates, you said you're going to attract some deposits from new loans generated by the small bank. So this is a pretty complicated profit function, but we can solve it. And so we get first order conditions. I only put those up because what's the first proposition? Well, we're going to look at two types of equilibria. We're going to look at equilibria that are unconstrained, where both of the deposit rates move freely. And we're going to look at a constrained equilibrium. So in the unconstrained equilibrium, we're going to find an RL and RS that solve both of those first order conditions. We're going to find that RL is less than RS. And we're going to find that the lending standard of the large bank is going to be lower than the lending standard of the small bank. But then it's also possible that we can have equilibria. And this is going to, in the model, this is going to depend on F. It's going to depend on how high the interest on reserves rate is. When interest on reserves is low, then the large bank would like this, which rate is smaller than the small banks, would like to set a lower and lower rate. Eventually it would like to set a negative rate, but it can't, so it gets bound at zero or it gets bound at S. If the CBDC rate S is there and it's positive. So you can also have a constrained equilibrium and the real difference is that now the large bank sets a rate S and the small banks rate has to be greater than the large banks rate, which is S. But we still have this idea that the lending standards lower in the large bank because they have a lower opportunity cost of funds because they have a higher market share. Okay, so this is just a picture of the deposit and the lending market. I'm through the model, I'll be able to get through most of the results in the next, within 10 minutes. So this is just a picture of the deposit market. And what you do, you see here, and I think this is, again, I'm trying to defend the model before I even get the CBDC, is that for low F, we're in the constrained equilibrium. So what does that mean? It means high market share for the large bank and the markets and the large bank's interest rate is non-responsive to increases in interest on reserves. And so the small banks rate is rising, but they have a small market share. And so what we get is this non-responsiveness of the market of the bank, commercial bank deposit rates to IOER in the low region. And then it becomes perfectly responsive once we pass this threshold. And so if we actually look at US deposit market, so this is just a weighted average of deposit interest rates. And you can see that when, before the crisis, when the federal funds rate is reasonably high and we're arguing we're in the unconstraining equilibrium, the actual rate moves pretty well with the federal funds rate. Maybe there's a little bit of a lag and Jerome Powell mentioned this idea that there's likely a lag, but in our model tracks this pretty well. In contrast, when we moved to after the crisis, the financial crisis, we have this idea that the federal funds rate went up and up and up and up and actual rates didn't go up at all. And that is consistent with our model where it's very non-responsive for low rates of federal funds. Now our rate goes up a little bit more than the actual rate does, but nowhere near as much as it sort of should have. And again, I would argue that the blue rate would have gone up too. We don't model any kind of a lag and it probably would have gone up if it wasn't for COVID and there were, F came crashing down again. Okay, so we think the model matches key features of at least the deposit market. Okay, so now let's talk about the impact of a CBDC with a positive interest rate So one quick question, Rod. So your model has prediction about this stickiness of rates being different for large and small banks. Do you have data about that? We started to play around with this, but I don't have any report. I mean, you're exactly right, Antoine, that there's testable hypotheses screened in here. We also have some predictions for the loan market that are also testable and that's certainly where we'll pursue that. Yeah, good point. Okay, so this is a short talk. So please take a look at the paper. So we have these full tables of comparative statics on deposit rates and market shares and loan quality and total loans made and all that kind of thing. But just to get through, I'm just gonna look at some pictures. So these are granted special cases, but they're the cases that are consistent with the vast majority of the parameter values for the model. And so essentially what happens when you introduce a CBDC interest rate and I'm looking at the constrained region because in the unconstrained equilibrium that the S is just fluctuating around doesn't make any difference. So it's the constrained case that matters. And in the constrained case, S becomes this floor on the large banks rate. So the large bank rate has to move one for one with increases in S. The small banks rate adjusts because it adjusts as an equilibrium rate but it doesn't move up as fast as the large bank rate does. And so we end up, the large bank ends up gaining market share, all right? So we increasing S causes the deposit market rates to rise. The black line is just the weighted increase. But we have this, I don't know if it's a negative side effect but we have this implication that the market share of a large bank becomes bigger and bigger as this occurs. Okay. And if we end up looking at what happens in the lending market, it's much as you would expect, both of the rates are increasing. The deposit rates are increasing. So the cost of the loans is going up. So their loan thresholds both rise. They require higher quality loans to approve loans. That's what you're seeing here. But we have volume, switching from the small banks to the large banks. So the large banks loan threshold, they're being more picky, but they get more customers. Okay, because they've attracted more depositors to what's happening in the deposit market. So the large bank share increases, the small bank share decreases and the total effect can actually go up or down. But for this parameterization, it happens to go down. And we give conditions in the paper under which that's true. Okay. So now if we move to a convenient CBDC, like I said, I'm gonna actually talk about this with that S equals zero just because it's easier to think about the two things and isolations for a moment. Because as I mentioned, I motivated this idea that we can think about improving the store value aspect or we can improve the medium of exchange aspect. So if we move to a convenient CBDC, the way we do it is the following. There's this convenient, now CBDC is something that's maybe on this specially built platform that the central bank provides. And it gives same access to everybody. So it has some convenience value V that we assume it's the same for everybody. And individuals still have their convenience for commercial bank deposits, which are zero at the small bank and independent draws at the large bank. But now when you go to a bank, you can use whichever is most convenient. So your convenience is now gonna be the maximum of the two. So it's gonna be V at the small bank and it's gonna be maximum of Delta and V at the large bank. And this is now gonna change your decision. So V used to be zero in this decision of whether or not, what draws are gonna have you go to the small bank, now you're gonna have to, it's gonna depend on R at the spread plus this common convenience parameter V. Okay. So again- Just to let you know, there's three minutes. I'm good. I'm good. Okay. Not. Okay. Like I said, I wanted to defend the model and then the implications follow fairly quickly. So now if we introduce V, this convenience, so we have, that's what this picture is showing. So first I'm looking at the deposit market and what V does and it's, I think it's pretty intuitive, is it takes away the competitive advantage in some sense of the large bank, right? Which used to be the only bank that had convenience. And so the higher V is, the more eats away that competitive advantage. And it allows the small bank. And so we're staying here in the constrained solution in this part here, but it allows the small bank to lower its interest rate. Okay. Now, as it lowers its interest rate and the central bank or the large bank, sorry, it keeps its interest rate flat because it's constrained and this is at zero here, but it would be at S if S was positive. What we have is we end up having a change in shares so the small bank share actually goes up. Why is that? It's dropping its interest rate, but it's not dropping his interest rate as fast as the convenience value is rising, right? So if you go back to, if you go back to this picture, so V is going up, RS is coming down, but it's not going down one to one for V. So we end up getting an increase in the market share of the small bank and decrease in the market share of the large bank. And this is something that we refer to as this idea of leveling the playing field. So that same idea is reflected in the lending market relatively, I guess I shouldn't say obvious, but intuitive reasons. And so one of the interesting things too about this is that for a given convenience value, for high convenience value, it ends up increasing the deposit rates. So for a given value of F, you get a higher deposit rate. So this ends up enhancing monetary policy, transmission through IOER2, the deposit market. Okay, excellent. So now just I have a couple of minutes left. And so let me just, this is just where we're headed. As I mentioned, what the paper does so far I think provides a useful model for thinking about central bank digital currencies that's tailored, I think, pretty much to the US situation. We talked about this idea of interest bearing and convenience. And we essentially cataloged how those things would impact markets, but we haven't said anything about what we should do. And so we're starting to think about design. And so the last picture I'll show is what we've done so far, which is create this idea of thinking about deposit or welfare. So deposit or welfare is equal to the sum of what they get in deposit markets and the sum of what they get in convenience in the market. So this is per dollar deposit or welfare. So you'd multiply this by the amount of deposits. But if we look at the equilibria and we trade off convenience value and CEDC interest rate S and solve the model sign for both of those features, I only showed you the two in isolation, but solving them in both of them, we can actually trade out these indifference curves, okay, which give sort of the additional welfare from different values of these parameters. And so the idea is that we can ultimately think about picking a point on that indifference curve. So we're gonna be looking at, we can create a similar picture for an ISO profit curve for the firms and ultimately we're working towards trying to understand what the best choice of these parameters are, but that's in the future. Okay, so what I did was provided a model, CBDC put a lower bound on deposit interest rates and higher CBDC interest rate increases deposits but reduces the small banks market share. So it increases inequality in the bank size. I should mention CBDC doesn't have to disintermediate banks in this model. So it's very similar to end El Fado where the CBDC rate interest rate rises, the banks don't lose deposits. They just raise their commercial bank deposit rate and they keep their deposits but they're forced to compete more. Own convenience value 11 is the playing field that can also enhance monetary policy transmission. And we're working on thinking about how we should choose these parameters. So thank you. Thanks, Rad. I now quickly turn it over to Monica to discuss the paper. Monica, the floor is yours. Can you see my slide as well? Thank you. So thank you very much for asking me to discuss this nice paper. Let me first summarize the paper in my own words and then I'm gonna make comments. So you'll hear it in how I would see the setup. So there's a central bank that fixes abundant reserves and it sets a reserve rate F. There's a large and a small bank. Think of these as risk neutral banker agents. So these are not firms that are owned by households. They're agents that initially hold just reserves and they issue deposits. They don't have equity. And there's a large bank that has an exogenous share ML which is larger than half of the reserves. So that's how the large bank initially is large. It's by assumption they start out with the larger share of these reserves or deposits. These banks have market power in loan markets. So there's a mass one of entrepreneurs. They have a technology. So they have ideas on what to produce but they need to hire workers. These entrepreneurs are assigned to a bank. So there was an initial assignment to the bank that they're at. And so they can borrow from their own bank a loan of fixed size, which is $1. So there's no choice of the loan size. So each entrepreneur is gonna wanna borrow this $1 and then do a risky project. These are IID risky projects. There's each of these projects has a quality. So QI is the quality of the project for entrepreneur I and those are drawn from some distribution. And then the projects have payoffs. Either they pay out A with some probability QI that's the quality of the project. So it's your success probability is the quality of the project or otherwise zero. These entrepreneurs are really bad at bargaining. That is sort of my interpretation of what they do in this model. They really suck at bargaining. They pay the entire dollar that they're getting from the bank to the worker. So the wage of the worker is one the entire dollar that these entrepreneurs are borrowing. So the wage is not an equilibrium outcome in the sense that there's an equilibrium wage. It's just there's bargaining and the entrepreneur has no power. And so just hands over the dollar to the worker. And the same thing is happening with their loan rates. So the bank just extracts everything from this entrepreneur. And so the entrepreneur pays a loan rate which is the expected payoffs minus one. Then banks choose how to splits their assets into reserves and loans. And so they're only gonna make loans if the loan quality is the marginal loan quality is higher than some cutoff level Q-star because they can always choose between reserves and loans. So now these workers have this dollar in their hands. So they work today for this dollar but what they wanna do is safe in deposits. And so here's where the choice of the bank is gonna take place. So these workers choose whether to deposit this dollar in the large or the small bank. The banks differ in the ATM network and their branches, the reasons that Ron mentioned. So there's some difference in these banks. And so workers, the way this is captured is that workers get a convenience yield delta which is drawn from some distribution at the large bank. They get no convenience yield at the small bank. So this is a normalization. You can think of just it's important that there's a smaller convenience yield at the small bank. That's why the large bank gets away with paying a lower deposit rate than the small bank because they have deposits at the large bank come with a higher convenience. And so they get away with a lower deposit rate. And so workers with a high that have a draw of this convenience yield that is high enough, they prefer the large bank. So they are fine with getting paid a low interest rate. It's just that they get this high convenience from having their deposit at the large bank. And so the large bank we can compute their deposit share is just using this distribution of shocks for the convenience yield. That tells you what's the share of workers that has a high enough delta so that they stay with the large bank. And so that determines the deposit share of the large bank. And so now there's a sequential game between the large and the small bank. Initially in period zero, they choose their deposit rates. So each bank chooses their deposit rate then banks make loans. So they essentially choose their quality cut-offs for the successful ability of these entrepreneurs. And then the workers get paid. And so the workers choose where to deposit their payments. And so in equilibrium, the large bank has a lower deposit rate because it has the higher convenience yield. It attracts a higher deposit share than the small bank and it makes more loans because it has more deposits. It now can make more loans. So the large bank has a lower quality cut-off. So now you introduce CVDC in the setup and it has a very nice feature that it affects these banks differently. And it matters how you design the CVDC. So the idea is that CVDC is administered by commercial banks. CVDC with a large bank comes, so it inherits basically the convenience yield of the large bank delta. This was the higher convenience yield. So if now CVDC pays a low rate, nobody will use it, nothing happens. But if CVDC pays a higher interest rate than the large bank was paying in the original equilibrium, now the large bank has to match what the CVDC is paying to stay in business. And so what happens is that the large bank gets even more deposits and it grows in market share. So again, a CVDC that just inherits the convenience yield of its host bank and pays some interest rate will just benefit the large banks. Instead, if the CVDC comes with its own convenience yield that is somewhere in between what small and large banks have in terms of convenience, then you'll see an equilibrium in which deposit rates and market shares and lending standards of these two banks will converge. And so that in that case, you're helping the small banks get closer to what the banks, large banks do it. The impact on overall lending in this model is ambiguous. So it could go up or down. Okay, so let me... Give me a three-minute warning. Okay, so I'm gonna make some comments. So what explains the differences in deposit rates in this model? The paper motivates the convenience yields through larger ADM networks of large banks, more branches. So it's more convenient to walk to the bank because they have more branches. So in some sense, the way this reads is as if the large bank has the better technology that they just produce the better products. And in equilibrium, the large bank gets to charge more because they have the better products. And so I can in principle get differences in deposit rates without market power because if what we're seeing in the world is differences in deposit rate that arise because if you go to JPMorgan, JPMorgan is just a better deposit, then I can get this without market power. And so I was wondering, it would be good to discuss this more in the paper. Is this market power or is there differences in the quality of the product that they offer? And I would think this, it would be useful. And so Rod, I didn't see the result from the last slide in the paper. So here's my comment asking for exactly that. So it would be useful to characterize what the efficient allocations are because it would matter for welfare conclusions what this delta really is. And so this goes back to Ricardo's question also earlier. What is this delta? Is this a feature? Is this really a technology or is this market power? Let me comment on relationship lending. So here the banks start out with an exogenous clientele of borrowers. So these entrepreneurs are just assigned to these banks. And then it's a special assumption to say that both banks, the small and the large bank, they're facing the same distribution of project qualities. So in some sense, so they're making loans and they're both facing the same distribution. So in some sense, they're not fishing from the same pond. There's not one distribution of project qualities in the economy. And then these two banks make loans to whoever is out there, entrepreneur and has a project quality. But they each have their own pond from which they're fishing. And so this is sort of a special assumption and it would be interesting to discuss this more. Also in the model when the workers choose the bank. So the only choice of bank that happens in the model is at the point where the worker gets paid and then chooses the bank. So that decision is only based on the deposit interest rate of the bank because the workers wanna save. So it matters to them how much they're getting on their savings. But if relationship lending was exclusive as it is in this model, then you would think that bank customers at the point in time where you're choosing where to bank, you would not just compare deposit rates but you would also compare lending standards. But because you know that the only loan you're gonna get is from your own bank. And so it matters to you how much this bank will lend and in particular whether you'll be what once you draw your project quality whether you will be the type of entrepreneur that gets a loan or not. That's gonna be important in the choice, ex ante. And the final comment wanna make is this feature of the model that it doesn't have an ambiguous result for overall bank lending. So here introducing CVDC can make overall bank lending go up and down. And so we know that often that's the key for whether CVDC is overall desirable or not. And so we have papers that go in both directions. Some papers argue that you introduce CVDC to get less small overall bank lending. And then the last two bullets are papers that argue that you get more bank lending when you introduce CVDC. So that we know that the welfare conclusions really depend on this. And so this is not a criticism of the paper at all. This is not how it's just that this is a key this is the debate essentially is CVDC what is it gonna mean for bank lending what increased bank lending or lower bank lending and maybe it's good for this paper to say we don't know but it would be really important to know what happens with overall bank lending. Thanks Monica. Rod, did you wanna respond to the discussion? Sure, so thanks Monica. That was fantastic summary of the paper and great comments. So I'll just try to say a few words about each of them. So first of all, yeah, it's not so much that market power leads to the lower rates. It's being a better bank, which we are just assuming is happening for whatever reasons the ATM story maybe isn't a great storage for the digital world, but just this idea that somehow larger banks can provide more and better services. That's what gives them more market power and allows the large bank to charge a lower rate. Yeah, I understand your point about not that they're not fishing from the same pond. I mean, they're just, they're drawing from these common distribution but they only can take loans from their captive audience. That is their existing depositors. And that's where we're getting this market power in the lending market from. Why we're essentially allowing the bank to extract all the surplus. I mean, this could still be a bargaining problem, but the borrowers don't have any outside alternatives. The choice for a lend depends on the rates. Like you say, it also depends on convenience. So it's not just on the rates, it depends on the convenience parameter, but you're absolutely right that in this version of the model, the depositors do not think ahead to the future, to like a future stage when they're gonna be coming on to an entrepreneur and they have to go to a, they have to figure out where they're gonna go ask for a loan and they might be thinking that if they're at the lower bank that would be better because the lower bank has a lower lending standard. So if you were thinking ahead to another period in the model which currently doesn't exist because we don't have a fully dynamic model, you're absolutely right that the depositors should be in deciding where to deposit. They should also be thinking to a future that we don't model where they want to know, where they are in terms of what their best chances are to get a loan. Now, mind you, as I think out loud, the depositor gets all the surplus taking from them anyway by the bank, so actually they don't care about that. But anyway, your point's well-painted. Can I just say that if you're really thinking of this as a technological advantage that large banks have and the word market power is a little bit misleading because they're just offering, they have a larger share because they have a better product. It's not market power. Okay, yeah, fair enough, fair enough. Okay, and then just I guess the last point is that you're absolutely right that for our general specification, the overall impact on lending is important and we care a lot about that, but we can characterize conditions on either the distributions that are driving convenience parameters or the loan values and on the other parameters and we can identify the situations in which the loans increase in which lending increases and which lending decreases. So taking your advice, we'll put more effort into that, recognizing that that's an important thing that people would like to extract from the model. So thank you for that comment. Okay, thanks. So we're gonna now open up the floor for discussion. So I do see a few panel, sorry, yeah, panelists hands raised. I'm gonna go to them in order in just a second, but reminder for other attendees, feel free to put some questions in the Q&A box and I'll try to sprinkle a few of those in. So the order I saw the hands up from the panelists, I have Charlie, Catherine and Ricardo. So apologies if the order is wrong, but that's the way I saw it. So Charlie, over to you. Thanks, really nice paper on the, there is another dimension along which market power works in your model and that's this question of, what's the likelihood of the reserves coming back to you? And that seems to be a pretty important dimension of your story. And that's actually like Todd, that's a really, really, really old issue in, older even than me, which is how old it is in banking of creation of money, is it a factor of an individual bank or of the banking system as a whole, depending on where it comes back, does it come back into the same bank or does it come back into a different bank? And you've got a leg up on that very traditional version of the story by actually having a calculation by these banks about whether they want to, what the rate is at which the money should come back to them versus coming back to some other rival bank. But what's missing in that argument then is once you acknowledge that this is a shortage of reserves that's going on, we've got these banks now acting as oligopolists, not only on the other side of the market, but also on the side where they're playing with the reserves and hoarding them or not hoarding them. So my question is this, how does your story change if we acknowledge that there exists markets in which banks can trade reserves? They've traded them back and forth among each other, which they do. And if we have that market out there, which is a competitive market, even though these banks are monopolistic or oligopolistic on other dimensions, they're still playing operating in the same market for lending and borrowing reserves. Is that gonna muck up the distinction between large banks and small banks for you? Not, those are all good points, but I don't think it does. I mean, one of the things that I didn't emphasize is that when we're thinking about a world of, if we're thinking about what the opportunity cost of a loan is, not thinking about this component due to retained deposits, but in the world of high reserves, sorry, that rate is essentially determined by an administered rate, it's by interest on reserves. That's what we call F. And I didn't introduce new notation when I showed that slide of the pre-crisis era, but in that era, the rate that determines that what we would think of F as being is not interest on reserves, but actually the federal funds rate, okay, the rate at which they could loan those reserves in the interbank market. So the story is compatible with that. I just wasn't careful in clarifying that in the high reserve regime, we're thinking of F as being the IOR rate and then the low reserve regime, we're thinking of F as being the interest on the actual federal funds market rate. And so all banks think of the opportunity cost of reserves as the same, no matter what size they are, and they don't worry about this increase or decrease likelihood of having it come back to themselves. Yeah, I have to think about that. Like they do in the context of when they're making loans. Yeah, but okay, that's where the puzzle would arise. Yeah. Thanks, Trevor. Oh, I guess then it's me. So I was wondering about your convenience here. So because who is setting this convenience here? So for interest rates, it's very clear. So it's the central bank setting the interest rate and yeah, adjusting this interest rate. The convenience here, you assume that it's the same for all customers independent of the bank, but CVDC is something that's quite technical. So I wonder whether you would have the same relation with small banks and large banks that large banks are offering better service also in terms of convenience that I think then the results would break down. And so I think it's really key that somehow the central bank can adjust or set these convenience here by the features of the central bank. And I find it very interesting because these are the things that are extremely hard to model and when you think about CVDC it's probably not the interest rate but it's the other stuff that's around that will determine take up. And yeah, I find it really hard to think about it what it does mean and practice this convenience that's the same for all consumers. Yeah, so thank you. And let me try to respond to that. So first of all, if you think the convenience yield is gonna be representative of the host bank that is the first version of the model. And then we get to this second version of the model where we're imagining that the central bank can produce this version of a CVDC that has this common convenience that's independent of which bank hosts it. And so first of all, I should emphasize we're not thinking of them as setting a convenience parameter, right? It's more that they design a CVDC and the platform on which it transacts possibly that generates convenience for people. And so we're really thinking, it is high level modeling for sure but we're really thinking about convenience CVDC as maybe being something like a programmable CVDC. So we could think about a CVDC that's operating on some kind of maybe not necessarily distributed ledger platform but some type of permission platform that allows the operation of smart contracts or various things. And just as an approximation, one way to think about that is that that type of convenience and access to those markets and those types of financial services might be worth the same to everyone. It's just our way to do it, right? And but I think it is very worthwhile to think about a CVDC that's offered in a way that it really just sort of blends in with the existing products of commercial banks. I have a savings account, a checking account, a CVDC account versus a CVDC which still offered through commercial banks because for various reasons, who's going to do KYC, customer service, all that sort of thing. We might want to imagine the central bank or the commercial banks are going to be the conduit through which consumers access CVDC but the CVDC could live on this new kind of common platform. So I agree with you, it's somewhat ad hoc but I think it's a, what we're trying to get at is differentiating those two sort of big ideas about how a CVDC might work, yeah, thank you. Turn to Ricardo, then I have Todd and I have a couple of questions from the attendees I'd like to get to. So Ricardo. My question was about the convenience yield. So it's a bit of a follow up now to that. But so, you know, we introduced the deltas. Now that's a bank specific thing, you know, branches, ATMs, whatever example we want to use. We introduced, you know, in the version of the CVDC convenience yield is an asset specific thing. So that, you got a little bit to that with the programmable and so on but then it makes you think, you know, what's a central bank doing? Is it, you know, opening branches, making it available to a small bank? You know what I'm saying? Then it becomes a follow up with that and the reason I think it matters is at some point you made the comment, the parallel with Andolfo, where, but Andolfo is different. There's this market power problem, the central bank offers a higher rate, just the threat of it being there doesn't mean the process have to flow, just, you know, the market reacts and then the process don't flow. But, you know, if you make it an asset split, if you stuck to the story that the V is the institution specific, then that story doesn't go anymore because, you know, only the Fed can do it. The only way you can get the convenience yield is to draw your deposits and go there. So these details kind of matter for, you know, Yeah, I agree, but I think I can defend it. I mean, again, it's, so we're not, we're not really thinking about a central bank digital currency in the second case, in the convenience case. We're not really thinking about it as a situation where, you know, like in the Banking for All Act that the government is leveraging the post office and, you know, issuing the money through its own outlets. We're still thinking about it being offered through commercial banks. And so, for example, we don't allow people to hold only CBDC. They go to one bank or the other to get it. But we're thinking, and again, this is just how we're thinking about it, how we modeled it here. I'm open to thinking about it different ways. But we're really thinking about it as sort of access to platforms, like how money's transacted. So I do, I go back to the programmable idea. You know, what's the point of programmable money of, you know, things like stablecoins? It's that you can transact them on platforms that you can't transact regular commercial bank deposits on. And so we're just envisioning this idea that the central bank might issue a form of money that gives people sort of common access to these markets and doesn't depend on things like how many ATMs. Although I realize that's not the best example, but it doesn't really depend on the features of the host bank. That's what we're trying to get at. But I'm open to thinking about different and better ways to do it. I mean, the thing is, sorry, if you go that way, it's hard to think about debit cards, which is the closest thing that we have these days. It doesn't matter what's printed on the card. And you just, you know, whoever takes a debit card, they take them all. So what is this? I mean, what is this convenience yield? I mean, yeah, you could say it's having a branch nearby, but I mean, I'm not sure that's the big deal these days. I don't know. You know, there's lots of online, fewer online banks with no branches at all. And they're doing pretty well. So what is it? Well, I still think, I mean, again, I guess I could try to defend this all day, but it's, you know, there are other aspects that come along with scale. I mean, just you could think about resilient or the likelihood of default, you know, the large bank your deposits are safer, even with deposit insurance, you don't necessarily want to go through having your bank default. I mean, I think we're just arguing that there may be these sort of scale features that make large banks or some banks more services associated with the deposits, therefore more convenient than other banks. But, you know, your point's well taken. If we're gonna make that a key feature of the model, we should defend that as well. Last thing, and I'll shut up. But I mean, you also need the bundling, right? I mean, you have this strong assumption that you're bound to one bank for everything. So in principle, if one is good at payments, I can just go there and get my other services from the other one. You know what I'm saying? Well, there's a realignment, right? So once the deposit interest rates change, people can go to whatever bank they want. So they're not stuck at the bank forever. They're all... I'm talking about the bundling between whatever services the big bank is giving you and the payment aspect of it. So if you worry about the run and you think Chase is safer, you know, you keep your big deposit there and just transfer a little bit to use the payments. Yeah. Yeah, no, thanks, Ricardo. We'll think about these points. We're short on time. Maybe I'll take my questions offline and David, if you have a couple of minutes to go to the Q&A. Yeah, sure. I think, you know, Rod and Monica both have said they could stay over just a few more minutes. So what I will do is, Rod, give you a couple of questions here. One from Janet Jiang about the effects on banks. Is the effect on the market share work in the same direction as the effect on profits? The conjecture that she offers is no. You talk about leveling the playing field. Is this more about market share or about profits? Thanks for that question, Janet. So we're just right in the middle of computing bank profits. So one of the questions that we've gotten in the past has to do with, you know, what is the overall impact on bank profits, thinking in terms of sort of financial stability concerns that we don't want the CBDC to reduce bank profits to the point where banks may be at a higher risk of defaulting. We just haven't done that yet. But that's exactly where we're headed next. And like I said, in terms of thinking about our overall design characteristics and trading off, you know, maybe convenience and interest rate, in terms of depositor welfare, which of course is a concern. We also want to think about, like I mentioned at the very, very beginning, this idea of thinking about an isoprofit curve, you know, between the small and the large banks and trading this off against overall profits, you know, financial stability. Yeah, these are exactly the types of things we're thinking about. So yeah, thank you for that question, but I don't have an answer at the moment. Another question from David Rappaport. Two exercises in the paper of a flavor of retail versus wholesale CBDC. Yeah, with the retail CBDC implementation, the commercial banks are not disintermediated but offer the CBDC to depositors. Can you elaborate on how you envision this option to be implemented through commercial banks? Well, I mean, I think this is one of the models that people think is most likely just the idea that if we're thinking about a CBDC in terms of sort of a standard, you know, account, you know, giving consumers access to reserves, then we're thinking about a situation where you would wake up one day and at the Bank of America, just the name of bank, you would have a savings account or deposit, a checking account in a central bank account, with the difference being that the money that's in that central bank account would actually be central bank liabilities. And so you could spend that money or probably move between the accounts as you saw fit, but the commercial banks would have this fear of, I think as you mentioned, of disintermediation, where they wouldn't want to see all of the deposits flow into this central bank account. And so they have to compete by raising their interest rates. And that's largely what the equilibrium analysis does is it sorts that all out. Thanks, and I got one more question or more, I think a comment that's going to the discussion about the convenience. This is from Larry Wall. I can understand differential convenience yield tied to bank apps. So large bank apps can develop, or large banks can develop better apps and link to better services. One could then imagine that the Fed would supply the app, which is equally good for big and small banks. But once you talk about commercial banks providing consumer services, when problems arise or providing KYC, then it becomes hard to see how the central bank could equalize the quality of services across banks. But if the commercial banks aren't provided any additional services beyond that provided by the central bank, then why are commercial banks involved in issuing CVC kind of related? And at the end there, but I see Larry down there, so. Yeah, so I'll just take a lot of that on board. So the discussion of the apps is maybe useful for our defending their convenience aspect of the larger banks. Yeah, I don't know. I mean, I think we could have our model, we could alter our model so that the central bank convenient CVDC was done as a central entity. We just have to allow the possibility that people don't join any commercial bank. We just haven't done that. But I just find it unlikely. I mean, central banks aren't allowed, first of all, to take deposits directly from individuals. I don't think it's their business model. And so I just think that the most natural way that this would ever occur in at least a shorter medium term would be that the central bank digital currency in whatever form it takes would still be offered through commercial banks. So that's the way we modeled it, but yeah, we should certainly be open to modeling it alternative ways. So I appreciate that comment. Can I follow up? My point was more, I don't see how you can equalize the convenience yield. If the commercial banks are involved, they're there to help provide some services that the central bank can't or doesn't want to provide, then not all the banks are going to do it equally good. How correlated that will be with size is a different question, but they're not going to do it equally good. And so the convenience yield doesn't completely drop out, I don't think. Okay, yeah, so sorry, now I understand better. But this is just an extreme version, right? So the idea is that we're coming up with this central bank digital currency that eats away at the advantage of the large bank. It doesn't have to be pure like we've modeled it. I don't think, I think we could easily alter that so that it's not exactly the same for everyone. It could still depend on, somewhat on the host institution, right? We're just, this was just, yeah, I don't think we're really dependent on that on the, in some sense, extreme way to be modeled it, yeah. Okay, well, I think that exhausts the set of questions that we had from attendees. So I want to thank both Rod and Monica for the presentation and discussion today. I think, as you can tell, generate a lot of interest and appreciate that. I will turn it back over to Russell to sort of wrap things up. Thank you, David. Thank you, Rod. Thank you, Monica, everyone. So next time we will have wristband holding, hosting Andy Levin of Davmo and will be discussed by Galty Eggerson of Brown and Andy will be talking about something opposite from something today, which is charging interest rate from using DVDC. So stay tuned and check our website and see you next time. Thank you.