 So, good morning also from my side and welcome to the fourth panel of this conference which will look at CBDC and monetary policy implementation. And I'm happy to chair a panel that reverses the usual gender balance in this topic. So the first presenter will be Noa Lama-Storf. Please, Noa, there she is. Okay, that looks good. Perfect. Thanks. Yeah. Hi, everyone. We organized this for including our paper to this really great conference. This is joint work with Tobias Lindsat and Cyril Monim. And since Tobias is working at the ECB, the usual disclaimer applies. So this paper is about CBDC, monetary policy implementation in the interbank market. I mean, I don't have to motivate a lot why CBDC is interesting. It's definitely high on the agenda of, I would say, many academics and basically all central banks worldwide. So far, most of the papers or many papers are focusing on the macro implications of CBDC. So for example, the paper by Williamson, Houston, Sanchez, or Tua-Dal. However, there are fewer papers on the micro implications and notably on the money markets. Although this conference greatly shows that also this front is developing quite a lot. So what we do in our paper is we study the effects of CBDC on the interbank market for reserves. So we take the general equilibrium model of banking and payments, including CBDC. And here we are using the Tua-Dal recent JPE paper to do that. And we add to that model end of the day refinancing shocks in the spirit of pool. And we do that to analyze the effects of CBDC demand and different design features. We assume that CBDC is a liability to the central bank and it's channeled via banks. So to make it available to their clients, banks basically need the cash to acquire it. So this is basically equivalent to a 100% reserve requirement. So we do not really stack a stance on how the market share of CBDC is formed. So we take that as exorgety giving. But we study what happens when this market share of CBDC is increased. And we find that if this market share is increased, this will drain reserves. And when reserves in the system are scarce, it will increase the interbank rate. However, the effect of CBDC remuneration on the demand for reserves and the interbank rate is somehow ambiguous. And as you will see in a minute, it depends on banks' landing costs. So how costly it is for banks to find profitable investment projects. And then we also study two important topics that are pretty prominent in these days, policy discussions, which are caps and also tiered remuneration. And we find that both tend to decrease the demand for reserves and the interbank rate. And they actually have similar effects as lowering the remuneration rate, which is makes CBDC less attractive. Okay, so I already mentioned the model is based on the recent Jewett-Arle paper. But we simplify this paper a little bit in the sense that we have perfect competition in the deficit market. And we have sellers accepting only two means of payment. So this is deposits and CBDC. But then we add to that model an interbank market. And also the banks in this model are facing end-of-the-day liquidity shocks in the spirit of pool. And these shocks take place once the interbank market is closed. So then the banks have still the opportunity to turn to the central banks' landing facility in case they need more reserves. Okay, so very important now a model are entrepreneurs or the relationship between entrepreneurs and banks. Entrepreneurs need to borrow bank deposits from the banks in order to buy investment products to produce something. So one bank leads to N entrepreneurs. And to do so it inquires some cost, C of N. Because it has to find these entrepreneurs, it has to negotiate maybe. Okay, and then the entrepreneur uses these bank deposits to buy investment products and to produce something. However, a fraction of gamma of these entrepreneurs will be hit by a reinvestment shock. So in order to keep producing, they will need to buy additional reinvestment goods. So this is the red Y here. And if they don't do that, they will end up producing nothing. And this gamma, so this fraction of entrepreneurs that got deficits from the, that got loans from the bank, is bank specific and distributed according to some distribution function. And once these entrepreneurs are hit by the reinvestment shock, they will have the opportunity in market one. So this is a decentralized market to go to sellers to buy the additional needed reinvestment good. And there are two types of sellers in our economy. So fraction omega one of these sellers only accept bank deposits. And a fraction omega two only accept CBDC. And since these entrepreneurs are funded by the bank, it will need additional funding to buy these additional needed reinvestment goods. This reinvestment shock for the entrepreneur is basically a refinancing shock for the bank. So out of these N entrepreneurs, a bank has funded. A fraction gamma will be hit by this refinancing shock. And a fraction omega one will need deficits to buy this additional needed reinvestment good. And for those gamma and omega one entrepreneurs, the bank can just issue new deficits. And for the other entrepreneurs who need CBDC, so this will be gamma N times omega two, the bank will turn to the central bank to acquire more CBDC and give it to the entrepreneurs. And while doing that, the bank is facing new reserve requirements. And in order to fulfill these, the bank can either use excess reserve that it still has, or it can turn to the central bank to the central bank's lending facility. And importantly, this refinancing shock takes place after the internet market closes. So if the bank then realizes that it has, that it needs more reserve, it can turn to the central bank who offers the lending facility at the lending facility rate. And in case the bank still has excess reserve, it can deposit them at the deficit facility, earning the deficit facility rate. And these two rates set by the central bank will determine the interbank market rate. OK, so this timeline gives you an overview of the model. So each period is split into two periods where two markets, the market one is this decentralized market and market two is decentralized market and the tradition of largeson right. And in the market two of the last period, banks give loans to entrepreneurs, occurring that cost C of N. And then in the beginning of the next period, there's an interbank market open where banks can trade reserves among each other. And after that interbank market has closed, a fraction gamma of the entrepreneurs will be hit by this reinvestment shock. So we'll need additional reserves because, I don't know, machine broke down on the weekend or something like that. And these entrepreneurs will meet sellers who sell these reinvestment goods. And a fraction omega one will only accept deficits. And a fraction omega two will only accept CBDC. And the bank will give these additionally needed funds to the entrepreneurs and then face its new reserve requirements. And in order to fulfill those, since the interbank market is already closed, it can turn to the central bank and either borrow from the central lending facility or deposit at the deposit facility. OK. And in addition to banks, entrepreneurs and sellers, we of course also have buyers in our model. And these buyers buy on the market one. So the decentralized markets where they also meet either deficit or CBDC sellers. And these first order conditions will determine the demand of buyers for CBDC or deficits. So in the demand for deficits and CBDC will depend on the interest rate, of course. Because if so one unit or X units of an asset A can buy X times the interest rate. So the interest rate on deficits or CBDC will determine the purchasing power of deficit and CBDC. And then if the interest rate is higher, the purchasing power will be higher. And here the deficit rate is an equilibrium outcome. So the commercial banks will set the deficit rate in order to steer the amount of deficits that buyers want to hold. And the interest rate on central bank will be set by the central bank. And the interest on CBDC, sorry, so RE will be set by the central bank. And we assume that if either the interest rate on CBDC or deficit is increasing, or if the market share of sellers accepting either CBDC or deficit is increasing, then this will increase the demand for either deficits or CBDC. OK, but at the heart of our paper of our model is the bank's problem. So suppose the bank finances n entrepreneurs and draws through this refinancing shock gamma. And we assume, and in the paper be sure there's a little bit more formal, that the productivity of these entrepreneurs is always large enough, such that the bank finds it profitable to refinance these entrepreneurs. OK, and out of the n entrepreneurs, the bank has financed. Gamma times omega 1 will need deficits to buy. And each entrepreneur will need one of our deficits in order to buy one unit of the investment good. And for them, the bank needs additional reserves in the amount of chi times that amount. So chi is the reserve requirement on deficits. And for the omega 2 gamma n entrepreneurs, we need CBDC to buy the reinvestment good. So they each need one over RE CBDC to buy one unit of the investment good. And for them, the bank needs the exact same amount in reserves. And the sum of these two will determine the overall amount a bank will need. And this is basically the spool shock in our model. So this is the additionally needed amount of reserves a bank will need after this refinancing shock has taken place. However, before the shock takes place, the interbank market opens. And here the bank chooses the amount of reserves it wants to trade. So this is yj, not knowing whether it will belong or short in reserves after this refinancing shock. And if it's long in reserves, it can deposit the excess reserve that's a deficit facility earning that deficit facility rate. And if it's short in reserve, it will need to borrow additional amount of the reserve paying the lending facility rate. And whether a bank will be short or long in reserves will depend, of course, on the size of this reinvestment shock. So this gamma. And if gamma is relatively small, specifically lower than gamma bar, then a bank will still be long in reserves. And if it's higher than gamma bar, a bank will be short in reserves. So you will see in the simulation that this gamma bar is really important. So this is the threshold above which a bank will be short in reserves and below it will still belong in reserves. And the first auto condition to that gives us the money market rate, which is basically a weighted average of the deficit and lending facility rate of the central bank, where the weights are basically the probabilities how likely it is to access one of the two facilities. And as it's also quite typical in these models, since all the banks are the same, the interbank market clearing condition applies, then in the end no trade will be happening on the interbank market. However, still a money market rate will be determined by the central bank's facility rate. OK, in addition to that problem on the interbank market, a bank also chooses the amount of investments or the amount of entrepreneurs it wants to fund and the amount of reserves and deficits. And it does so in order to maximize this object, where the first term here is what the bank gains after production takes place. And the second term is what the bank occurs. So this is the cost to find new entrepreneurs. And then it has to pay for deficits and also for these newly needed deficits after the refinancing shock has taken place. And this is subject to two constraints where the first is just the balance sheet identity. So in loans to entrepreneurs and reserves need to be equal to deficits. And the second is the reserve requirement. So excess reserve need to be larger or equal to zero. OK, and whether these excess reserves are larger or equal to zero, that will determine two different equilibria. So one equilibrium regime is the one where the reserve constraint is binding. So there are no excess reserves. And the other one is one with abundant reserves. So excess reserves are larger than zero. So let me now turn to the simulations. And first, we will look at what happens when the CBC market share increases. And remember, so the CBC market share omega 2 is exogenous in our model. So we just look what happens when this market share is going to be increased. And in this graph, we can differentiate three different zones. The first zone is at the very beginning when omega 2 is very low. And you see in the graph on the bottom left that gamma bar is equal to 1. So that means there is no refinancing shock that is larger than omega 1 that brings a bank to being short on reserves. So there are so many excess reserves in the economy that a bank will never be short on reserves. And you see here that the amount in the lending facility here, this orange line, will be equal to zero. And the amount on the deficit facility will be relatively high. Also, you see that the money market rate and the deficit rate are constant. And they're equal to the deficit facility rate. So this is basically a flawed regime. However, as omega 2 increases, more CBC is used, and more reserves will be needed. So the demand for reserves increases. And at some point, this threshold gamma bar will be lower than 1. So now there are some refinancing shocks that are so large that a bank will be short on reserves. And you see also that now the amount of the lending facility will be larger than zero, and the amount in the deficit facility is decreasing. And since the demand for reserves is now increasing, also the money market rate will increase. But also the deficit rate will increase because banks want to attract more demand for deficits. And for the demand for deficits, there are actually two effects. So first, there's this very direct effect as the share of CBC meetings increases, the share of deficit meetings decreases, so less deficits are needed. However, in order to increase the demand for deficits, banks will increase the deficit rate. And that will actually increase the demand for deficits. And you see there's a small region here where actually this latter effect dominates. So the demand for deficits is slightly increasing in the share of omega 2. However, at some point, the first effect or this direct effect will dominate, and the demand will decrease again. And as the share of CBC meetings becomes higher and higher, more reserves are needed. And this gamma bar will be decreasing until it reaches zero, which means that even the tiniest refinancing shock will make a bank become short on reserves. And at that point, you see here that the amount in the deficit facility will be equal to zero, and the amount in the lending facility will be relatively high. And also, the money market rate will be quite high, and it will, I hope you can see that. At some point, it will be constant again and will be equal to the central banks lending facility rate. So now I haven't mentioned loans yet. So you see that the loans, so this is the blue line here, will decrease in the share of CBC meetings. And that is because now funding entrepreneurs will become more costly for banks because they will need more reserves. And that's why banks will decide to give fewer loans out to entrepreneurs. So to summarize the effects here, we can say that the increasing the share of CBC meetings will drain reserves in the economy and will, therefore, increase the money market rate and also the deficit rate. So I think this is rather intuitive. However, when we look at the effect of the CBC remuneration rate, this is a little bit less intuitive. Because as the CBC rate increases, CBC has a higher purchasing power, which leads actually to two different effects. So the first effect is a funding effect, which says that to fund the same amount of entrepreneurs, now banks will need less funding. Because if in the case that entrepreneurs need CBC to buy more investment goods, CBC has now a higher purchasing power, and therefore they will need less CBC. And that will decrease the demand for reserves, and it will also decrease the interbank rate. However, there's a second effect, which we call the lending effect, which means that since now it's cheaper to fund entrepreneurs, banks will fund more entrepreneurs. And that will, again, now push the demand for reserves up and, therefore, also the money market rate. And which of these two effect will dominate, will, in the end, depend on the lending cost. So how costly is it for a bank to find new investment projects to find new entrepreneurs? So when we simulate that, you see on the left hand side the effects for low lending costs, and the right hand side for high lending costs. And in the case of a low lending cost, where this lending effect dominates, you see that loans are actually increasing quite a lot when we increase the CBC rate. And that will mean that banks will need more reserves to buy these additional needed, yeah, to buy more CBC in case of refinancing shock. And since the demand for reserves is increasing, also the money market rate is increasing, and also the deficit rate is increasing, because banks want to attract more deficits. If this funding effect is dominating, however, you see that loans are still increasing a little bit, because it's still cheaper to fund entrepreneurs. However, since this funding effect is dominating, overall the demand for reserves from banks will decrease, which will also decrease the money market rate and the deficit rate. However, the amount of excess reserves, so this is the deficits minus the loans on the bank's balance sheet, which banks held ahead of this refinancing shock, they will decrease, because loans in both cases increasing deficits are relatively constant, and here a little bit decreasing. So we can say that a higher CBC rate will tend to drain reserves, but the effect on the money market rate is ambiguous, and specifically it depends on the bank's landing cost. Okay, so now let me come to one, yeah, maybe like policy exercise. So in the paper we study both the caps on CBDC and also tiered remuneration, here we'll only focus on the caps. So on the simulation you will see next, we have two different caps, a low and a high one, and we assume that each entrepreneur can only buy from one seller, because otherwise it could just go to a lot of sellers never hitting this cap. Okay, and if the purchasing power of that cap, so if the cap times its interest rate is smaller than one, and one is what the entrepreneur would need to buy an additional investment, it has to liquidate some of its initial investment. An aggregate liquidation in the economy is what a single entrepreneur needs to liquidate times this refinancing shock, times all the entrepreneurs who will need CBDC to buy additional reinvestment goods. And here you see that the effects of increasing omega two and the CBDC rate in general are the same, but now when we compare a relatively high cap, which is the solid line in this graph and the relatively low cap, which is the dashed line, you see that in both cases, investment through the dotted line here and there will be lower when the cap is relatively low. So we can say that a tighter cap will decrease investment. And that is because entrepreneurs will have to liquidate some amount of the initial investment, and therefore the effective productivity of entrepreneurs will be lower. And banks taking this into account will give loans to fewer entrepreneurs because of that. And you also see that the dashed line for the money market rate here is also lower. And this is the case since now banks give out fewer loans, fewer reserves will be needed, and that pushes down the interbank rate. And you see that the effect is especially large when the CBDC rate is relatively low. Okay, and the amount of liquidity or liquidation can be interpreted as a measure of inefficiency on our model. And to see how that aggregate measure of inefficiency evolves, we can see that once the market share of CBDC is relatively high, a low cap will increase the amount of inefficiency. So we increase the amount of liquidation. And here what is more interesting maybe is that the amount of liquidation is decreasing in the CBDC rate. I mean, it's higher in the case of a low cap than in the case for a high cap, but it's decreasing in the CBDC rate. So in principle, the central bank could choose a higher CBDC rate to undo the detrimental effect of a cap if it wants to go so. Okay, so now let me conclude. So what we do on our paper is we build a model adding end of the day refinancing shocks to a general equilibrium model of banking and payments which includes CBDC. And we find that when we increase the CBDC market share, this will drain reserves. And when reserves are scarce, this will increase the interbank rate. However, the effect of CBDC remuneration on the interbank market is ambiguous. And as you saw, it depends on the bank's lending cost. And then in the paper, we study both the effects of caps and also tiered remuneration. And we find that a cap on CBDC will lower the loans and increases the inefficient liquidation in the economy. And in the paper, we show that a tiered remuneration system is basically equivalent to a reduction in the CBDC single rate. Thanks a lot. Thank you very much, Nora. And the discussion will be done by Anneke Kosser from the Bank for International Settlements. Yeah, thank you. Thank you very much. As the slides are loading, let me start with thanking the organizers for putting together this, I think, very relevant conference. And I really like the idea of bringing together like the researchers and the policymakers. So I think this is really a great event. I also would like to sort of thank the authors, but also the organizers, again, to giving me the opportunity to read this paper because I must say I learned a lot. I am not a person with a background in monetary policy in the bank market. I'm also not a theoretical modeling person. I'm more empirics, I have a background in understanding payment habits. So I learned a lot. And I think that's also to say that please do not expect very in-depth comments from me about the model. But what I would like to do today is really to maybe come up with some suggestions on how to maybe enrich the model, if possible, with some assumptions or ideas really reflecting what the current statuses of central banks in their thinking about the motivations of issuing a CBDC and the design choices. So it's really more about suggestions to align it with the policy discussions that we see at central banks today. So I'll start with a short summary. So the key question of the paper is really what is the impact of a CBDC uptake on inter-bank market rates? And then they do some scenario analysis or some further assessments of different remuneration policies and holding limits. In terms of modeling, what it's basically doing, it's combining elements of the money market model of Berenson and Monif in 2005 and the CBDC model of Chiu et al from 2023, I just saw in the slides, but in a paper I saw 22. But anyway, it's combining elements of these two models. And what I think is key, what are the key assumptions to this paper? It's really like the behavior of the banks. And the assumptions are that banks have no market power when issuing deposits. And then in the first investment round that they face, they will finance some entrepreneurs with deposits. And then banks are assumed to be subject to a reserve requirement really as a fraction of the deposit issued. And then banks acquire these reserves by either issuing deposits or borrowing reserves on the interbank market. Then at the end of the day, when the interbank market closes, they are faced with a refinancing shock where some of these entrepreneurs need more money, either in the form of deposits or in the form of CBDC, depending on the demand from the market and the end users. If this demand is in the form of CBDC, then banks are assumed to buy these CBDCs with reserves. And these reserves would then be subject to 100% reserve requirements. So not fractional reserve requirement, really 100%. That's, I think, the key difference with deposits. And then it's also assumed that banks only have access to central bank deposit and lending facilities and may actually borrow these reserves. So the findings, I think what they really show is that an increase in CBDC use such drains reserves may increase the interbank rates. Banks might also increase deposit rates to increase deposits. And then they also acknowledge that central banks can undo the effect by supplying more reserves. Then when looking at the effect of remuneration, the conclusion is, well, it depends because there's two different opposing effects. There's a funding effect and an investment effect. So the funding effect is basically that because of this interest remuneration in CBDC, in the end, less CBDC would be needed for the same outputs, so less reserves would be needed. At the same time, because this funding becomes sort of cheaper for banks, banks can fund more firms, which would then mean more reserves are needed. And that's the investment effect. So it's really about the final effect, really depends on the size of each of these two different effects. And then they look at the holding limits and they find that holding limits reduce the interbank and the commercial deposit rates because banks, they require fewer deposits to buy reserves. But again, there they say central banks can reverse the degrees in investment by higher CBDC rate. Finally, I also looked at, they look at tiered remuneration and what they find is really that this generates a similar effect as just one single rate, which would then be a little bit lower. So my comments, I think I have three comments. And like I said in the beginning, these comments are not really about the model itself or the details of the model, but more about, yeah, how some of the key assumptions fit within the current discussions, the current policy discussions among banks, within banks. So one of the first assumptions that I, if I interpreted right, the paper is saying in our model, one can interpret the CBDC as a type of commercial bank deposit carrying a hundred percent reserve requirement being earmarked for CBDC, these reserves are remunerated at the rate I. So I try to visualize that assumption based on an earlier paper produced by Raffael Auer and Burma 2021. And if I understand it correctly, this assumption is basically saying that end users, they can hold retail CBDC and bank deposits and both would then be a direct claim on the commercial bank. Within the only difference being that the commercial bank, the CBDC deposits would be 100% backed by reserves and the traditional deposits would be fractionally backed with well, only a portion of reserves. And then they would hold these reserves at the central bank. So I think this construction is also known as like being a synthetic CBDC or a rigid stable coin. So I'm not really sure if this is the right way to go or maybe I misinterpret the model. Could well be because I think a CBDC is in fact a direct claim on the central bank. So it would actually not show up on the balance sheet of the commercial bank. So it would actually not be affected by any bankruptcy of the commercial bank. It would not rely on the soundness of the commercial bank at all. So that's a key difference, I think. So then here my question comes in, would this model and the findings differ when really thinking about the CBDC as being a direct claim on the central bank and not showing up on the balance sheets of the commercial bank? And of course, I mean, there are all kinds of differences in these two different type of constructs. I mean, there are regulatory, supervisory, legal, insurance related differences, but are there also monetary policy related issues? I'm not sure. I'm not the expert here. But maybe there are in terms of crisis because again, if it's a direct claim on the central bank, it would not be subject to any claims in terms of bankruptcy. My second point of question is about the stickiness in deposits because there's a trade-off between remuneration and riskiness of the banks. So because concerns about this outflow of deposits, it not only depends on the features of the CBDC, but also on the riskiness of these deposits and the remuneration of these. So there might be differences across banks. And how does the model allow for that? And there might even be differences between uninsured deposits and insured deposits. So again, also there, could the model allow for that? And would the results be different in that case? And now finally, another assumption of the paper is that no sellers ever accept cash in the economy. And to me, I think that's a very strong assumption. And I was really wondering how would the results change when allowing for a competition between CBDC and cash? And to me, that's important because when what we learn from central banks is that CBDC is seen as a way also to enhance financial inclusion, especially in developing economies, but also in advanced economies, it's also seen as a way to enhance digital inclusion. So, and to digitalize P2P payments, which are currently paid for by in cash. So how would the results differ when accounting for that and including cash in the model? I mean, I can assume that might limit the decline in the demand for deposits or maybe increase the demand for deposits if it really draws more people into the official financial system. So, actually to further sort of, yeah, show that I think CBDC will have, of cash, CBDC will also have an impact on CBDC. I took this graph from the results of the annual CBDC survey carried out by the BIS. And if you look at the third driver, so this is showing the key drivers of CBDC work by banks. And financial inclusion is actually among the key drivers of the CBDC work, especially in emerging markets and developing economies. But also in advanced economy, it's an increasing motivation. So I think that makes it pretty likely that CBDC will not only impact the demand for deposits, but also the demand for cash. And then this final graph, I think also is to substantiate that claim, is that it's actually a study done by a couple of ECB colleagues. They looked at substitution of bank notes and overnight deposits in three hypothetical scenarios. And then on the left, you see three scenarios, and yeah, three scenarios, and then the digital take up by sector. A scenario one, the one on the left is a scenario where the uptake of CBDC is really limited. CBDC is only used for retail payments, but to a limited extent. Level B, the middle scenario is really assuming a very high uptake of CBDC, where CBDC is used for retail payments, but also for store of value, no caps, no remuneration. And scenario three is sort of a kept scenario where the uptake is assumed to be like, where every household is assumed to have like this maximum within this cap of 3,000, 3,000 euro of digital euro. And then that visitors can also use CBDC only for retail. And you can see that in every scenario, CBDC is really found also to substitute cash. So yeah, my question here is really how would the model change or the results change when accounting for this? Let me stop here. I think again, it's a great paper. I learned a lot and I look forward to hear yeah, all those views on these three points. Many thanks. Thank you very much. So I would open the floor for questions. One is there, one is there. You got money? Should I start? Okay. So my question concerns the issue that the central bank might just treat this demand for reserve from CBDC as an autonomous factor in liquidity management. So try to fully accommodate this. And if I remember correctly, the size or the share of the firm suffering your refinancing shock is fixed, so it's not not stochastic. So is there any aggregate uncertainty about the reserve demand emanating from this reinvestment shock? And if there is none, why wouldn't the central bank just try to preemptively accommodate this additional reserve demand? Thank you. So two quick questions. So going back to the discussant's point about cash and you don't have cash in the model. My understanding of the model is if you had cash, it would have similar impact on reserve requirements as CBDC. So one way to kind of get around that would be maybe think about cash and CBDC as a composite. And the second is for Omega two, that drives the market share of CBDC, but it's independent of the cap when you look at the CBDC cap analysis, right? So it would be nice if you could get to a model where if Omega two was, sorry, if the cap was zero so that you couldn't hold CBDC at all, you kind of get to the case we're in now, right? You don't have CBDC. Sorry, I didn't understand something of the model. So he explained a lot about the refinancing shock, but how about the goods market? I don't know where these entrepreneurs spend this money. I mean, part of this is coming back to the banking system, I guess. I didn't understand this part. How much of these reserves that are drained come back actually after the refinancing payments? Sorry, the reinvestment payments. Yes, one question on, there was lots of emphasis on the substitutability between the various forms of deposit, cash, CBDCs. And this refers, if you want, to the broader concept that the BIS is working on, on the singleness of money. And we were discussing with Nikkei yesterday about exactly this point, and I want to hear the panel opinion on this issue. I mean, we all assume that there's one-to-one exchange rate between all this form of money, but in reality this is not by construction or because we want this to be. Or there are market forces that makes this to be the case. But in a world in which the new form of cash, the digital cash will have potentially lots of other features such as programmability, how can we ensure that these implicit exchange rates remain one-to-one and don't deviate from that? Thank you. Yes, thanks for a very clear presentation. I just had one question about the interbank market. Just clarifying questions. So I understood that some banks will need more reserves and other banks have a surplus. So I thought I was like, that seems nice because we know that some banks have more reserves than other banks now. But then you said that there's no actual trading on the interbank market. So I just didn't, I don't know if I understood that correctly, but I just wonder why then in the end is there no transactions, but is there a price spin down and how does that work? Yeah, just a short technical question that also Anniket took up in her discussion. Namely, does it show up in the balance sheet of a bank or not? Since the digital euro is a claim on the central bank, it does not, actually I agree. But yesterday we learned that the digital euro amounts are counted as part of the reserve requirements. So in somehow banks must take this into account. So when they have reserve requirements and the deposits with the central bank are of course a claim in the balance sheet of the commercial banks, but at the same time, these wallet balances somehow taking into account. So I would like to hear from both of you how you think that works and whether that affects the findings. All questions. So if not, then you have time to answer this pretty broad range of questions. Okay, yeah. So first of all, Anniket, thanks a lot for this great discussion. This is super helpful. And also thanks to everyone for all these questions and comments. So maybe so let's start on the balance sheet question. So we don't assume that the CBC will show up in banks balance sheet, but CBC will be channeled through banks and banks will need reserves. So they bring the reserves to the central bank, take this CBC from the central bank and give it basically directly to their clients. So it won't stay on the banks balance sheet. So to the question on the interbank market, so why there's no trading and that some banks are short and some banks are long in reserves. This will be the case after this refinancing shock takes place and then the interbank market is closed. So then the banks can only turn to the central banks facility. That's why there's no, because ex-unter the banks don't know which how large this refinancing shock will be. So that's why there's no trading on the interbank market. And there was a question on the substitutability of deposits CBC and possibly cash. So exactly because we believe that there won't be perfect substitutes because maybe of different features of CBC maybe programmability or something like that. That's why in our model there also no substitutes. So that's why some sellers only accept CBC and some only accept deposits. So that's actually, yeah. So the point and the question on the model. So what happened? So if I understood Mariana's question correctly. So what happens to the reserves in the secondary financing round? So what happens is that so entrepreneurs get the CBC and they will pay to the sellers to get the additional reinvestment good. And sellers then can use the CBC in the market too. So the centralized market to consume and to spend and to give it back to buyers and buyers can use then use it to obtain deposits from the banks. So the buyers can revert it back to reserves and then get and reserves basically cash on our model and then can turn that to deposits. That's how that channel goes. So then your point on the, so whether Omega two should depend on the cap actually I mean it's a good idea because we could think about it. I mean in general this Omega two will depend on many features will also depend on the CBC rate. So I mean it would be super interesting to do it and organize it but this will basically be another paper I think because it opens up so many more questions. And then whether we have aggregate uncertainty. So I mean we don't. So this liquidity shocks are bank specific so we don't have any aggregate uncertainty and you are perfectly right that actually in our model it's like an autonomous liquidity factor and of course in reality it's very likely that the central bank will react because in now in our model the amount of reserves is basically fixed. So if more reserves are needed that will decrease and that will immediately increase the interbank rate but of course in general the central bank could and probably will react and increasing the supply of reserves. So I mean we had similar comment before we probably should work on that. Also what is like the optimal amount of supply maybe. Maybe if I can add on the one to one conversion I think the central bank always has means to ensure this one to one. We are able to ensure one to one conversion between reserves and cash and it's elastic supply at the same price. So I think that's something that would also work in the case of CBC. So this concludes the first paper and then we can move to the second one which will be presented by Gallo. So thank you very much. Thank you for the organizers for accepting this paper. If I'm not mistaken this presentation is the only thing that stands between all of you and a very nice weekend or maybe not that nice that will be this in critique. So I just hope that you're gonna stay with me for a while because this paper in Hawaii in a certain way it touches on several of the points that have been discussed by previous papers both today and yesterday. This is joint work with a couple of great guys from Ango España and the use of this term is applied. So again, I will not get into why CBDC matters and in this paper we are gonna concentrate on the role of CBDC into the operational framework. For those of you who were yesterday in the keynote speech by Panetta, there was this question by a gentleman from Politico about a certain paper that makes these claims about the operational framework, that is the paper. So I mean just for that I hope that you will have because the paper has been relatively, has received a lot of attention by the press because essentially we deal with something that is very topic out today, which is the operational framework. As probably most of you know, if not I will tell you a number of central banks are currently or have been engaged over the last 12 months in discussion about which is gonna be their operational framework. By operational framework essentially, we mean which size and composition you want to have of assets and liabilities of the central bank and how do you want to steer your, I mean how do you want to change your policy rates in order to steer the interbank typically the target rate in order to deliver on your price stability or whatever mandate that you have as a central banker. So the point is that CBDC may have a considerable impact on the, sorry, on the central bank. On the central bank. So in order to, essentially I mean I will present here. So we are gonna try to answer several questions. I mean some are broader which is like if you read the popular especially I mean three, four years ago, but still I mean in this year if you read the popular press you will always find an op-ed by someone in the financial times making the claim that CBDC, I mean the introduction of CBDC I mean and that you hear more often that not from representatives from commercial banking unions that CBDC is gonna lead to a deposit crunch because there will be this shift of deposits from the commercial banks to the central bank and hence that will lead to a credit crunch and that's gonna make credit more expensive and that's gonna reduce credit and that's gonna slow down economic growth. So then we should be very careful about CBDC. The first question that we want to make is that within the context of a general equilibrium standard New Keynesian model, so the benchmark model that people around used to analyze monetary policy transmission, is that true or not? The second question is, okay fine, but how does it depends on the operational framework of monetary policy? I guess that many of you know that since the crisis, not the last crisis, but the previous crisis, so since the great financial crisis, most central banks in the world have been operating what is typically called a floor system. So essentially it's a system characterized by huge central bank balance sheets, a lot of excess reserves and then the interbank rates typically back to the deposit facility rate in the case of the euro area or the interest on excess reserves in the case of the Fed. So which are gonna be the consequences or which would be the consequences of introducing that CBDC on the operational framework? And then finally, we want to touch on the issues related to the Marcus and Dirk paper about, is there a way to make the introduction of CBDC neutral and there's two tasks not having any real impact on aggregate variables. So this is the kind of question that we are gonna try to address. And in order to do that, what we are gonna do is to introduce kind of a heterogeneous edge and tractable heterogeneous bank model into a standard New Keynesian structure. So and then with an OTC search and matching the market in order to understand the interbank. The interbank, but to me, I mean particularly, I'm very biased because I'm one of the authors, but really I've written many papers in my life and some are better and some are worse. What I really like of this paper is the calibration. The calibration and the figures, I mean the aesthetics of the figures, but this is because of my co-author. But leaving aside the aesthetic of the figures that you will see in a while, the calibration, I mean, it's very nice. It's, I mean, you will see. So which are the main results? Because I mean, I'm sure that at some point I will start losing track of the presentation and then I will run out of time. So then this is important that I can convey the main results at the very beginning. So first there is the issue that is true that introducing, I mean, it was made by the previous paper. Introducing SEVDC will necessarily, I mean, one important caveat is that I'm an engineer. I'm not in the marketing business. I know nothing to marketing. I know nothing about customer experience. And then I have no idea whether SEVDC is gonna be a success or a failure. And neither do my co-authors. So then we don't make any assessment about how successful SEVDC is gonna be or why it is gonna be successful. Essentially, we are very agnostic and we consider a continuum of scenarios about SEVDC take up in order. And some of them, SEVDC is extremely successful. In other, SEVDC is maybe less successful or it's super successful but has been capped by this kind of holding limits by the central bank. So we don't take any stance on that. So then in any case, depending on how successful it is, what necessarily will happen is that the more successful it is, the larger the drain on cash and deposit that the introduction of SEVDC will produce. But then the issue is that for the, let's say, in the first phase, given the huge volume of excess reserve that exists nowadays and that are forecasted to exist in the foreseeable, at least close future, then this is just gonna produce a reduction in the excess reserve but we are gonna keep on living in this kind of a floor system. However, there is a level, I typically call it the NIE but people typically call it the KING at which the volume of excess reserve is large enough and hence the behavior, the dynamics in the interbank market change and the interbank rate will lift off from the deposit-facility rate and we will shift to a kind of a corridor system similar to the one that we were operating in the Euro area prior to 2007. But even in that case, the point is that even in a situation in which reserves start to be scarce, then what will happen is a substitution from funding from deposits to funding from the central bank. But, I mean you may say, yeah, but funding from the central bank is gonna be more expensive but this is a partial equilibrium approach because the real interest rate of the central bank will be given by the whole stance of monetary policy so then depending on the whole stance, the funding should not be necessarily more expensive because the monetary policy reaction of the central bank and that's part of that. So that is an important issue. So the answer to the first question is do the deposit crunch lead to a credit crunch? And the answer is no. Or yes, but to a very relatively small extent and the reason of why it is happening is unrelated to the dynamics of the central bank. It's more related about the kind of Brunemeyer-Nippel dynamics about the remuneration of the total wealth of the household. So it operates through an entirely different channel. But then of course there are reasons and there are reasons why you want to preserve the kind of floor system. And in that I side with what Panetta said yesterday. The central bank has tools to preserve a floor system either through outright purchases of bonds and credit operations. Notwithstanding and that's notwithstanding a deposit is not the same as a TLT row because a deposit is a non-collateralized instrument for the banks whereas TLT row are collateralized. So that opens the door and if I have time I will talk a little bit more about that on the particular designs of this kind of credit operations into the role of preserving the floor system. So this is more or less the roadmap of what time we want to say. Okay, great. Thank you. So the model can be summarized by this. I cannot walk, it's a pity, by this slide. Essentially what's the idea of the model? What you have is that households can save in CBDC, in cash, and in deposits. And then the banks can be funded through deposits. By deposits you can take both into account kind of retail deposits and wholesale deposits, even corporate deposits. I mean for the purpose of the model that is irrelevant. And actually for the calibration we are gonna lump them together. Banks can be funded through the lending facility of the, so the marginal lending facility from the central bank or by equity. And essentially the business of banks is to just give loans to or purchase securities from the firms and that is the way that firms, I mean this kind of Modigliani-Miller structure from the side of the firm holds in the model so it's irrelevant whether firms are funded through loans or through equity. But all this funding come from the commercial banks. Commercial banks can also invest into government bonds and they can also save into central bank reserves. And then the central banks on the liability sides is gonna have central bank reserves, cash, and CBDC. And similar to the previous discussion, CBDC is just a direct, the same as cash is just a direct claim on the CBDC. And CBDC for most of my presentation is gonna be unremunerated, so 0% return on CBDC. And then the central bank on the asset side can have credit operations with the commercial banks or purchase bonds from the government. I'm not gonna get into the whole of all the details but in general, all the details as in the standard New Keynesian model, except those related to the heterogeneity of the banks and the interbank market. So this is where I want to talk. And I want just to make a point about the demand for CBDC. How do we construct the demand for CBDC? One possibility is to go to this kind of a Kiyotaki right, Lagos, micro-founded demand for money in order to understand which are the intrinsic value that CBDC will give to households. Another which is this one is to say, I have no idea why people are gonna demand CBDC, honestly. I'm not gonna make any welfare assessment in this paper because I don't feel bold enough to say, so I'm gonna consider similar to other papers this kind of liquidity in the utility function. For some reasons, people would like to hold CBDC and this is gonna be controlled by this parameter eta DC. And essentially what I'm gonna do is a number of comparative statistics about both steady states and transitions or long run situations and transitional dynamics for different scenarios of the take up of CBDC controlled by that parameter, okay? Then how do the banks, how do the banks work? Essentially we built on previous paper with Thos Karate here from the ECB and Dominic Taylor also here from the ECB and Carlos and myself, which the idea is is a relatively tractable model of bank heterogeneity. So essentially what we assume is a continuum of islands. In each of the islands there is a band. This bank interacts with firms. I'm not gonna repeat the funding and the investment of the banks but it's pretty complex. But essentially the core idea is that this is one of these morning afternoon models in which in the morning you get some deposits but then in the afternoon maybe you realize that the productivity in your island is higher or lower and then you want to have more liquidity or less liquidity. And then in order to get that liquidity or less liquidity, you go to the interbank market. A contrary to the previous paper, the people in the interbank market here are more hard-working and the interbank market is open. So then you can go to the interbank market and borrow or lend. So that's the good role of the interbank market. That's the role of the interbank market here. And then banks are subject to this kind of leverage constraint that essentially limits the maximum size of the loans depending on your equity, okay? So the important thing is to understand what is the role of the lending facilities of the central bank. The role of the lending facilities of the central bank is that they just give you an outside value in the bargaining that you have in the interbank market. I mean, I don't know if any of you have ever worked in a commercial bank. I have. You have seen this kind of interbank, how the interbank market works. So essentially, I mean, now not anymore but in the good old times, you just took the phone and said, look, I have this liquidity and I want to place it. Which is the, I mean, you get quotes and then you place it. So essentially the point is that the quotes that you get depends on which is the what the people in the search and matching literature call is the tightness, is the relationship between the volume of lending orders and the volume of borrowing orders. Because that gives you the idea of which is your market power. I mean, your bargaining power in that would be more precise. I mean, when you call and you know that if I hung the telephone, I can call another guy and the other guy is gonna give me a better quote than if you don't give me a good quote, I call the other guy. However, if I've been extremely lucky that you took the phone for me and I know that if you don't give me the price, no one else is gonna give me, then I have to stick to your price. But then there is always an outside option and which is the outside option? The outside option is that if no one else wants to hold the telephone for me, I can always go to a central bank and deposit at the deposit facility or borrow at the borrower. So the way that monetary policy steers in any kind of corridor is by giving these outside options into the bargaining protocol between the borrowing and the lending banks. So in a way, with this kind of a stupid example, I have explained more or less the way of this kind of OTC, search and matching markets, Alfonso Alagos or Bianchi Abidio that are underlying to our structure. So the bottom line of that, that it works, yes. So the bottom line is at the end, the interbank rate is gonna be a linear combination between the two facilities of the central bank, but the position within the corridor is gonna depend on the tightness, is gonna depend on the volume of excess reserves if you want like that, it's gonna be linked to that, okay? Then there is a way of pass through from the interbank rate to the deposit, but I will skip it. Then the central bank just operates a fixed corridor and then we'll have this, and follow this kind of tail rule for monetary policy. And then the balance sheet of the central bank is the one that I described, so essentially it's the bone holdings. In the marginal lending facility, what do you have? You have the total volume of borrowing orders that didn't find a match. So then it's important to discuss which are the different kind of operational framework that you may face. A floor system, a corridor system, or God forbid that even a sailing system. The idea of this regime is that in a floor system what you have is abundant liquidity condition, so the tightness goes to zero. So essentially all borrowing banks are matched with lending ones, and the most lending bank deposits at the central bank. So essentially the deposit facility rate is the key policy rate because it controls the interbank rate. I've wasted, I mean, many hours, thank you, with journalists trying to explain them that the MRO now is a totally inconsequential for the transmission of monetary policy in the euro area, but they're still in Spain, they keep on reporting that the MRO is whatever and said, okay, fine. So the corridor system is what it used to be, and in that case, I mean, the corridor is key and the two facilities play a role because you are in the middle of the system. So this is just, I gave you a brief overview of the operational framework, and now the quantitative exercise. As I said, the calibration is great, and it is what I like the most. How do we do the calibration? We don't calibrate that for the today because CBDC is not gonna be launched tomorrow. So essentially, we need to understand which is gonna be the operational framework in the medium run, let's say 10 years ahead. But I don't know the future. If I knew the future, I wouldn't work in a central bank. I would work in a hedge fund. But as I don't know the future, I work in a central bank. And then the only thing that I can use is this amazing survey called the Survey of Monetary Analyst that is run by the ECB before any governing council and that make all kinds of crazy questions to the financial analysts. And among these is how large, which facilities, which interest rates, whatever, expect to have in the euro area in the next months, in the next quarter, and in the long run. So essentially, this is how we calibrate the model. We calibrate the model for the long run. That part one. And then the second part, I mean, then we draw from the literature about this elasticity of substitution. But then a key issue is that this is data and model. The black line is a model. And I mean, don't tell me that this figure is not super beautiful. I mean, it's about me. I really like it. So the dark blue is very far in the past. It's like 1999. Some of you were not even born in 1999. The red hot part is the present. So essentially, this is the link. I mean, this is the data. Between the interbank rate that in our case is not the stair because with the stair there is a problem of the leaky floor because non-banks can intervene. So this is a repo. So it's a collateral. But I mean, don't care about that. Trust my words. I mean, I know what I'm talking about. So this is the interbank rate minus the deposit facility rate. And then you see that relationship. This is an equilibrium relationship. This is not a demand curve. This is a hole. So this is the demand curve cut by the supply curve in different moments in time. And this is the equilibrium relationship produced by the model because we calibrate the model to replicate the data. Of course, and this is something that concerns very much our dear colleagues from the Fed, it can be that this historical relationship has changed with time because the Basel 3 regulations or higher liquidity demand by banks have pushed that to the right. If that is the case, you should see the results on this paper as a kind of an upper bound. So everything will happen before we expect that in the paper. OK? So then we calibrate, and this is what I really like, we really calibrate the whole aggregate balance sheet of the financial sector and of the central bank in the euro area. I mean, this is not that we put a couple of numbers. I mean, this is a very serious calibration exercise in order to replicate the balance sheets of banks and of the central bank. I wish I could spend more time. I will focus on this slide because I think I will run out of time in a couple of minutes. But this slide kind of conveys some of the, at least the first part of messages that I was highlighting. This, as I said, is just a combination. I really regret that I cannot move there. But anyway, so in the x-axis, what you see is the amount of CBDCs in circulation in the steady estate in the long run. I have no idea how much it's going to be. If we impose a 3,000 euros holding limit, it will be equivalent to around a 6% of the GDP. So maybe it will stay, every European citizen will want that. And it will be 6%. Maybe it will be a huge success, it will be 40%. Or maybe it will be a failure, and it will be 0%. And then what you see is, depending on that, so don't see that as an evolution in time. This is an evolution of time. This is a continuum of alternative scenarios. You see it like that. So what you see is that the larger the take-up of CBDC, the larger the reduction in both and cash and deposits. Both of them are going to be. Of course, if CBDC would be remunerated, I mean, it would impact even more. But this is non-remunerated CBDC then. As you go increasing the amount of CBDC, you go reducing the amount of excess reserves. I mean, this is almost autological. But then there is a point in which you jump from a floor into a corridor. I mean, even if people cannot hear me, but I will show it here. Those are the points in which essentially what is happening, this is a very strange figure to see. Because what you see as almost constant is the interbank rate. So what you see in this figure that seems to be moving is not the interbank rate. It's the deposit-facility rate. Because in a way, as I was saying, the interbank rate in a steady state is given. And this is a model in which the natural rate is endogenous. But in a way, it is endogenous to the monetary policy decisions in order to preserve price stability. So what you see that is decoupling for the interbank rate is the deposit-facility rate. It's kind of a twisted logic of a steady state. But this is where you move from the floor to the corridor. But the bottom line, I mean, the most important issue is given by panel F. So in panel F, you see that even if it's, there is a certain decrease in the volume of bank credit, bank equity, and output, quantitatively, it is tiny. It is very small. This is a quantitative exercise. And this is very small because essentially what is happening is this substitution from deposit lending to central bank lending. And the reason why it is falling, it's just because the remuneration of wealth of the deposit. So this is why it relates to Dirk and Markus' work, which is that if you remunerate CBDC, it is possible to find a way of remunerating CBDC such that from the point of view of the household, the remuneration of wealth is neutral. And then if you operate in a floor or in a saline, then essentially you find no effect. During a corridor, you will always have some losses of output and efficiency. But those are related mainly to the fact of seniorias. The fact is that during a corridor, I mean during a floor system or a saline system in a steady state, this amount of remuneration, the seniorias doesn't change for the central bank. But within an up floor, there are always, I mean, the orders that go to the marginal lending facility and the orders that go to the deposit facility and then you make an extra cut. I mean, that would be, you tell me, I'm not getting what you say, don't worry. I mean, this is like, I don't know if you have seen American Beauty, that at the end of the movie, there is this issue that you said, probably you don't know what I'm talking about, but don't worry, one day you will know. So this is the same. So if you want to know, read the paper. So let me conclude. I mean, we could talk about transitional dynamics, but the figure is less beautiful. But I will just make a point about the transitional dynamics, which is totally kind of counterfactual. I mean, it's very anti-treative, which is, if you introduce CBDC, which is something that is gonna compete mainly 0% remunerated CBDC, that is something that is competing, again, cast mainly what you will observe. According to this model, in the first decade, it's a huge increase in the demand for cash. How it's that possible? Due to the magic of the New Claimsian model, which is, as the introduction of CBDC is gonna be recessionary, even very likely recessionary, in the long run, it is gonna be deflationary in the short run. And given that it is deflationary, that means that it pushes prices down. But then monetary policy will react and it will also reduce nominal rates and real rates down. So then from the point of view of the household, you will find that you want to get more CBDC, but then the deposit, the remuneration of deposits, the real remuneration of deposits is going down due to the action of the central bank, whereas the real remuneration of cash is going up because inflation is going down and then the real return on cash is going up and then you want to hold more cash. Is that something realistic or not? I have no idea, but this is what the theory would predict. Thank you very much, bye. Thank you, Nuno, and now the discussion will be done by Lea. Yeah, thanks a lot for the opportunity to discuss this very great paper. And actually I have to admit, I was already quite a big fan of the underlying framework. So introducing CBDC into this framework, now I find this paper even more interesting. So maybe shortly in a nutshell, I think it is a very elegant, yet comprehensive New Keynesian model that includes a realistic central bank balance sheet and interbank market dynamics via search and matching frictions. And that allows us to combine the analysis on the implementation of monetary policy and analysis of its macro effects. I can only one framework, which makes it very flexible and kind of able to speak to a lot of ongoing debates in the literature while also contributing to a debate that so far has been neglected a little bit by the literature. But as we see, it's also by the other papers in the sessions or in the conference is now coming up more and more as well. So maybe very shortly, let us recap on the results. So we have non-reminerated CBDC as a baseline. We have a floor system in the future and in the baseline, CBDC-induced deposit crunch will not imply a credit crunch. However, if not managed, CBDC may affect the operational framework. But I think it's already a quota to kind of to Panetta. I mean, we can undo what we have done. So a pre-CBDC floor system can easily be maintained by increasing the reserves, either we are lending operation or bond purchases. We have a kind of, we have a contractory effect that's small, but it's kind of the, I think also here, just the tendency of the trend matches that maybe kind of recovered by renumerating CBDC and then we'll end or we can recover the result of Bruno Meier and Nipol to kind of have the equivalence of private and public money. I also try to spice up a bit the presentation with some pictures given that's the last one before lunch. So the story is really quickly, this is actually what Dolly gives me when I described to him a New Keynesian household, a typical New Keynesian household. So he loves consumption, he hates work and he loves money. We have money in utility via CES aggregator. And what's central here is that money is either banknotes that are non-renumerated, CBDC, that's also non-renumerated and deposits. And I think this will really matter, the remuneration of the different types of money because that will actually create this wealth effect. Then we have this kind of island-bank-firm story that was explained really well by Michalo. So we have kind of different types of banks, different types of banks, firms, combination on different islands that have this idiosyncratic productivity shock and are either less or more productive in each period. So this will determine their different types of borrowing and lending orders and the balance of these borrowing and lending orders determines kind of then where the interbank market rate is. So if it's balance, we're kind of in the middle of the corridor, if there are more lending orders than the interbank market rate is pushed through the floor system and also kind of the demand from the deposit facility becomes much more important. And of course the other way around, if you have like a lot of borrowing orders then the lending facility becomes more important and the interbank market rate is pushed to the ceiling. So that's kind of the quick recap of the story. I have two main comments and maybe let me also note that already two weeks ago this paper was discussed by serial money in the money markets conference of the ECB. So I will be complimentary to his comments and I'll really kind of focus on the macro effects. So let us look at the results that we have a non-remunerated CBDC that reduces output. I mean, it's not big, but I think the main channel is important here. How does this happen? So we have remunerated deposits and then household shifts from renumerated deposits to unremunerated CBDC. So overall the average return on liquidity will be reduced. This increases kind of overall the opportunity because of holding money and makes it less attractive. So of course I will hold less, or I'll hold less generally less liquidity. This leads to a decline in lending to firms and also to a decline in production. But what I was asking myself is where did the money go? So let's have a look at where did the money go? So we have this less lower remuneration and now households get kind of lower returns on their savings. However, this lower kind of lower return on savings profits kind of the counterparty who has to pay less interest. And actually this counterparty here is the central bank. So kind of this leads to kind of higher central bank profits that are redistributed to the government. So we have higher transfers to the government and the government balance sheet is very simple. We have constant bond holdings. So an increase in the profit of the profit from the central bank will lead to a decrease in taxes. So overall we end up back, ah here, we end up back kind of at the taxes level. So it's basically kind of just a redistribution within the household balance sheet. So why does it matter? And I think it matters because we shift funds from the friction, very friction loaded or the kind of the friction loaded banking sector into kind of a completely frictional sector. So we have a, because the government it's, we have constant step to GDP. It's quite passive, it's constant step to GDP at his non-distortionary taxes and no government spending. So what I was asking myself is how much is of this world effect is coming from kind of having frictions in the banking sector, but no frictions in the fiscal sector. And I think actually the paper of this important channel, this Expansionary Fiscal Channel is an important channel also in the paper by Badang Kumhoff and is also present in the paper by Bologna Dahl. I think a second related point to this is that we have kind of all the assets that households hold, the monetary assets. So we have kind of liquidity, kind of we have and that are motivated through this money in the utility function. But in real, so everything that we kind of, we have money, kind of we have all the assets to use as money as a medium of exchange but also to store value. And I think here the store value is a concept of also what's really creating this wealth effect. So my question is kind of, what would the presence of an additional non-monetary asset kind of how would this affect it? And would it maybe just simply shift from liquid assets to illiquid assets? And I think if I have time, I want to very shortly focus on my second point is so on CBC balance sheet adjustments. So we have four ways in which CBC issuance can kind of affect the central bank balance sheet. Either we have kind of the demand coming from banknotes and then it's really neutral. We just have a kind of a very, very simple or kind of easy liquidity swap. Just the households are holding a different type of central bank liability. If it's coming from deposits, we have kind of either a decrease in reserves and this will also just be a liability swap. It will not affect the size of the central bank balance sheet. And for some reason, kind of we want to keep the level of reserves to a certain level or we want to keep it constant. We can also have set CBC issuance either by an increase in bonds or an increase in lending and this would then imply kind of an increase of the central bank balance sheet. And I think this is kind of very important in all four possibilities are also in this framework and in this paper. And I was actually wondering about the macro effects of the different policies that are maintained to keep the floor system. I think this could be explained a bit more because I'm wondering the policies affect the economy via different channels. So do they have a different macro effects? So I think that would be interesting also to hear a bit more about this. And I think it's also very important because really the literature needs a more systematic analysis of those general equilibrium effects of CBD central bank balance sheet adjustments. And I think the framework has all the ingredients. So kind of it's less of a comment but more of a suggestion to do further research with this paper. And I think it would be doing kind of analyzing this question. Maybe it would be also interesting to include a few other elements. For example, kind of also, I mean a more fiscal, maybe a more sophisticated fiscal sector. Maybe also introducing this kind of government bonds as collateral requirements and maybe additional regulation. But I think then it would be a very important contribution also to the literature because I think this is also where still some work needs to be done. And with that, I think it's a very elegant, very flexible and a rich framework and that I think can be used to do much more additional research with. Thank you. Thank you for the discussion and we have time for a few questions. So there is Oscar, Dilk, Matteo. Yes, thanks for these very good last presentations of the conference. Kalo, I would like to like your simulation as much as you do. And I think for that, I would like to be convinced why it's okay to not have any liquidity regulation in the quantitative exercise. Because I think, I mean, I've done some simulations of Euro area data using bank level data and we find it actually is the LCR requirement that's the first that becomes binding for a majority of banks when they consider how to adjust to CBDC. So that seems also relevant in your framework. I have an online question and then I would hand over to Dilkette. So the question online is from Paolo Fegatelli and he says, your model assumes that for bank deposit funding and central bank funding these are virtually equivalent apart from the different rates. However, in the real world we would also need to consider the opportunity cost of holding appropriate collateral to access central bank liquidity and the different regulatory treatment or deposits compared to central bank borrowing for example the LCR and the NFSR ratios. And could you envisage a manner to include these implicit costs which might be substantial in your model when banks run out of reserves and need to switch from deposit funding to central bank funding? I think this matches a little bit with what Oscar had as a question. Thank you. I have a bit of a disagreement about the intuition of the non-neutrality and I think I'm closer to what Leah said here. So we know from what we discussed also in the morning that with the CES liquidity aggregate you will never be able to get neutrality and we understand why because you have a non-linear rate of substitution between those two things so if you change the composition of liquidity the spreads will change and everything will change, that's for sure. So that is the fundamental source of non-neutrality. You emphasised the return on the household savings. I don't think that is any source of non-equivalence here because as Leah emphasised I mean you have essentially regarding equivalence or whatever, right? The household absorbs all the wealth effects and there's not per se any friction in the household savings. They choose to adjust their portfolio because the spreads on these different assets change because of the CES liquidity aggregation. That is the only source of non-neutrality. The other source that you have in your simulations but I think you don't need to have them is this frictional interbank market. But if I understood correctly that is not truly frictional in the sense that there are some resources lost in the process of matching the two sides of the market. It's just a mapping from the interest rate that the central bank sets into the interest rate that the banks face because those weights change with the tightness of the interbank market but if the central bank were to adjust its rate appropriately the central bank could always make sure that the rates that the banks face are fully insulated relative to the situation before the introduction of CBDC. So you could perfectly insulate the banks from everything CBDC related if you wanted to and then it's only the CES aggregator that has any real implications but of course you fix that rate that the central bank sets if understood correctly and therefore you get this additional boost but that is by assumption and therefore I think also your real effects on output etc are probably mostly reflecting this talk assumption about what the central bank does and to some extent for sure also CES but I understood when I saw your paper last time that the CES part quantitatively doesn't matter much but conceptually it is the only friction I think that matters. Hi. So one of the main results of your paper is that the introduction of CBDC causes a mild recession and your argument is that the interest rates set by the central bank in general equilibrium will be reduced but then we started talking about CBDC in a period of zero lower bound and from what I understand in your model there's no zero lower bound so like the presence of this CBDC like how would it change the result if you had zero lower bound in your model considering that you wouldn't have this general equilibrium adjustment of the interest rate. Then I would give you the opportunity to respond. So first of all I mean I want to be very thankful for the discussion as Lea was saying the paper was presented a couple of weeks ago and then it's not always easy to make a discussion of when you have already a discussion of a paper but I think something that I like of a discussion is when you learn something about your own paper and that was the case and that connects with their comment so let me go in order in the particular order with all the points the point of the fiscal reaction is a very good one that I have never thought about so that is something that we need to think about whether results would change or not with different fiscal reactions and revisit the Bardera and Kunhof and the Burlone in that particular dimension so that is a point well taken. Then about the presence of non-monetary non-monetary assets and the more general issue of collateral requirements I don't know how to answer I mean in the sense that I would need to think more about that about what would be the implications or how can it be introduced I mean it's not a simple question to answer and I don't want to give a rush answer then the Oscar San Paolo questions about the liquidity regulations and the collateral requirements are totally true I mean those are important things but the way as I was saying the way that I see them essentially is that if those they will shift the demand for for reserves to the right essentially so there are several papers there is a paper by and there is another one by a bunch of guys from the New York Fed including John Williams Domenico Giannone I think is Garrafonso but I'm leaving some co-authors which they just discuss these kind of things both theoretically and empirical which are the determinants of the demand for reserves and then I mean I totally agree but then in that case that would mean a shift of the demand for reserves to the right and that would make all the quantification would be an upper bound because essentially all this mechanism would kick out earlier that says for a smaller amount of take up you would already have that so so that is true then let me jump one second one second Dirich's question and I move to the last question about the issue of the zero-hour bound in the first version when we started the paper we started analyzing the case of a zero I mean moving from a zero-hour bound and whether you would be converging to a zero-hour bound so they were quite present in our mind but given that in the current economic environment what the markets at least expect of the likelihood of a zero-hour bound is much diminished we have shifted this kind of normal situation but I mean the presence of the zero-hour bound would matter to a certain extent but the issue that this is not a paper about how I mean such as Andy Levins or Michael Bordeaux statements that CBDC allows to defeat the zero-hour bound or Ken Rogoff because you can charge negative rates because I mean having cash you will never you will never have it but we have also we never publish that paper we have used this framework to analyze this kind of issues about how banks suffer with a zero-hour bound but it's true that we decided not to explore that in that paper because I mean we saw that less policy relevant and then coming to to dear points I need to think about that because because I'm not entirely I'm not entirely sure about I knew Jorge was telling me about that and we were discussing it like last week this week I mean like three days ago for one hour about that point but I would need to because I'm not entirely I'm not entirely convinced about the issue of why because the point is that essentially you can for a floor system you can prove that if you remunerate that if you remunerate as you would do in the counterfactual scenario without that all the questions are the same so then I don't see why this issue that CES doesn't allow you to I mean that always breaks that I'm not entirely sure but the fact that I'm not entirely sure doesn't mean anything I mean it can be right, it can be wrong so but if you don't mind I would like to ask you a little bit about that later so thank you very much thank you very much for the discussion and for all the questions thank you very much as well so before we end I have still a few words to say so first I would like to make you aware of a special issue of the Journal of Economics which Michael Frankel and I will act as guest editors so submissions are still open so it will be a special volume on CBDC and if anybody of you has a paper or knows somebody who has a paper you are invited to submit submissions are open until 8th of January and it will be peer reviewed and we aim for a quick turnover with an online publication by the middle of the year the other thing is such a conference cannot happen without a lot of help from many people so first I would like to thank the presenters the discussions but also you as the audience who came here to Frankfurt and I think it makes always a very nice and lively discussion that people are on site so thanks very much for that then I would like to thank our media team so Stefan Seitz Anja Zinsch and Isabel Schmidtknecht from Digi Communications as well as the TVN team for perfect technical support so one tends to underestimate the work and the efforts that are needed that everything is broadcasted in a nice way so the recordings will be on the ECB's website so if you want to get back to a discussion or to the questions you will be able to see that in a few days and this is really very much appreciated then of course I have to thank Elisabeth and Nina Willenberg who did the administration so Elisabeth mostly on marketing and also paper submission I think everybody of you has got an email from Nina at some point in time so this is on top of her usual work and this is really something that is very much appreciated and then of course Dirk who had the idea for this conference and to address the ECB to set this off and also was instrumental in bringing people together and then of course my ECB colleagues Tony Arnaud Massimo Ferrari Arnault May and also Peter Hoffmann who we as a team across business areas which is always kind of an effort in the ECB I think we had a great collaboration and it was really very enjoyable and we are all happy that this conference was quite successful so before we part I would like to wish you a good trip back we will have a buffet lunch outside so please go there enjoy some lunch and I hope to see all of you again at some point hopefully quite soon thank you very much