 On behalf of the organizing committee, I want to thank everyone for joining us today. In this series, each session is hosted by a different institutional partner. Today's host is the ECB, and I will now turn things over to our moderator today, Katrin Asimaka. Thank you, Jonathan. So I'm Katrin Asimaka. I'm head of the the Monetary Policy Strategy Division at the ECB, and I'm very pleased to chair today's seminar with Martin. So let me first start with a few housekeeping notes. So the presenter has been allotted 25 minutes, and then there will be a discussion of 10 minutes, and then there will be a Q&A with the floor and the panelists of about 20 minutes. During the Q&A session, the panelists can unmute themselves and ask questions or make comments. The other attendees, please use the Q&A area to ask questions. I will then select questions from the Q&A box and ask either the presenter or the discussant to answer them. Please also note that this conference is being recorded, and the video will be posted on our website. Let me now introduce Martin. Martin is professor at Stanford University, and he has published many papers in top economic journals, and I have looked at his publications and I found that he has worked on every hot topic in monetary economics. So starting from housing, going to interest rate and credit risks, the effect of collateral, heterogeneous beliefs, distributional effects of monetary policy. So it's probably no surprise that he also has a paper on central bank digital currency, and today he will present the paper on credit lines, bank deposits, or central bank digital currency, competition, efficiency, and modern payment systems. So Martin, the floor is yours. Please share your presentation, and we are very much looking forward to your talk. Well, thanks very much, Katrin, for the kind words, and thanks a lot to the organizers. First, for putting together this very nice series, and then for having us. This is joint work with Monica Piazzese, who's also here, so you could ask questions in the chat that she can reply to while I'm speaking. So this is a paper on motivated by the discussion of central bank digital currency, where there is a rapidly growing literature with different concepts of CBDC. In this talk, I'm going to focus on the concept of interest-bearing reserves account for everyone that would compete with bank deposits. And our angle on this is that we want to say this is something that comes into the market for liquidity, where the other products are not only bank deposits, but also credit lines supplied by commercial banks. And just to give a rough idea of importance, this is deposits and credit card limits at US commercial banks over the last decades. And the idea is credit cards are an important means of payment. This is mostly household credit cards. And so this does not include all of the credit arrangements that banks have for liquidity purposes with, say, asset managers and firms, which would make this number even more important. So the argument is going to be that commercial banks add value by providing liquidity. And to do that, they exploit a complementarity between bank deposits and credit lines. Now, if you add a CBDC, the way that your proposals work, these will not be complementary to credit lines. And therefore, they're going to be beneficial only if they're much cheaper to produce than deposits. And this argument is going to rely on an externality among different liquidity providers. So it's going to help also think about stable coins and money market funds, which are other intermediaries that only do deposits and not credit lines. And so do not exhibit this complementarity. Now, we add to exciting and growing literature of theoretical studies of CBDC that have asked under what are the conditions when the CBDC doesn't really change anything or it changes various things. The new element in our analysis is studying the interaction with credit lines. And therefore, we also relate to a large corporate finance literature on the role of credit lines that in particular use this idea that when liquidity needs are hard to predict, then a credit line can be a very useful instrument. This is what's going to be our model as well. And this complementarity of deposits and loans, loans for liquidity purpose at the individual bank level, has been pointed out in the literature and studied empirically. And this is going to be central in our analysis, which of course is going to be about the general equilibrium and the interaction of these different types of intermediaries. Okay, so let me go straight into the model. So this is going to be basically a new classic growth model with some liquidity shocks and frictions that people then need to navigate. So we're going to have a continuum of households who work and consume goods. And every period, there's going to be a fraction VC of households that consume, whereas everybody works. And then when I have a preference shock, that's going to select the guys who work. Preferences were represented using a discount factor beta and then utility from consumption goods and labor, which is standard belt growth type. And aggregate consumption is going to be the sum over all these people. We're then going to have a continuum of competitive firms. And those are going to come in two flavors. There's consumption good producers that use capital on labor to make the consumption goods. And there's capital goods producers that use the consumption goods one for one to make the capital. And here, we're also going to have these liquidity shocks. We're going to say that there's some fraction of the I selected by an ID productivity shock to try. And that tells the person whether they can produce or not. And the reason for having that is to have this additional need for liquidity in the accumulation of capital. And we think of this as broadly as standing for asset management, portfolio rebalancing, things like that, where also liquidity is needed. And now the, so it's all neoclassical growth model except for these shocks. And then we're going to have these parameters V that are going to be central, going to describe the predictability of the liquidity needs in the economy. If we was equal to one, then everybody always knows that they're going to consume or produce and that makes life easier than it is not known. Okay, so why are these liquidity shocks a problem? Well, we're going to have liquidity constraints. And the way we're going to build those is by saying that the time is going to move in half steps. So t minus half t t plus half. And in the non integer periods, the households and firms trade goods consumption investment is purchased and also production occurs. But then the key is that only the banks trade assets, whereas in the integer periods, households and firms trade assets and the factors of production are paid. So you see that the role of these non integer periods is to have some like a device to mimic decentralized exchange as so did this sort of multi sub periods structure is also in other model of money, particularly Lagos invites a model. And here it's done with these different types of periods in a growth setting. And then the in order to transfer funds both in and out of these non integer periods, the households and firms going to have to use payment instruments. And so that's different for buyers and sellers. So the buyers of goods here, these households and capital producers, they're going to need the payment instruments before buying. And then because the liquidity needs are unpredictable, only some share V gets a chance to buy it. They got to take this into account when selecting the payment instrument. And then at the sellers that are the producers of goods, here you need the payment instruments after selling and their liquidity needs are more predictable. They just need to store the funds until the next half period in order to pay the wages and the rents. Okay, so where do these payment instruments come from? This is going to come from banks and banks also going to need payment instruments to meet customer outflows. So there's going to be liquidity constraints in both layers, both at the end user layer firms and households and at the bank lab. So let me say more about the payment instruments. We're going to have competitive banks that offer two types. One is deposits and deposits have the feature that you've got to hold them before trading. Then you spend them if you need them. And otherwise you keep them and hold them again for another half period. In contrast, a credit line is something that you draw down in order to receive a loan if you need it and otherwise you don't need it. You don't use it. We're going to quote prices per unit of liquidity provided. That's the way the pricing is going to work. And then you can think about the trade off here in a world where these liquidity needs are hard to predict is that if you arrange a credit line then that avoids holding deposits that you may not need. That's going to be the key advantage of credit lines. And then there is a cost. So then we're going to see how that works. Now we've got to take a stand on how the banks are organized. So we're going to assume that equity issuance is costless for the banks as well as for the firms. And the financial frictions in the banks and firms are going to be one for the banks there is a collateral constraint. That is the debt the deposits have to be less than some value fee times the value of the assets. And this is going to capture the idea that payment instruments have to be safe. And then second, we're going to say that there is an asset management services Kappa needed per unit of assets. And these are the goods that must be purchased at some price P. This is in order to capture the idea that any delegated asset management be it in banks or in firms is costly. And these are resource costs. These are not just wages but these are resource costs. So this production of services requires capital and labor. And then you see another piece of intuition right here which is that these firms because they have these asset management costs are going to like credit lines where they don't have to hold the deposits for liquidity purposes that they might not need. This is going to have these asset management costs. Okay. Now before defining equilibrium I have to say what are the assets that are available to the households in the integer periods where they can trade assets. And so in addition to the payment instruments households can directly hold capital and they can also hold bank equity. And importantly there are contingent claims on all the preferences and productivity shocks that the households receive. And so these contingent claims are one period ahead that the shock gets realized in the half period in the middle. Okay. So what that means is that we can then get a household sector that works like a large family that ensures the members against these idiosyncratic shocks and that owns the banks and the firms. And then we don't have to worry about the wealth distribution even though there's all these different people that get different shocks. And so that's sort of using complete markets to complete markets against with a period delay to cancel the wealth distribution as a state variable similar to what in the Lagos Wright setup is done by this Valrasian market where there is a linearity. Okay. So then we are going to look at symmetric competitive equilibrium that consists of prices and allocations and asset positions with many identical firms and banks maximum utility shareholder values and markets tier. Just comment on this. So this is putting in place a very stark liquidity-centric view of banking, not because we think that's necessarily the right view, but it's sort of zeroes in on what's new in this paper. And so in particular, we put some independence of the savings and liquidity provision features of the economy with two assumptions. One is that in this what I've described, we're basically saying that banks are in terms of their investing only good at credit lines they're actually bad at holding capital because you pay these asset management costs when you put capital in the bank but not when you hold it outside the bank. Second, and this is an additional assumption that's in the papers, we restrict preference in technology so that a capital is held by banks that if held by banks to back payment instruments is less in equilibrium than the total capital. This means that you don't have to have in order to provide the liquidity and back the deposits with these collateral constraints, you don't have to accumulate capital in order to provide enough liquidity instruments to make the economy work. So that means that then the size of the banking sector, which is an indulgence object in this model, is going to just reflect the demand for liquidity and not say these are kind of authorable things. Okay, second point, I've talked about capital as something that households and banks invest in, but this we can interpret broadly because in this model, Modigliani-Miller theorem and Riccardo's approval is going to hold except for any trades that involve payment instruments and so you could just think of the banks holding firm and government debt and not just capital. Okay, so now we have the model on the table so I can characterize equilibrium and this is relatively easy in this model because you can basically write down a social planner problem that gets you the allocation and that social planner problem has a special resource constraint that has liquidity costs omega on the different ingredients there, consumption, investment, and output. Of course, because we can divide by one, then principles is only two things that matter, but for the interpretation and for deriving it in the paper, it's helpful to split it into these three things and so then you can kind of see how the real effects of the payment system work in this model. If you have a more costly payment system that is equivalent to having a less efficient production technology and then how does the allocation of the economy react to that? Well, just like it would in the new classical growth model and that may include effects that differ by sector. For example, if you set up the system so that it's particularly expensive to do investments, say because credit lines are particularly expensive and then these firms that don't want to use deposits find it more expensive than the payment system discourages investment and there'll be less capital accumulation, etc. Okay, so that's, now you can imagine, so the paper has lots of equations where we can derive these omegas for different types of systems and do comparisons and this is what I'm going to do now in the remaining time, I'm going to show you some comparisons. This will all be steady-state welfare effects for different payment systems. I'm going to make the assumption that the liquidity needs are the same. This predictability of the liquidity needs is B, the probability that the liquidity shock shows up that is the same for the households and the firms, the capital producing firms here. It's to simplify and then I'll not walk through all the busy work algebra but I will just show you some balance sheets for foreign after trade that's going to give you the intuition for how the different systems work and then you're going to see how that's going to show up in these liquidity costs and therefore in wealth. All right, so step one, what if banks offer only deposits? So let's just assume there's no credit line. Let me shut that off. Well, then how does this economy have to do the trade? Well, the buyers, these households and these capital goods producers are going to have to hold some deposits because they want to purchase stuff in these intermediate periods where they can trade assets and the banks are going to supply that. So there's going to be a balance sheet of the bank here where the bank is providing deposits. There is a collateral constraint so it also needs some equity and it holds capital in order to back the deposits. So that's how things get set up before trade and then after trade. So trade means now V of the buyers get selected in order to buy and so V times D money goes to the sellers and then 1 minus V times D remains with the buyers. Those are the buyers who did not get lucky and able to consume. And so the bank nothing happens here because the bank just has the banking system all together has this D on the books. So all of the trades that happen there on the books of the bank. Okay, and so then how does this translate into costs? Well, then what we can do is we can relate back the deposits to the consumption and investment desires that motivate them. So in particular if these buyers what they need to accomplish is to provide C plus I consumption plus investment and they need to then hold C plus I over V in deposits in order to make that happen. And then sort of you can see I think that then now one can do the busy work and the algebra and come up with this resource constraints with these omegas which are the liquidity costs in the equivalent planner problem. And so that gives us the properties of banking with deposits which is that the liquidity costs are going to be relatively high if the needs for liquidity are unpredictable if the V is relatively small. This is because you have to hold though these sort of precautionary deposit holdings which are costly. This is especially the case for the capital goods producers who are going to not natural savers. And I'm going to just comment here that this is when I say bank then it's always the banking system. So kind of interbank flows wash out. This is something that we've written about in different paper but here this is the sort of interbank liquidity management is all netted out is all identical banks because of the symmetry. Okay good. So that with that we can go to credit lines and how credit lines might improve life. So what happens with the credit line? Well there's still buyers households and a couple of good producers and they know that they're going to have to trade and they're going to arrange for that and now they're going to arrange using credit lines. That means then before trades the balance sheets are all empty because the credit lines are contingent liabilities that are off the balance sheets. And then when trade occurs there's only all these buyers who have arranged for credit line with the banks that are now going to draw down loans and so that shows up as a liability on the part of the buyer. It shows up as an asset on the part of the seller and the bank is now going to have both on the asset and the liability side this v times l because every credit line that gets drawn that becomes a loan on the on the asset side and it becomes a deposit on the liability side. And then the bank has to have a bunch of capital because there is a collateral portion. And so that as you see already because there's all these empty balance sheets here the this cost of of holding assets is lower in this environment because you just get the liquidity when you need it and then the balance sheet costs don't come otherwise. And so then again we can derive by linking back to the to the macro quantities we derive these costs and this is the resource constraint. This is a resource constraint for the credit line economy and that basically has now welfare gains because all of these costs on the left hand side are smaller and it's basically that with credit lines you avoid these precautionary holdings which works as if there's higher TFP. And in particular because there is these these investment good producers are bad at holding assets this is particularly good for them so there is an additional kind of investment specific technical progress. Okay so that's the that's saying credit line's good. Okay now what happens when there is the entry of a deposit only intermediary and so generally we can imagine there's a new intermediary. It's got a different set of cost parameters there's a maximum leverage say fee star and some asset management costs copper star. And our leading example and this is the way that I'm going to talk about it now is this is a CBDC with a central bank that offers deposits at marginal costs given by k star over fee star. But you can also think about it as the entry of Libra which or there's competition between different Libras. All right so two sort of preliminary observations. One I mean this is going to help only if the new technology is better. That means that if this this cost at which the the deposits can now be produced is smaller than what we already have. That could come because of cheaper asset management or because there's better ability to commit say because it's the government. Okay now second point if banks offer only deposits and the new technology is better then it's always great. Okay because then what we're going to have is that all the depositors are going to migrate to the central bank commercial banks disappear. Here there's no loss from that because they provide no value beyond the liquidity provision now the CBDC can do the liquidity provision better so so great. Notice also that the investment actually will go up because now liquidity is cheaper the the resource core is smaller. Okay and now the the the subtle question is what if we have banks that not only issue deposits but also provide credit lines and there is the entry of the CBDC. Okay so this is now I can not only do the picture because there's some subtle pricing issues that Piro talked about. So now when there are these two payment instruments to choose from there we got to talk about the buyers themselves choice of payment instruments. So the first observation is if the bank deposits and the CBDC are priced the same then the bank customer is going to be indifferent. I'm wanting you have three minutes left just excellent yeah thank you and then so so generally there is kind of a potential price there's a price effect and then and there could also be a quantity effect in the sense that households might choose to hold deposits now that the deposits are cheaper relative to the credit lines. What I'm going to do in the in this talk is I'm just going to focus on the case where after the central bank has entered the buyers still use credit lines okay all of them and that there's a condition in the paper that guarantees that is basically saying that this that the predictability is relatively small. Okay and in the paper we do sort of the word the other case as well where there is a quantity adjustment as well. Okay so now imagine that there is these two products and what is the response of the commercial bank well it's going to choose to still issue the deposits it's better than equity and then it's going to match the higher interest rates that people can earn at the CBDC. In order to remain profitable what it's going to do is increase the price of the credit lines to break even okay because it now has this higher funding cost it's no longer profitable for the bank to invest in capital so it's not going to do that and so then we can look at what that means for the asset positions and so it's actually helpful to start down here where what happens after the trade so the because the buyers and sellers there's still there's nothing on the books here because they just come in with credit lines right they continue using the credit lines what happens after trade is they draw the credit lines and the sellers get as an asset the loans and then the bank is going to have the the credit line as an asset on the books but now the problem is that the banks because it can't it doesn't issue deposits against capital anymore it can't keep all of the deposits on its books here some of them are going to migrate over here to the central bank okay and that is an outflow from the banking system through the trade to the central bank and the banks have to prepare for that and they do that by holding up front here some additional liquidity at the central bank okay and so that then there's kind of two effects that that happened one is that there's this pricing of the credit line that is now different there's no more expensive because because of the competition from the deposits and second there's this liquidity holdings that the bank need in order to manage everything and then when you compare the costs and I'm not giving the full formula but just just the key inequality then you have the result that the CBDC is going to improve welfare in this case if and only if this equation holds which means that the CBDC is sufficiently cheap that the cost parameter is kappa star over fee star is sufficiently low relative to the cost of the private sector and then that that sort of translates into higher TFP otherwise not okay so it can happen here that because of this this competition um very welfare goes down can i ask you to wrap up so we are yeah i'm almost right um so uh so kind of there's competition for deposits that distorts the price of the credit line and this works not only with CBDC but you could just have free entry of this deposit only intermediaries the the mechanism is this externality among these different types of providers who jointly support the transactions that people do okay and so in any hybrid payment system that like in a deposit only system you're going to have these extra costs and so just to summarize the this is CBDC even if the technology is better for producing deposits is only beneficial if it's actually much cheaper to produce deposits than for the private sector okay thank you thank you very much fine um then we can move on to the discussion and i would like to ask Dirk to to share his screen and the discussants will be Dirk Neepert so he is a director at the study center and professor at the University of Bern and yeah so the presentation is up Dirk you have 10 minutes and i also will remind you of three minutes before the time is out thank you very much Katrin thanks to the organizers for this great series and for having me discuss this very interesting paper by Martin and Monica so the argument i think in in essence is that credit lines help economize on balance sheet length cbc doesn't and when balance sheet length is socially costly so shows that the emphasis here is on socially costly then CBDC may be detrimental and the interesting part that Martin emphasized at the end is really that nevertheless you might have an equilibrium in which the firms or some of them coordinate into holding that stuff CBDC although in spite of these externalities that it creates the outcome is is worse for society at large so i want to briefly focus on two aspects the first one is that this is and that's what i find interesting it's sort of a different result right i mean there's a difference between banking sectors with credit lines and without credit lines and the argument is that this difference is affected when we introduce this CBDC so it's like different different i want to look at that for a second and the the second thing that i find very interesting is this last point that that you can coordinate into something which is yeah worse than the previous equilibrium just because you have introduced CBDC so for the first part for the CBDC argument let me let me first state what what i think at least according to my thinking about the issue is is sort of a simple way to think about neutrality so an equivalence result um let's forget about credit lines for a second and just think about CBDC versus no CBDC so the first set of balance sheets here shows your world in which we do not have CBDC there are some firms that have some deposits on their asset side in the balance sheet and there are some banks that issue deposits um the question is what does um what does it change how does the equilibrium change in that setting when you introduce CBDC and one way to think about this is that it does not change um because essentially the firms could rather than holding deposits they could hold this new means of payment CBDC which would now be a liability of the central bank what about the banks well they have lost the deposits as a source of funding but instead they either have reduced their reserve holdings at the central bank as a compensating measure or alternatively they have simply gotten new financing from the central bank so the central bank here is a pass-through entity in the sense that we at the private sector we hold claims vis-à-vis the central bank the central bank lends to the banks and then in the end the banks you know don't really for them the world doesn't really change of course under some conditions so that's that's sort of where I come from that I think well CBDC doesn't really change anything of course there's conditions attached to that and there's there's many reasons to believe that this equivalence does not apply here's a list of some potential sources of equivalence failure for example CBDC and deposits might not be substitutable as a means of payment which means that you know for the firms this is not as good a means of payment as this one or the other way around then of course you would not get equivalence or there's different resource costs attached to using either deposits or CBDC as a means of payment against you would get non-equivalence or this refinancing of commercial banks by the central banks doesn't fly because there's collateral constraint there's informational issues etc etc or and that really gets to the heart of the story today I think is that gross positions gross balance sheet positions matter for some reason and this could be the balance sheet length because there are some asset management costs as Martin and Monika assume or there's some ownership structure issues you know some assets used to be owned by that guy now they're owned by somebody else that could also be a source of some mechanism of non-neutrality now how do credit lines come into play here I think Martin was very clear to show that and now I do the other diff now I go from the diff regarding CBDC versus no CBDC I know go to the diff regarding credit lines versus no credit lines so what is the effect here on the one hand and again Martin was very clear about that credit lines basically increase velocity in the economy right because on the one hand the firms need to hold fewer precautionary balances so these balances these deposit balances are turned over more often than before because you just use them when you really need them on the other hand the holding period for that stuff is shorter you only get them when you really want to transact and not already a month before that as you would do with deposit accounts you could also argue in that society issue that the ownership structure again has some consequences here so that's clearly a change of credit lines versus no credit lines there is an argument in the paper that credit lines also generate some complementarities and there I'm not that clear exactly what what the idea is in the paper a Martin Monika write that when a credit line is used the transaction generates an asset that in turn backs deposits this complementarity economizes on asset holdings and cuts cost by a factor of two now according to my reading in that model there's always of course one unit of assets backing a unit of liabilities and if you compare the cost ratio in credit lines versus no credit lines the cost ratio is totally orthogonal to the leverage ratio in that economy so that seems not to be the key mechanism to get different costs and also the two that is mentioned here is really just a reflection of the different holding periods with credit lines you have a shorter holding period than without credit lines so I'm not that sure about this complementarity issue but the main point that I want to make here under that heading is really that credit lines do not change the equivalence result per se that I like that I lined out before it's true that credit lines reduce the product so to speak of banks cross positions times to holding period that's clearly the case because of this higher velocity but at each point in time you could argue that whatever deposits are outstanding either in the world without the in the world without credit line or in the world with credit line you could still substitute whatever deposits are there by cbdc plus central bank refinancing as I highlighted in the beginning of course subject to the caveats mentioned but what I'm trying to say here is that these are two orthogonal things right you can think about credit lines on the one hand and you can think about cbdc equivalence on the other and the two are not necessarily tied into each other you can think about cross positions under giving a role to credit lines as as we see clearly there is you could also think about cross positions clearly undermining the cbdc equivalence result for reasons I highlighted before but you can think about them differently they're not necessarily coming to better in the paper they are because essentially I think the assumption that whatever balance sheet you have it has the same cost across all the agents in the economy is that nexus between these two assumptions but that's that's not really necessarily built into the story per se yes okay given that all these beautiful balance sheets that I that I spent hours creating a late take I will not show it to you and let me get to to my final two slides so so I think there's two ways to read the paper the first one would be to say this is a paper about the non-equivalence of cbdc then I'm not that sure because the main argument really concerns the relevance of cross positions for credit lines versus no credit lines as I tried to argue before and such relevance need not per se implied at cross positions also undermine the cbdc equivalence in the paper they do as I said but in principle I see them as two orthogonal arguments so in some sense if you want to argue that cbdc is non-neutral because of cross positions you can write down a much simpler model right without any of these credit lines you don't need them whatsoever you just directly go to the equivalence result I showed you before you say well the balance sheets are longer and that's it right so that's that's much more direct so I think the more interesting way to read the paper would be to say this is really about credit lines and credit lines have this interesting beneficial effect because they lead to this cross subsidization that martin emphasized towards the end and then some cbdc comes in and somehow things you know explode cbdc compromises them that is sort of I think a much more exciting way to read the paper it's sort of the opposite message of what we are used to hearing namely that the banking sector of course creates value but it might also create negative externalities that society has to bear in the end like charian felon's paper for example here it would just be the opposite somehow right I mean the private sector creates these the banking sector creates these positive externalities and now the government comes in the cbdc and screws up basically that's sort of the the message that they pursue that they they propose here and then of course the empirical question is is this really relevant how big are these social costs and these cost differences in particular on balance sheets and the theoretical question is and I think martin was open about that towards the end as well you know is it really cbdc that screws up here or what is it really because anything anybody who comes into into the game here without this cross subsidization business but just with offering a means of payment deposits would presumably undermine this beneficial equilibrium with cross subsidization so in the first place before cbdc do we actually have this equilibrium this beneficial equilibrium that cbdc can kill or doesn't that even exist thanks very much thank you very much for the discussion so I will next then ask martin to um respond to to the um the um issues that duke has raised and as munika is on the panel as well she may also want to come in thereafter we will commence with our q&a session and therefore I would like to invite you to post your questions in the q&a areas and I also would like to invite panelists to ask questions so um yeah maybe you can unshare your screen and then we can start with martin and probably munika responding to the discussion and then I will collect questions right so so great duke thanks a lot that that was that was very very useful um yeah and so let me first i said we agree about the readings so the way you read it that's also what how it was intended to be written we think is not primarily about there there's no equivalence result in fact i mean we the characterization of the equilibria works with a planar type problem and one can and it's very instructive to think about different systems that lead to to similar allocation um but but the sort of main thing that that we learned from doing this is that uh it is for for judging whether the all of these innovations not only cbdc but also libra etc and actually looking back also money market funds as they compete with banks that these are whether these are beneficial things is requires thinking about credit lines as well because those are important parts of the payment system and there is this um relationship through the bank's balance sheet between these things so that that's kind of that's what we think is a message um and uh so and now the next step we or many next steps should be taken to make this quantitatively uh more uh to evaluate this and i think that that's that's a kind of exciting one has to think harder about what are exactly the flows of payments between different types of instruments right for example if all of this let's say there was some segmentation of the market for payment instruments where sort of mostly one group of people just uses deposits and others use payment it was used credit lines and deposits together then maybe the the cost is lower etc so there's there's uh so using this framework and the and the forces that we've outlined i think one could make the model richer to get more of these empirical things and then that's uh that's something that that i think it's on that should be on the agenda good one it so if you are done with your response then i would like invite participants to to ask questions so at the moment i i don't have any questions in the q and a and um i'm also not seeing somebody raising their hands so maybe um let me start of course you you may know that the ecb is very eager in starting with looking at at cbdc and trying to to figure out what cbdc would mean so um provocative issue so i wasn't quite sure um i think dirk alluded to that um there is an issue of course with um the the um funding of the banking sector um and there might be friction so so you are just assuming that it goes to the central bank balance sheet so if there are frictions i guess your results will be even worse uh as regards to the welfare effect of cbdc so if the central bank accepts only very safe assets whereas the bank accepts loans that are have some credit risks and are are riskier then um you would get another friction and i guess your results would be even less favorable right yeah okay so yeah there's a good point that i mean that connects also to uh uh the paper by uh by uh um keister and uh sanches where in our we took the approach here to say the only thing that banks are good at is liquidity and there is no uh cost of raising equity and the banks are not good at funding particular types of firms and so that shuts out all of the potential disadvantages of this competition competition with cbdc that other papers have looked at is just about the liquidity and so but you're right i think for especially for a quantitative investigation uh it'd be useful to combine these things i mean our uh model is sort of geared would only work in a in a world where uh all the banks are very large yet competitive and everything is super securitized so that there's no friction on the asset side which is a um start world and there are these other things but both uh cost of raising equity and uh and uh sort of advantages of banks lending particular customers that's something that that's probably relevant and so to quantify the relative uh importance of these things that that that would be interesting yeah thanks so Bruno has a question thank you very much um thanks a lot for this this great paper and it's great discussion um it seems that we are from the paper we're learning something about central bank currency we're learning something about what happens when firms can have accounts at the central bank but but i'm not i i didn't get what we learned about central bank digital currency it seems that the technological aspect of the currency is not really the focus of the paper uh am i wrong yeah so that's so i guess um that's all swept up in this cost parameter and i would say the the fact that there is no digital creates this uh creates more of a potential that that that the new entry uh either from the central bank or from some other deposit only intermediary is actually produces deposits more cheaply okay but it's it's in this very reduced form way that it shows up yeah that's that's right yeah i think yeah that taut has a question yeah martin so i was thinking given that we live in a world where there are money market mutual funds and if we think of cbdc maybe it's it's a lower cost way of providing deposits but only a little bit lower not wildly better could i interpret your model is kind of saying either we should adopt cbdc or we should ban money market funds that is if if one is is desirable the other is desirable yeah so that exactly so okay and do you have a sense sorry go ahead yeah just i mean there's sort of an interesting there's two two ways to use the pay the model to think about money market funds i think one is to sort of think of the 70s when they started right where they are the situation the way you can think about it through the lens of the model is that um there was a cost of producing deposits uh that that increased basically because of the inflation and regulation queue and the money market funds could come in to provide cheaper deposits but the banks were not allowed to match right that's different uh from the the model right now but the model could be useful that and then in the model what would happen is that you get massive this intermediation and it's very costly for society now the current situation with money market funds so the way i look at it from looking at the empirical literature is that that there these guys they came in for historical reasons because of regulation is not clear what the cost advantage currently is because i mean we see banks offer money market deposits accounts they can kind of replicate whatever these i mean the custodian banks can kind of replicate what the money market funds themselves are doing okay and so then that that's basically then i would say exactly what you said so right now there is not obvious cost advantage it's just creating this competition is making it costlier for liquidity management of the economy as a whole that's the argument of the paper and so then that's not a there's no point in having them okay thanks yeah i would have a question to Bruno which is not really a question to the paper but it's something that i have been hearing very often and that i am struggling a little bit with it's always this technology question and of course as Martin said so as a monetary economist you put in some cost parameter and that's it and so you don't think about whether it's a token or whether it's an account or whatever and so you say well there is a cost and and we have so i was wondering do you have ideas how to go about modeling technology choices whether it's distributed ledger or a token or accounts beyond going for a cost parameter that might differ between technology technological choices i wish i did but at least there is some some advances which have been done in trying to model blockchains you know one of the possible ways in which central bank digital currencies would operate would be with consortium blockchains or you know or even centralized blockchains so at least we could try to build in that direction to try to understand the technological infrastructure for central bank digital currencies can i say something i mean i would put the focus on the use so that's what my reading of kister and sanchez is that's what they do for them digital means you can make certain payments with that instrument but not others right so they have different vendors you know there you can settle with different instruments they have several instruments they have physical currency and those are not very good if you want to buy something online so the way they model it is you know they have different liquidity properties depending on who accepts it so that shifts you know that that's an answer you know it shifts it to the use rather than the production side i think the production side is harder to make a case that it matters a lot for these high level questions like you know we've been doing monetary economics for many years without thinking about clearing the fed funds market so it's not it's not obvious to me that those details will matter for the high level questions that these guys are talking about but you know for example there is let's take one one concrete example otherwise it's a bit difficult in general you know if you think of international transfers of funds this is something that currently is not digitized so it works with old technologies you know banks calling each other and using swift to transmit messages and this this is really costly it takes a very long time and it's very inefficient and one could think of having a blockchain a consortium blockchain to make this process more efficient so it's not exactly the same question as what Martin was talking about i realize i'm i apologize for answering it beside the question but but i think there are ways to try to model the technological improvements associated with digitization well this seems to have broken the ice so there are a number of questions now from from the the q&a and i'm going to to summarize a few ones to have them jointly so there is george ponzer asking how the results affect if one considers that a cbdc could foster competition and payment cost reduction and financial inclusion then there is the question whether cbdc um inadvertently could introduce a credit constraint in the economy given that's under control of the central bank and related to this blog is also a question um that's referring to narrow banking so uh whether these arguments also relate to narrow banking so maybe you can answer these questions yeah great okay so first um on competition so yeah that this model everything's perfectly competitive it just zeroes in on the technological features um and there's other work by starting with the work of david and alfato that has argued that you know undercutting commercial bank market power in the deposit market is something that cbdc would be a force that makes cbdc improve welfare that is probably an important for um feature uh given the measures of market power that we have um in our context there is this additional question of market power in provision of credit lines and and so those that is present so there's if you look at um capitalization of these credit card companies which you can think of it i mean those are not guys that have credit uh line drawn loans on their books but rather they just provide the technology to uh to make this all happen that there's network externalities there there's larger ends uh and and so that is something that's uh there's additional force that we want to think about it and roughly i mean the way this would work is that if there are if the banks get taken to the cleaners by uh by kind of these service providers with network externalities then the synergies between deposits and credit lines that we discuss are weaker because there is an additional imperfection there in the credit line market that i think is something that very interesting to um to talk about and that also actually goes to the question of the role of technology because uh you might imagine you know in a digital world the way that market power works is different i think that that's an important question for the future uh then about the overall credit constraint so that in our model because of it's the banking sector is really quite uh that financial frictions are just this um collateral constraint and you can always raise equity and so there isn't an overall credit constraint for the economy in fact that in the sort of the optimal capital stock uh is not affected um by some credit constraint directly it only comes from liquidity demand the effects on that um and then so interesting about narrow banking yes so i mean you can read our results about deposits only versus um credit lines as a result that says uh narrow banking is not good enough because it doesn't provide this uh flexible payment instrument that is credit lines that has an advantage in in our model okay obviously we do not have other features of narrow banking that have to do with financial stability that would maybe provide a counter for us but uh there is a force here that says narrow banking is not good enough we want on the asset side of banks uh loans drawn uh from credit lines thank you so i have another question from james mcandrews so jamey i you have uh your microphone unmuted you can talk thanks very much great paper uh thank you a lot and i i'm still trying to kind of understand the market structure as you see it it seems that uh you know all the asset management has been moved over to money market mutual funds and banks uh consist solely of these uh credit line uh you know they create liquidity solely via the credit line in your in your main thing and i'm wondering do you think the credit lines are the marginal source of deposits in the economy or in you know in practice and and when i'm having difficulty reconciling is in practice we see you know for example as you were saying the credit card banks are separate from most deposit banks and they they fund themselves via asset back securitization um and the vast majority of backing for deposits is is some sort of asset management by by banks a lot of that consisting of reserves which you know paying interest right so um and we get into kind of the david and the lotto uh new market uh you know structure market imperfection uh point of view but uh i'm wondering uh have you thought about a bank that that does combine the the asset management and the lines of credit together and and how how a uh because what i'm thinking is that you know if the rate of interest is set by the central bank then that's kind of the that's the base rate of interest so there's no inefficiency uh credit line borrowers would be willing to pay the safe rate of interest plus the benefit of the credit line and so it wouldn't be it wouldn't uh detract from the liquidity created by the credit thanks okay great so there's a couple of interesting points there so so first about um so the way i think about our um the the model without the entry by anybody who's doing deposits only that would be then then the banks uh optimally combine both activities there's some asset management and there is some credit line provision right and i think i mean roughly speaking in in a system where uh there was universal banking unfettered by regulation in the 1970s like it was in many european countries it sort of looks like that um and then the US is kind of special because it's got these money market funds extra um and and then the cbdc i view it as sort of okay now in all these other countries we're going to do something similar that that's kind of the way i think about that okay and then you'll raise this interesting question about interest rate policy this is something that we're actually working on so what what we haven't thought about uh fully is is there some interest rate policy that can strike an interesting compromise between uh benefits of cbdc just from from uh making having cheaper costs and not generating this externality on the commercial banks and this is so i don't i i know in um if sort of leverage is sufficiently high that's not true in fact so the optimal then you can show that the optimal policy uh is the is then one where you just don't do the cbdc under the conditions that i showed but there may be more generally uh some some room for for short additional results that that that we haven't done it yet but that's interesting so so we're coming close to the the end of the seminar but there are still two open questions so um i read the one the shorter one which is by angela reddish and she asks about the um empirical significance of the result and how you would uh think about um assessing them and the other question is on regulation by larry wall so he is unmuted he can ask that i would also uh ask you to to keep your question short okay this is as much an observation as anything uh because the climate had been made money marcus funds aren't serving any purpose well perversely we're creating a purpose for it in the u.s with the central bank uh quantitative easing uh greatly increasing the amount of reserves we're also increasing the amount of deposits that wouldn't be a big problem if the commercial banks could costlessly hold reserves but post the global financial crisis we've imposed a whole bunch of costs on holding uh reserves or increasing the balance sheet so now the combination of money market funds and the central bank reverse repurchase programs provides a way to deal with this increase in reserves without sticking the banks with the increase in their balance sheet and having to bear those regulatory costs okay great yeah thanks so so angela first on the yeah so i think um so the model basically says how the that these uh externalities that are at the center of the analysis should show up in prices the relative prices of the different instruments um and uh moreover we would expect those to depend on the kind of the patterns of payment flows among different participants and so i think for an empirical approach we want to think about um sort of add some more detail on uh who is using which payment instrument and at what prices that's roughly that the way this would go and and this may mean making the model a little bit richer and having different types of agents that uh partly maybe use things as in as imperfect substitutes that this model is is quite styled but but that is the direction i think which we take this um and then so larry uh yeah so i mean that there is um i always think that the us uh monetary and banking system is uh is the most complicated because it's got all these different regulations that lead to private sector responses that um that in in other context we don't really have and so so i agree that to really to really understand the details of the tradeoffs in the us uh as they occur historically and or even or now uh various other uh governments um peculiarities have to be taken through account so that i'm important to them thanks a lot martin so this ends the q and a session i would like to thank the the presenters and the discussant um and then i would hand back to jonathan to close the meeting um thank you again martin dirk and catrin as well as all of you for participating today you'll be able to find the slides and the link to our youtube video on our website cb and dc.net we hope that you will join us again next month on may 28th when the imf will host our next session until then i hope you have a nice weekend thanks everyone