 Good morning. I am Katrin Asenbacher and I'm heading the Monetary Policy Strategy Division at the ECB. I'm very happy to guide you through this morning's session on central bank searing systems and cryptocurrencies. We will have two papers. One is presented by Maria Suntagianetti and is on money market and bank lending evidence from tiering adoption and the second one will be presented by Simon Meyer, the coming battle of digital currencies. Both papers will have discussant, so the first discussion will be done by Diana Bonsi from Banco de Portugal and the second one by Morten Beck from the Bank for International Settlements. Without much ado, I would like to start the first paper. Maria Suntagianetti will present the paper who has been done with ECB co-authors. Maria Suntag is a professor of finance at the Stockholm School of Economics and she has been serving as associate editor of several international journals, so she is a well-known name in the profession and I'm happy to give the floor to Maria Suntagianetti. So thanks a lot for having me and I'm very sorry not to be able to be there with you. This is a joint work with Carla Vattarilla, Miguel Bocinca, Lorenzo Burlone and Julian Schumacher and what we ask in this paper is whether money market segmentation in normal time can hamper the transmission mechanism and monetary policy. So this is partially a new question, meaning that theoretically and to a lower extent empirically, people have been analyzing whether in crisis time when the money market freeze banks that have been funding themselves mostly through the money market land less. We will consider a period in which banks were not particularly constrained because also the European Central Bank was providing liquidity and what we ask is whether the fact that some banks remain more in the periphery of the market might have constrained their lending policies. So and the idea here is that access to liquidity and the having uncertainty in the ability to access liquidity may lead to bank precautionary behavior and limit bank lending. So let me tell you how we attempt to ask this question. Basically what we want is some shock that affects the money market and how integrated the money market is. And for doing this we use the introduction of the tiering. Basically the idea is that the tiering gives banks incentives to reallocate liquidity. The fact that the banks have a stronger incentive to trade in the money market would decrease the money market segmentation and we ask whether this matters for bank lending. Of course in front of this audience I do not expect to have to go in the institutional details but basically the ECB awarded each bank exemption from the negative interest rate. Of course there were banks that had more access liquidity than the exemption and there were banks with a new exemption. So we would expect that these would give banks incentive to trade. In the paper we use the particular timing through which the tiering was introduced. So the first time that the ECB has been discussing the tiering was in a speech of President Mario Draghi at the end of March 2019. This was just the hinting of the tiering that was however interpreted by the market as a credible announcement. So a first effect of the tiering in this case is a positive wealth effect. So in principle the tiering could have affected the transmission mechanism through this positive wealth effect. We show that at least in our sample this doesn't matter that much. There were other two steps through which the tiering was introduced. In September 2019 all the features through which the tiering would have been implemented were revealed but the actual implementation happened at the end of October 2019. As I will show you at least for the transmission mechanism this is the crucial date. Why? Well in the money market banks transact claims that have a very short maturity. Therefore market participants had uncertainty about what would happen to the cost of borrowing interbank market or ability to employ liquidity until the very implementation. So we use the rich data resources of the European Central Bank and specifically Anacredit and MMSR. So let me get directly into the results to give you an idea of what we find on the effect of the tiering on bank wealth and then on the transmission mechanism. So others that have explored the tiering before us have mostly focused on the wealth effect and what one would expect in the cross-section of bank returns if the tiering indeed announces bank wealth we should expect that the banks that benefit the most from these exemptions that is banks with more excess liquidity experience higher valuations and this is basically what we find for the salient event of the introduction of the tiering that is when the tiering was initially announced and then when the tiering was actually introduced in October 2019. But there is another important effect that is what we will focus mostly on that is the fact that market segmentation in the money markets decreased. So what you observe here is that after the introduction of the tiering banks with unused allowances are able to borrow more in the money market. On the contrary banks with unused allowances are lending more. The effects are there both in the secured and insecure part in the money market. Market segmentation probably implies something that we want to see in the extensive market that is a number of relationships as opposed to amount of borrowing. This cannot be done for the secured market because there is a central counterparty system. But if we look at the unsecured market what you observe here is that in the first graph banks with a new exemption are increasing the number of borrowing relationships that they have with other banks. Instead the banks that without unused exemptions that is these are the banks with excess liquidity are increasing the number of lending relationships with other banks. So on the extensive margin these again suggest a reduction in market segmentation. So we can basically show these effects also in a multivariate setting in which we control for a variety of fixed effects. But what this multivariate analysis shows is banks that were more exposed to the tiering because they had a larger proportion of unused exemptions are those that have higher net borrowing after the introduction of the tiering system. And what we see is that of course as we would have expected given the incentive to trade these banks excess liquidity increases after the implementation of the tiering. Somewhat interestingly these banks also seem to hold relatively less government bonds after the introduction of the tiering. Why? Well probably they have weaker incentive to hold this security because they being better integrated in the money market they might not have to they might not need this security for instance to access the tier row. So let's look now at the transmission mechanism. So how can this shock affect the banks willing to invest to land? Well there are at least three channels. One is the wealth effect because we do observe that banks with higher excess liquidity that can exempt me that the extent that I expected by the market to make a better use of the exemption have higher valuations. But there are other also other two mechanisms that can be applied. One is that we expect negative interest rate policy to affect the transmission mechanism through a sort of hot potato channel. Meaning that since it's costly for banks to hold on to the excess liquidity these banks would have stronger incentives to lend. The fact that now some of the excess liquidity is exempted from negative rates might imply that banks lend less. So to some extent there might be concerns that the introduction of the tiering is contractionary. The noble channel that we introduce and test is this money market uncertainty channel. That is the reduction of the market segmentation increases reduces the bank's uncertainty to be able to borrow short-term liquidity if needed and my increase willingness to lend. So it is basically a reduction in precautionary behavior. So let's first try to see which of these mechanisms is a play. So what we are looking here is the volume alone of a given bank to non-financial corporations in a given month. So here we are using current credit so we will be able to absorb differences in demand with high-dimensional and fixed effects and therefore we will try to test whether the supply of credit related to different cross-sectional characteristics of the banks matter. So what you observe here is that what we call the exposure to the implementation of the tiering that is the proportion of unused exemption of a bank seem to be positively related to the supply of credit after the implementation of the tiering. So we evaluate whether the other two channels that I've just introduced are a play. So first of all wealth effects. So the wealth effects are expected to be larger for banks with higher tiering savings. That is the banks that before the implementation of the tiering add enough liquidity to use all their unused exemptions. We don't see much related to these wealth effects and we also consider this of the potato channels. So does the introduction of the tiering and the fact that some excess liquidity is exempted from negative rate reduces the incentive to lend for high excess liquidity banks. Again we find no evidence of these hot potato effects. So what do we do in the text of the analysis? Basically we explore how robust this finding that banks that we find that expose become better integrated in the money market lend more is. So what you observe here is that we can control for demands either by including the interaction of industry location size and time fixed effect. Assuming that all firms in these kind of clusters have a similar demand for credit or we can also absorb a demand using firm and month fixed effect. Across these specifications we find that banks that benefit more in terms of integration in the money market from the tiering increase their donor supply by around four or seven percent. Now you may think that what I have shown you so far is a bit of black box. I've been interpreting these banks that have relatively higher extensions so have a stronger incentive to borrow exposed in the money market as banks that were exempted at the margin. And this is of course consistent with a lot of evidence but can I test this better? So what we do here is we look at samples plate and cross-sectional variation between banks that face the exante higher borrowing rate in the money market. So if the idea is that if constraints in the exante access of the money market really explain the lending behavior exposed. I should find that the effect is driven by banks with high unused exemption with above median interest rate. And this is what you find across all these specifications. Basically for banks that face relatively high interest rate for borrowing in the money market before the implementation of the tiering these are the precisely the banks that lend more to non-financial corporations after the introduction of the tiering. So the effects that we find are there not only for the quantity of credit but also for the loan maturity and the lending rate. So here look at the loan interest rate. Banks that have higher news exemptions before the implementation of the tiering and the face high borrowing rate in the money market are those that decrease the loan rate to a lower extent as opposed. And these banks also extend the maturity of their loans after the introduction of the tiering. Now what you might be wondering is okay there is a reduction in precautionary behavior presumably due to the integration of these banks in the money market but is these good or bad? Are these banks taking too much risk or are just lending to borrowers that are credit worthy and board constrained? So basically we look at who gets more credit. So whatever you see in this table is that we are in column one to five but we are splitting the sample in above below the median based on the characteristics that you observe here and that are presumably related to the riskiness of the borrower. Basically we observe no evidence that the borrower with high default probability or lower productivity are getting relatively more credit. So we don't think that we are capturing excess risk taking or zombie lending. Instead what column six to eight confirm is that the banks that were relatively constrained extent that is for instance a bank with low bank capital with presumably less access to the money market are those that extend relatively more credit after the implementation of the tiering. So let me conclude what does this paper shows? Well the paper shows in the context of the introduction of the tiering that money market segmentation can actually really hamper the transmission mechanism of monetary policy and that the introduction of the tiering seemed to have benefited the transmission mechanism reducing the most constrained bank's precautionary behavior. Thanks a lot for your attention and I look forward to Diana's comments. Thank you very much for your presentation Maria Sunder. So I then would like to give the floor to Diana for her comments and you as a discussant you have 10 to 15 minutes for your discussion. So thank you so much for the invitation to discuss this great paper and participating in the conference. I'm really sorry I cannot be there today but it would be invisible to travel today but it was still great to have a chance to go deeply through this through this paper that Maria Sunder just presented. So I think the presentation was very clear but at the same time it's a very dense paper very polished with a lot of results so let me try to take one step back and make things even even more simple than what Maria Sunder did. So I guess everyone here knows that in 2019 the EC implemented this tiering system on access reserves and if we go back to the explicit statement on why this was done the goal was explicitly to support the bank-based transmission of monetary policy but at the same time making sure that the negative rate environment was still contributing to an accommodative stance of monetary policy. So how is this actually achieved? So the idea behind implementation is as simple as it could be so you have this access reserves that at the time were being remunerated or actually charged 0.4 interest rates and so the idea after October 2019 was okay let's make an exemption let's allow for an exemption and so the amount of access reserves that would be up to six times the minimum requirements would get a 0% interest rates and everything else would continue to have this negative DFR attached. So what does this mean? I think about different examples, different settings, banks with different conditions to see how exactly this would matter for them. So let's think of a bank that has one million euros of access reserves and half of those access reserves would be subject to the exemption. So what this means is that instead of having to pay four million euros per year to have these deposits parked at the ECB now this bank would have to pay only half of that. So this kind of alleviates pressure on the banks and exerts this wealth channel that you have something to describe and the channel that actually the literature has focused more on. Let's think now about a bank where okay that there are access reserves by coincidence are fully covered by the exemption while that then this bank is now able to deposit everything at the ECB at the zero interest rate but then where things become much more interesting and I think this is where all the empirical strategy of the paper really contributes and allows us to understand some of the puzzles that could exist is to think about these other banks, banks that wouldn't have enough access reserves to meet the exemption. So these banks actually have some room to increase their access reserves and get these zero percent interest rate instead of being charged to do so. So having laid out that the very simple pricing mechanism how does this actually achieve the bank based transmission of monetary policy and so there's these three channels that were just described in Madison's presentation. So there's the first one the wealth the banks net wealth so for the banks that have this excess liquidity of course now they have this profit positive impact on their net wealth and this should encourage the banks to lend more. At the same time there's two other channels okay the second one is is what Madison has just labeled the hot potato liquidity channel so basically what this means sorry what this means is that the higher value that excess liquidity now has might actually now encourage these banks that have unused exemptions to increase their reserves and this would make them lend less but at the same time there's a counterfeit another effect that that could make these banks with excess liquidity to actually lend more and this is exactly where the paper focuses on is the argument that by improving the functioning of the money market through the reduction of uncertainty this might encourage that the banks that were previously constrained to actually now lend more because they face less uncertainty actually they can borrow through the interbank markets in a much clearer way. So when we go through through the results that the paper delivers on all these three mechanisms so sure there's some evidence of this positive wealth effect even though it doesn't seem to be the major driving force but with results that have been shown for other adaptive restrictions seem to be at work but where I think it's more interesting is to see how these second and third channels actually interact and so this possibility that the tiering was contractionary that is not at work at all in this setting what happens is that there's a reallocation of liquidity to the banks that have the new exceptions and these banks lend more and the reason this happens is due to this improved functioning of the money market. So this is a very polished complete paper I think I mean many things were covered in the presentation and many more are in the paper so at this stage my comments are going to be kind of at a relatively broad level and actually thinking about future work and things that could be done. So first I will talk a little bit about more about these mechanisms then about local versus aggregate effects then about how exactly was the money market functioning and then about I mean what are the long-term implications of all this. So when it comes to the mechanisms at work right we have these second and third channels the hot potato and the one related to uncertainty. So the way if we think about these channels they're going to affect maybe two different types of banks and this is what the last tables that were shown tried to address. So for the banks that were possibly unconstrained they might have chosen not to have the same amount of of excess reserves as they could because I mean what would be the point that would be something that would be charged 0.4 percent interest rate. So these banks might now face this opportunity and think okay now I can increase these reserves and to do that to have more money deposited at the ECB these banks would lend less and this would be the banks that would have somehow the possibility to choose which level of reserves they had and they were actively choosing not to have that as much reserves as possibly that they could have now with exemptions. Where things are perhaps more interesting and relevant is for the constrained banks it's for the banks that didn't have excess reserves because they really didn't have the space to do that these banks were somehow constrained in their access to liquidity and so now that the money market is working in a better way these banks can lend and they can borrow from other banks or and with these excess funds now they can lend to the economy okay and as it's written in the paper which one of the two mechanisms is going to prevail is really an empirical question. So the results in the paper are really suggestive of the idea that it's the second effect that dominates much more on average and in the last table that we saw there's a great effort to to look into this its originating to the different conditions that the banks face. I still think that it would be interesting to be able to say something more explicit which I think is already there in the paper but I don't think that it's so much explicitly discussed is looking specifically to these unconstrained banks how do they how do they react and so I think the results are very explicit for the constrained banks and possibly this is the mechanism really at work but the way the results are stated are much more subtle for the unconstrained banks and I think it would be very interesting to see how these banks that have room for a more strategic reaction how did these actually change. Then I think it would also be interesting to to think a little bit about local versus aggregative facts okay so I mean as we would expect from a paper written by by this team of co-authors empirical strategies we're really very carefully executed the causal interpretation is out there but I still I mean given that it's a paper with so much policy relevance I mean especially for all of us here in the audience today I think that in some parts having a bigger picture of the effect even if we have to to let go a little bit in terms of the benefits of precise identification I think it could be helpful and there are specific questions where I felt that I would be curious to learn more about these more aggregate implications of of of what happened after tiering was implemented so for instance one thing that it wasn't clear for me when reading the paper is actually what happens to total liquidity in the banking system okay did actually that the entire liquidity increase or or maybe not I wouldn't see a reason why it would but this is actually not so so if the liquidity doesn't increase then the additional lending would come from reallocation okay and so this means it's a relative effect it's okay some banks like more but perhaps at the expense that other other banks are lending less and so while while it's very interesting to see that okay when we compare the different banks and the before and after the different announcements we see the effects it would be interesting to have a broader picture and to understand a little bit more about this valuation I think there's a lot of evidence going into this but seeing it in a more explicit way I think it would be very interesting another thing is of course we have much better data on on on corporate loans right and especially now we've done a credit we can do this very impressive analysis and and and and really go deep and understand what is going on in terms of how the banks are financing the firms but this is not the only thing that the banks do and so looking into other asset categories where of course I mean we cannot go as strict on fixed effects on or even granularity but still I think it would be interesting to have a broader understanding of the implications of the tiering and so looking into what happened to mortgages what happened to consumer loans on which the TLTRO targets were actually very explicit and so what happened in these credit markets as well I would also be curious to understand what happens to government bond holdings so one thing that government bond holdings are always or almost always a control in all the regressions in the paper and very often they come up as significant for many of the outcomes so this suggests that something might be going on here in terms of bank decisions of of holding government bonds and while it's important to control for that I wonder if it would be interesting to analyze that in itself if there's something interesting happening through through a channel linked to the software then then then my third comment is about how the money market was working at the time so here I think that the paper really offers a very important contribution so we don't have many papers looking into how money markets influence bank lending but those that we have they're focused on crisis periods and so this paper really offers a very important angle by looking at what happens when the money market is actually not frozen the thing is it wasn't frozen but it wasn't normal either right so as the author said that the money market was dormant in in in some segments at the time so I think it would be interesting to think about how special the setting is right and so how many banks were effectively constrained and and how many banks and in which segments were things actually working as smoothly as they could be normal times and and and and these relationship analysis somehow helps us on that but but perhaps we could learn even more and for instance one way to learn more would be to also extend the analysis to other episodes for instance in 2016 the tiering had already been mentioned in one of the press conferences so it would I would be very curious to see of course we don't have all the data for that period but some of the data we do so what actually happened I mean what once the the topic was discussed given that the Draghi meant reference had so much effects in 2019 just a pure mention of that in a speech did the same thing happen in 2016 also trying to understand how many markets and banks reacted in the year area after announcements in other jurisdictions related to tiering in Switzerland or Denmark also figuring out how tier tier announcements changed some of these outcomes and so the the final comment I have is is about I mean the longer term implications of all this so when you read the paper the the results are all very positive and so I mean given that we're at the ECB and we all contributed to this decision it's it's very good to see that this seems to be the case but at the same time I think it would be interesting to to go a bit further and think about the median term effects of course in terms of identification becomes more challenging but I think it would definitely be interesting I think one thing that well surely for future work but trying to understand what happened when the DFR now became zero again at earlier this year I think this would be very interesting to explore as well I mean and to see what happened in many markets and how that changed lending decisions as well so so summing up I think I mean this is a great paper there's there's a very impressive evaluation of the multiple effects that the tiering could have looking into these different channels there's really no stone left in turn so all the channels are debated and evaluated the data is impressive of course and the results are very positive I just have some some some uncertainty about okay maybe if we have a longer term and our perspective are the results still so positive what are the broader more aggregate implications of these in other dimensions of banks balance sheets and and also how special the setting is and how can we extend it further to think about other settings in which these could work but thank you so much and best of luck for the paper thank you very much for the discussion so I would first like to give Maria Sunter the opportunity to react to the comments and then afterwards I would like to open up the floor for questions and I would like to particularly invite also the online participants to submit their questions via the chat so they will be read out by me so I will receive them and give them to the the author and yeah but first Maria Sunter so thank you thank you a lot for the excellent comments Diana we can work on them for months and what we try to say and I think this is the first evidence is that market segmentations seem to matter for bank lending and these segmentations are to some extent a feature of the euro area because it's the money market of several countries with that has a bit different features for instance from the Swiss money market or perhaps even the US money markets and I think this is to some extent the first evidence that I think is correct that the ECB and policymakers are concerned about this segmentation but it would be indeed interesting to look at the money market effects of the July 26 policy decisions and I agree that we should look a bit more at the aggregate effects and there is also Carlo there I understand perhaps also my other co-authors so I don't know whether they want to add something that's but I look forward for more questions no thanks thanks a lot Diana for the for the comments and they were all very useful as Maria Sunter was saying it's probably looking at you had a very fair point with that I was discussing with Julian here that is also present in the DOOM which is what what will happen to all these segmentation the money market dynamics now that yields all rates policy dates are zero or both zero so whether there is this detachment of the effect that we found so this is of course something we can look but at the same time of course the paper is really on the effect of segmentation so this is really one in a stressful period like one that we were in probably this was the measure concerning your other point probably very quickly on the sovereign bond and probably you were inting and potential carry trade effects coming out from the this wealth effect let's let's be so over the period actually the sovereign holdings of banks were decreasing so we will never have some substitution effect between lending and sovereign bond or money that were that were at disposal of banks because of the tilling and sovereign bond so this would actually suggest that there was no cavitate activity coming from the the exercise that I said but I don't know if Julian wants to add anything okay but thanks a lot for the for the discussion so any further questions in the room yes thank you very much for this very interesting paper and thanks a lot for presenting I want to to go into a bit of the comments that also Diana made I was wondering whether you can can explain a bit more how you identify the market segmentation that is before the introduction of tearing no because of the paper tries to exploit the reduction in market segmentation that is then triggered by the two-tier system was wondering I mean market segmentation sounds to me that there is some kind of constraint in the market and if I believe correctly from your regression results that is also related to or you identify that by having higher borrowing rates in the money market but what you see when tearing is introduced is that indeed these trades start to happen but at the same time that average money market rates or that they don't respond so it doesn't seem that these rates are indeed much higher so that to me doesn't really sound like the market was constrained beforehand instead well there was just not a big incentive to trade because there was a lot of liquidity in the market and now we have introduced an arbitrage opportunity so I was wondering how you interpret this that the money market rates basically don't respond that these banks do identify as constrained in the money market before the introduction of tearing they actually traded the average rate that all other banks are also trading afterwards so this is a good question and to me the fact so the tearing that was designed in a way that rates would not respond because otherwise the negative interest rate policy would be no longer there but indeed there is a bit puzzling that on average these banks are borrowing a slightly at the same rate as before said that it is a fact that the banks with less excess liquidity and unused exemptions had much fewer relationships in the money markets and as soon as this shock happens these banks have more links so the way in which we interpret this reintegration and a change in which the money market is working is basically that some banks establish more links and these banks are the ones that happen to have more unused exemptions and up and to have higher borrowing interest rate and I'm sure that Julian can answer this type of question much better so if Julian wants to add something that would be great okay no thanks for the question which is actually very important I think and Maria already made made the important point that of course the system was designed in a way that the average rate indeed would change because otherwise that would have sort of undermining the policy stance per se but at the same time what's what if you look at so if you go one level lower basically you look at the rates at which different groups of types of banks were trading in the market then you do see this sort of segmentation and we actually also document this in the paper which shows that the two groups of banks were in fact trading at different rates now the average rate didn't change that's how the system was designed but still there was already there was still this this this element of segmentation between these two separate groups and by bringing back basically one group of this group of banks had got this additional incentive to trade therefore also you know that is to me then evidence also that this segmentation was resolved to some extent now in the regression stand on the on the effect on the on the real effect or on the lending effect so to speak there we we try to approximate this by by dividing the banks again into groups that have particularly high rates and particularly low rates and and also there we see that those banks that ex underhead higher rates so that were in this group let's say of at least you know paying a higher price for their activity in the market that there we also see stronger effect so there's also because I think in the same direction then in the end of showing that by reintegrating those banks into the market by giving them this additional incentives even at a higher price let's say that also then led to these to these to these stronger effects in terms of their their lending performance okay thank you very much so we are just on time with the first paper