 Our next speaker is Hannah Winterberg from the University of Saint-Galene and Hannah will present a paper entitled Money Market Disconnect. Hannah, the floor is yours. Yeah, thank you so much for the nice introduction. So the paper that I'll be talking about is called Money Market Disconnect. And it is joint work with Benedikt Bayernsiefen and Antoine Ronaldo. They are also both at the University of Saint-Galene and Benedikt is also in the audience and he would be happy to answer any questions too, I think. Yeah, very topical problem, isn't it? Should we try to maybe stop the presentation and open it again? Sometimes this helps. Maybe I can go ahead with a short introduction, what the paper is about. So what we try to understand is whether there are frictions in the money market and how these frictions kind of affect different segments within this money market and whether that leads to market segmentation. If you think about the money market, I think there are two segments that come to mind, the secured and the unsecured part. And we have a look at kind of the interrelation of those two. And in looking at this interrelation, we kind of zoom in into the secured segment. And the secured segment is the repo market, so the market for repurchase agreements. And if you think about a repo, that is typically an over-collateralized, very short-term loan. So it's a one-day instrument between two banks and they basically secured this with a government bond. So as you see in that example, there are kind of two parts. There is this cash and there is the government bond. Therefore, a repo can basically serve a dual role as either a source of cash or as a source of collateral. And once this collateral becomes central to the repo, in that case we observe kind of a lower correlation of those two segments. When the collateral motive prevails, the repo is less funding connected. And we show that there are kind of two sources of this disconnecting mechanism that come to play once the collateral motive in repos prevails and these relate to the central bank framework. Those two aspects, the first one that I will be speaking about is about the access to the deposit facility. And the second one is the eligibility of those collateral assets. Thank you so much for quantitative easing programs. And what we kind of expect to find is that when there are repo rates that are lent by banks that have access to the deposit facility, they can become collateral driven just as repos that are secured by quantitative easing eligible assets. And when this happens, those become more disconnected from funding-based money market rates. And we think that our results are relevant for several monetary policies and have suggestive evidence implications for monetary policy path through considerations. Hey, it worked. So I think we contributed two strings of the literature. The literature on money markets is one of the first things that we contribute to. We show that the money market becomes more segmented when the repo market is predominantly collateral driven. And we identify two mechanisms that lead to this source of segmentation. And then we also have something to say about the literature on monetary policy. So what we show is that there is some form of money market segmentation and that this can be relevant for different monetary policies. And we kind of provide some suggestive evidence that this has to do with pass through. Maybe to shortly introduce the setting that we look at. So we are looking at the Euro area repo market and that repo market has several segments. Two that are well known, I think maybe just shortly to sum up. There's the general collateral or GC segment in which a repo, any government bond out of a list can be used. And there is the special segment in which already before a specific bond is predefined. And this graph that I'm showing on this slide kind of shows some different rates. And there are several things that I think are interesting. First of all, we see that the GC rates, which are here in red, they tend to fall below the deposit facility rate a couple of times. During the sovereign debt crisis and also in recent periods. And if you think about banks that can engage in trades in the GC repo segment or that can use the deposit facility, that gives us an interesting trade off or an interesting decision problem between those two options. And then there's also the special segment. And here we have those various different government bonds and some are eligible and others are not eligible for quantitative easing. And we observe that those rates, they tended to drift apart due to those QE purchases. And that is another aspect that caused collateral to have to have a focus in the repo market. If we jump right in into the first aspect that we consider, which is this access for banks to the deposit facility. This is relevant for market participants and access banks are typically ones that are located within the Euro area. And for this analysis, we focus on the general collateral or GC segment. And we had already seen in this graph that the rates dropped below the deposit facility rate for an extended period of time. And this is something that causes an interesting decision problem. So if you are a bank that is a lender, a lender in a repo trade, then you can basically have two motivations. You can either be interested to have a very short term safe investment with a market based return. That would be a funding motive. Or your motivation can be to source a high quality collateral. Then your main motivation to enter this trade is collateral based. And of course this outside option of using the deposit facility is something that only banks with access to the facility may use. And if you think more closely about this kind of decision problem that they have, the incentive to deposit liquidity at deposit facilities is more attractive when the GC rate is below the deposit facility rate. And in that times they may deposit liquidity at the deposit facility to the detriment of engaging in more funding based repo trades, which already hints to this form of segmentation and motivation between the trades that we want to have a look at. So what do we expect to find? We expect to find an endogenous selection of these excess banks into collateral driven parts of the repo market. So there are different baskets in this market and some are for collateral that are more sought after, whereas others are for collateral that are less sought after or more general have a wider set of funds. And we observe that they kind of self-select into those that are more collateral driven, because when they are funding driven they also have this outside option that can be quite attractive at times. And then we also think about those other banks that don't have access to the deposit facility and we find that they have a higher need for cash immediacy. To kind of jump right in into the regression analysis, so what we do is kind of a panel regression and this graph basically illustrates the approach and the graph has basically two panels. You see on the left hand side the time periods when the GC rate is above the deposit rate and on the right hand side you see a panel that pertains to time periods when the GC rate is below the deposit rate. And we see two observations here, one for excess and one for non-excess banks. And what the graph plots is basically the change in the repo rate as a reaction after a change in the unsecured benchmark rate, which in this case is the EONIA rate. And this graph illustrates what we have in mind with this segmentation and something becoming collateral driven. On the left hand side, when the GC rate is above the deposit rate, then those two groups basically have a very similar reaction. Once the GC rate is below the deposit rate, the reaction of those excess banks is more or less muted. And we think that is due to them self-selecting into these collateral motives and then being less connected to funding-based market parts. The empirical results are basically a formalization of what we just saw in this impulse response. That's why I would like to very quickly go over them. So the panel regression formalizes these reactions and has a look at this time period when the GC rate is below or above and those two groups. And what we observe is that in general, those GC repo rates react strongly to changes in the unsecured benchmark rate. But those excess banks, they react less strongly and in particular in times when the GC rate is below the DFR. And what we also see in thinking a little bit more about this mechanism is that the reaction becomes even less stronger the further those GC rates are below the DFR. So if you think about this decision problem of the bank, for them it makes a difference how much further the GC rate is below the DFR and the trade-off between using a GC repo and the deposit facility becomes more important when the difference is larger. And this is exactly what we observe. And in the paper, we provide a variety of robustness checks for those two results that I've just shown. And the results remain significant across those specifications. We have a look at another aspect to have an illustration of this mechanism. We have a look at this kind of tiering program as some form of natural experiment. And that is kind of an out-of-sample analysis. And we have a look at those banks during this time period just before and after the introduction. And the tiering program, maybe to give a short introduction, what it did is basically an exemption of some part of excess holdings from the negative interest rate. So they would be renumerated at zero. And that kind of makes this endogenous selection even more attractive. And we find that just looking at this time period around the introduction, we find a fact that is in line with this idea that we had that this happens due to the DFR being attractive. The second aspect that we look at is the eligibility of the bonds in these repo trades for quantitative easing. And in this analysis, we are looking more closely at the special repo segment. So if you think of quantitative easing and the rules there, there is one program that is I think the most prominent one, the public sector purchase program. And we have a look at the implementation rules of this program and use this to define eligible government bonds. And we also do this retrospectively to define hypothetically eligible bonds. And this creates some form of difference in difference setting. Why does something happen to those eligible bonds? I think that has also been discussed in the literature is that there is a specialness premium for QE eligible assets. And this means that they are scarcer and therefore those assets or in our case, the repos secured by those assets, they become more collateral driven. And this being collateral driven kind of disconnects them from funding based money market rates. And yeah, another thing that we will be interested in is to see if we have similar reactions of those eligible and non eligible assets before the introduction of the quantitative easing program. To jump into the empirical results, the idea here is very similar to the results that I discussed before. So we look at those different reactions for two groups for the eligible assets and the non eligible assets and to time periods during QE and before. And what we observe is that those eligible rates, they react less strongly than non eligible rates to changes in the unsecured benchmark. And this is about half of the sensitivity. And another thing that we observe is that those two assets were not significantly different before the quantitative easing program. So we think this is an effect of the program. And we again provide a variety of robustness checks in the paper. Speaking of robustness checks, we also have a look at a couple of other robustness checks or extensions for these ideas. In particular, we have a look at these kind of rates that we have a look at for the unsecured market. So the results that I have shown you so far were for the Aeonia rate. And we use that as the benchmark unsecured money market rate. But there can also be other measures for funding based money market rates. For example, one thing that we consider is a combination of the Aeonia and the Esther rate, but also other derivative based rates. And the results are very similar. One idea that I'd like to talk a little bit more about is that we also have a look at the rate on the ECB's GC pooling basket as kind of the most funding driven secured market rate. And we do find something similar and that points to there also being a segmentation between the more funding based baskets and the less funding based baskets. And finally, I have been talking about kind of this access for the general collateral segment and the effect of QE on the special segment. But of course, these could also have inter relating effects on top of each other. And this is indeed something that we find in the paper. Like I shortly introduced in the beginning, we think that this has some implications or maybe suggestions for monetary policy. Because collateral has such a prominent role and collateral have been occurs in many situations. And one policy that seems to be interesting to have a look at in our example, which is why we included it is the tiering program. So just the tiering kind of makes depositing funds at the DFR more attractive and could kind of foster create an even stronger segmentation. Another aspect that could make depositing funds at the deposit facility more attractive is kind of some amendments to the capital requirements regulation. They could their reserves are excluded from the calculation of the leverage ratio and that could make them very attractive. And you could say, okay, why is that something that we should care about? Besides this segmentation that we talked about, another aspect that is of relevance is price determination. So if there is the amount of lending goes down and that's very strong, then there can be a negative side effect on price determination as has been discussed in the literature. Basically, when we did this analysis, we came about the idea after reading this quote, which is from former ECB executive Benoit Corrie. And he mentioned that there might be the case that under the central bank framework, some short term market rates would not respond fully to changes in key interest rates, which is very similar to what we have shown in this analysis. And going further that that could lead to a continued dispersion of short term rates and that could have a negative adverse impact on the transmission of monetary policy. And we kind of follow these steps to have a look at the dispersion and the transmission as like some suggestions that can be drawn from our results. So we observe that there is a dispersion in both market segments. So we compute a dispersion measure. We take a methodology that is similar to a paper from Duffy and Krishna Murthy that has done a dispersion measure for the US. And we compute such a measure and we observe that the dispersion increases when the GC rate is below the deposit facility rate. And also in time periods when the QE program was introduced and expanded or extended. And in the paper we also show some suggestive evidence for the pass through to lending rates. To conclude, I hope that I was able to convince you that there are kind of two aspects of the central bank framework that have led to a form of market segmentation in money markets and that this has to do with collateral being a central motivator for the trade. And those two aspects are the access of banks to the central banks deposit facility and the eligibility of collateral assets for quantitative easing programs. And these two aspects they do have this effect when the role of collateral becomes central and then those money market segments are less connected to funding based market rates. Thank you so much for your attention and I very much look forward to the questions and the discussion. Thank you very much, Hannah. The discussant is Sebastian Infante from the Federal Reserve Board. Sebastian, good afternoon. Good morning for you, I suppose. Can you hear us? I can hear you. Can you hear me? Yes, perfect. And we see your slides. Great. Fantastic. Fantastic. Well, thank you for the introduction and thank you for the invitation to discuss this interesting paper. I would have loved to have been there. I think this is a fantastic conference. I've only been able to attend virtually so far, but hopefully in the future I can actually go there. Because I think I look forward to the discussions with folks there. As was mentioned, my affiliations with the Federal Reserve Board, so naturally the typical disclaimers apply. So what's the motivation for this paper? It's basically just to underscore that, you know, repo has two legs. One is a cash and the other is collateral. And therefore the contracting, there's different economic motives that you may want to participate in these markets. One is to invest your cash in safe manner with high quality collateral at prevailing market rates. And the other is to source collateral specific or of a specific kind, if you will. And obviously these different economic motives to trade result in different repo rates and different contracting terms in general. And so the perspective of this paper is that will central bank policy choices such as the remuneration on reserve balances or asset purchases can affect repo lenders to incentivate that repo lenders incentives to participate in these markets. And therefore, you know, the actual prevailing market rates that we observe could actually depend on the lender's identity. If they have access to the deposit facility or the collateral availability, which depends on the amount of asset purchases that have been taken place. And so recognizing these spreads between repo rates, these wedges, if you will, can have an effect on monetary policy transmission. That's where this paper is going and that's where this paper is contributing to the literature. So what do they actually do empirically? Well, they basically the authors basically study daily changes of lender and collateral specific repo rates and daily changes and unsecured rates and the only is the main rate that they work with. And effectively, what they find is that rates are below the rate on deposit when the rate of repos are below the deposit facility bank lender and quality QE eligible collateral rate repo rates are less responsive to changes in unsecured rates and sort of a picture is worth 1000 words, you know, kind of explain this very well and sort of, but I'm just going to reiterate it here. What we have is seeing scenarios when the GC rate is above the deposit rate. And there you see there's no difference between access banks and non access banks in the dark and gray dots and plots and whiskers. And then on the right, we have when it's actually that a GC raised below the deposit rate and low and behold here, you can if you can see my clicker, we can see that, you know, access banks really just do not change your interest rate. Right. And so the interpretation is a large fraction of repo trades are collateral driven. Thus, monetary policy transmission, which operates in part through changes in funding rates. There are many other channels obviously can be affected by this prevalence of collateral driven a repose and that that then that consequently can be affected by by central bank actions. So that's kind of like the main takeaway of the paper. So, for today's comments, I have, you know, three more substantial ones. And but I'll start with the first which is sort of more. Let's say superficial. Don't listen to me. This paper, I already got this advice once before this paper is second round R&R and RFS congratulations to the authors. So you really just have to listen to the referee and the editor and sort of whatever I say, you know, just think it's a cute sort of reference, maybe for a future paper but just sort of just keep that in mind. And then I'm going to talk about sort of, you know, what is this issue of collateral driven repose, you know, how prevalent is it. And, and then I want to push back a little bit on what to the authors mean by collateral driven repose because as I'm going to explain the policy prescriptions are going to be wildly different depending on what is actually the collateral driven incentive. And finally, I'm going to briefly touch on the role of regulation. So, this is just, you know, the first point just to clarify, lower rates for collateral driven repose is a very well documented phenomenon. So, you know, starting with Duffy in 1996, and others who have subsequently, you know, worked on this area, they document that Treasury repo rates can trade special as below prevailing market rates, right. And this has been wildly recognized. Even in the policy space, where sort of recently the Federal Reserve has been publishing and promoting the SOFR, the Secured Overnight Funding Rate, and its construction actually recognizes that, you know, you can have rates that are below prevailing market rates and therefore you shouldn't consider them in the calculation. And so here I have in a specific, from a paper from colleagues here at the Fed, we have a snapshot of the distribution of rates for a given day on the x-axis on the, and the x-axis we have rates, y-axis we have, you know, take up. And then, you know, this green spike is the overnight R&P that, you know, Tony really unpacked in the previous presentation. And what I want to draw your attention is that there's a lot of rates in the DVP market, which is the centrally cleared market for seasoned collateral and on the run, which is very well below these prevailing market rates, and therefore we just don't consider them, right. Now there's a very crude sort of fix to the problem, but just to make the point that this has been sort of well recognized and policy makers at least in the context of construction, this in construction, this indices have already taken into account. So the more normal part is about monetary policy transmission. And sort of if that is really where we want to go, I think it's incredibly important to know what is the collateral driven motive, which I'm going to detail in the next few slides. So as I said, the paper talks about collateral driven repos, but never takes a stance on why lenders want to source this collateral. And so this is going to be relevant for the public policy implications. There are a few answers. One, it could be source specific for short securities. That may not be that prevalent in the GC market, but in the SI and specific issue for cert for short. It could be that the profits that they're still profits if the borrowing rates of these firms is relatively low. Right, so in the matchbook repo, you know, they're still going to make a profit even the rate is low. And the other is the final one that I want to bring up is sort of the distribution source in distributing safe assets. And this last one is kind of really interesting because Angelou and Benedict actually have an incredibly interesting paper where they completely unpack the differences in repo rates and they call it they characterize the safe asset carry trade. And so I was very strict and I was very surprised that they didn't try to say whether it was the safe asset demand that was driving these wedges. And so in earlier work, I've also shown that, you know, when the demand for safe assets goes up repo volumes actually go down is a little bit, you know, answering Maria's question, you know, what happens with private provision of repos. Well, in fact, this is this paper answers that question. I welcome you to take a look. And more recently, I have a paper with called here in the Fed where reuse of Treasury securities also increases with a man for safe assets and this is the idea that the distributing more securities when the demand for safe assets is high. And so, if it is the in fact the demand for safe assets driving this what should be done. But as I said, there are many other reasons that, you know, could not just be safe asset demand there could be other reasons, and sort of I would invite the authors to tease this out. And one specific way that they can do it as a particularity of the contracting terms in the in the in Europe, is that effectively they had there are many different contracts that have different settlement dates of delivery of collateral, basically overnight tomorrow in the spot next. And again, you know, Benedict and and Angela do a really really careful job in studying this and their other paper of when they actually receive collateral. And so, if we thought that it's, you know, actually the reception of the collateral that matters, we would see a differential sensitivity for these different contracting types, and that would really bolster the message, right. And the other issue is like, I see on the program there was another interesting paper on on sort of the effects of securities lending program, and sort of if it is that quantitative easing or asset purchases more generally is in fact draining securities, then why isn't the securities lending program alleviate the problem. And, and sort of, again, this points to what is it that agents really want to source collateral, what is the problem. And again, this is much more informative to what should be what should be done. You could also think that maybe the securities lending program isn't is an effective could be reasons why you may want to change that, or other sort of policy prescriptions. So, so turning to my legs, my, my, my next slide and I'm sorry I'm speaking a little fast I'm going to make sure I'm on time. I want to talk about the rule of capital regulation Hannah briefly mentioned this and sort of you could do something about capital regulation and ameliorate this problem. I think that there, there's there's much more to say about this so I'm cognizant that in in EU mainly capital regulations are major measured on quarter end. But you know there's a lot of nuance and how banks respond to capital regulations you know you can think about there's international firms you can think that their investors also investor case also cares about these things. And so therefore, you know you may have these effects that are prevalent throughout the year. You know as a risk management perspective that makes sense. And so if you read carefully Duffy Christian murky they actually suggest that regulation could be an important part for this limited path through pass through that on our co authors are trying to promote. And so, if it is regulation, the setup of the market in Europe is going to be a great place to kind of inspect this. Why? It's because the three platforms where the data comes from broker tech urex and TS are essentially cleared repos. Right. And so what's the idea. The idea is that you have if you have a bank with a large repo liability, right in one of these centrally clearing mechanisms that it may be incentivized to lend and take a haircut on the on the on the rate basically accept the lower rate in order to net down their gross positions. Right. In the in accounting rules, you can net down your gross positions and repo if they have the same maturity date and are cleared on the same platform. And so effectively, this creates an incentive to you to you to want to deploy funds into a trade that you know in order so that your balance sheet just looks smaller for regular regulatory purposes. And in fact, recent theoretical literature, you know, starting with Duffy, Hey, Nagel and song have really emphasized the role of the SLR supplementary leverage ratio and low yielding balance sheet intensive activities such as reverse repo and positions in in Treasury markets. And more recently I have a paper with co authors in the Fed, where we actually show this empirically that effectively the SLR is really affecting firms participation in the US Treasury market. And so, if it is the SL, if you believe that it's the SLR or any other capital regulation that doesn't risk weight and you think, and then one possible solution is expanded central clearing which is also something that's been discussed. And so this would effectively allow agents to reduce the balance sheet size associated with this activity. Right. And so this is very, very different than from what. So the policy prescription would be very different than changing the central bank implementation frame. This is talking about changing the regulation. And specifically, there's the way that the regulation is implemented and the way that is calculated. And so to be able to be able to discard or rather support the role of capital regulation, maybe the authors may want to consider the net lending positions of these firms that are accepting low rates. If they are very, very, if they have a huge imbalance, specifically that they have a lot of repo borrowing, then they may incentivize to add that extra dollar to the lending in order to reduce the net balance sheet. Right. And then you could also look at different CCPs, because this is going to be isolated any CCP. And so you could potentially discard this, which, you know, would be would be incredibly interesting. So, that's sort of all I have. I, again, I think this is an interesting paper. I think that documenting thoroughly widening of spreads across money markets is something that, you know, it's good to observe it's good to know. I think the real, real punchline is, you know, thinking about how this really affects monetary policy transmission implementation and transmission. And to the extent that that is really what we care about. I think it's incredibly important to actually spell out what is this collateral driven motive, because that's going to have wildly different implications on what you actually to do to try to address this. So, that's it. And thank you for the opportunity to present and look forward to the discussion. I suggest we collect some questions from the floor and then let Hannah respond both to the discussion and to the questions. There's a question over there. Yes, thank you. Your mark for me on APG management. To build on the story that the colleague just presented, I think that in your current research you look at banks with access and banks that do not have access to the deposit floor. I think if you broaden it to non bank institutions, then you come also across the ones that have different collateral motives, i.e. search for safe assets, asset managers, pension funds. They hold large liquidity buffers due to regulatory standards, clearing markets need for cash liquidity. And we reinvest that in markets through reverse repo and we are searching collateral every day in large size. And that goes below the deposit facility. So the motives for safe assets are growing and growing and the demand for collateral also. So if you would look beyond only banks, but also into the non bank sector, then you find the real motives of why the GC rate is dropping below the deposit facility. And then also crossing over to the previous paper, there one argument was made to allow money market funds to overnight reverse repo facilities, not only for financial stability. But if you consider the non banks also part of the transmission mechanism and operating below the GC rate, if you open up to an overnight repo facility, then you also take away part of the inefficiency that is currently there caused by this deposit floor. Then you create a level playing field between banks and non banks, and also have a better transmission mechanism. So maybe something to research. Hannah, over to you to respond to Sebastian and to the question. Perfect. Yeah, first of all, I'd like to thank you very much, Sebastian, for your sorrow discussion. And we were very glad to have the opportunity to receive your feedback on this paper. I think your comments kind of started with this more general question why this is interesting, especially since there is some literature on like the levels of repo rates. I think what we bring to the table on top of this is that we are looking at kind of the sensitivity of this repo rates in response to unsecured, unsecured developments of funding based developments. And I think that is something that in Europe is maybe also of interest because of this discussion about kind of the benchmark rates because I think in the US was so far now there is a secured rate. So the repo market starts to be very close to this remains very close to this. Well, in Europe, we have this kind of Esther rate, which is still unsecured. So that kind of brings this trade off a little bit more into the focus. And I think another aspect that we look at that hasn't been discussed in the literature also on the levels is this aspect of access to the deposit facility for the European market. And yeah, I think you were also talking about kind of the motivation for banks to want collateral, so why their trade can be collateral driven. And I think there is another difference to the US market here that I think I can shortly mention is that there are various GC baskets much more than in the US where it's like Treasury GC and that is it. So there is another differentiation. So some banks can, for example, use those collateral in margin accounts where some collateral is valued more. And if you place their government bonds as collateral, you don't have a negative interest rate like you would have on cash collateral. That's why this can be quite attractive just as like one example. I think you were also mentioning these regulation aspects and I think these are very interesting aspects and there are many regulations that I think could be mentioned. The collateral plays an important role in various of them and so that will be a very long chapter. But I fully agree that this is something that is very interesting to look at in more detail. Yeah, so I think these were very valid sorts and well taken points and I think we will try to incorporate them into the paper. And I think we also had the point about the non-bank institutions. I think that is something that is very interesting and was also looked at in the US market I think. I think in the past and I think that is something that would be highly interesting. I'm not sure if we can find any data on this, but I would be glad to talk to you a little bit more about that. It would be very interesting. And yeah, I think to conclude, I'd like to thank you so much for the comments and if you would have time to send me your slides, that would be very amazing. Thank you so much. Thank you very much, Hannah, Sebastian and everybody. This concludes the session. Please join me in thanking all the speakers and discussants.