 Good afternoon and welcome back from lunch. My name is Marie Hurova and I'm a senior advisor in the ECB's research department It will be my pleasure to guide you through at this session at the conference Our money market conference would not be complete without a session on repo markets We will have two papers in this session today The first paper will examine how access to safe assets affects the fragility and lending behavior of financial intermediaries The second paper will investigate the issue of money market segmentation Presenters will have 25 minutes Discuss in 10 to 15 minutes and this will leave us some time for discussion Afterwards without further ado. I'll give the floor to Tony Honored from the ECB who will present his paper on liquid assets and financial fragility. Tony floor is yours. Thank you Thank you for the nice introduction. It's my great pleasure To be here. Thank you to the organizers for including the paper in the conference So this is joint work with Marco Machiavelli who has been until very recently at the Fed Board of Governors Giving us nice access to confidential data and he has just moved to UMass Amherst and We've been working on this For a couple of years and we've just finished a substantial revision of the paper So I'm very happy to show you the the most recent version of it so the so this is a paper on money market mutual funds and I'm going to call them money funds for short and they issue shares that are redeemable on demand and then with these funds obtained they invest in short-term debt and They come in two forms. So some of them are government money funds investing in liquid government debt And Others are prime money funds that on top of those asset classes They can also invest in more illiquid and typically higher yielding short-term private debt So think about commercial papers certificates of deposits and as a result of the Illiquidity and and risk of some of their assets We have seen that they're subject to once I mean both in 2008 and in 2020 There were once on on prime money funds in the United States So the question that we are trying to ask and understand in this paper is whether financial stability could be improved by providing liquid assets to prime money funds and thinking about this issue Publicly providing liquid assets Liquid and safe assets could affect financial stability in at least two ways so first we have the relationship between The creditors of the money fund and the money fund itself So if the money fund has access to liquid assets, it does not need to engage in costly liquidation To meet Redemptions to accommodate redemptions. So that would Could reduce the one risk of the money fund If we now move to the next layer to kind of the real effect of this Money funds fund the real economy Corporate firms so as money funds become Safe or less subject to one risk. They can also engage in more lending to the real economy That is even in in terms of stress that could be continued or even improved lending to the real economy improving financial stability via the second channel and the Ambition of this paper is to study exactly These issues both theoretically and empirically So, let me let me give you the results of the paper in one slide so on the theory side via constructing a parsimonious global game model of once on a money market mutual funds So in particular, we're going to show that the provision of liquid assets Which empirically will be by the Fed via the overnight reverse repro program Which uses the strategic complementarity in the redemption of withdrawal decisions and hence it reduces the one risk on Or the one risk of money funds As a result of the greater stability of money funds in a crisis These money funds are actually able to continue lending to the real economy So they are going to have kind of measures of of lending and and and this taking allowing them to extend funding to borrowers exactly kind of in in troubled times and This this lending will be in in terms of illiquid assets so that will be kind of liquidity and maturity transformation and I'm going to have much more details on these two headline results later on and then we take the model to the data we We believe we can offer quasi random assignments To a treatment that is the access to liquid liquid assets So we're going to have a treatment and a control group So some money funds have access to an extremely liquid and safe asset Which is the overnight reverse repo facility of the Fed and we're going to have treatment and control groups and then there will be an exorgeness a stress event That triggers outflows from money funds, which is the debt limit episode in 2013 so the United States is a peculiar Country in the specific sense that once you approve the budget You still have to change the debt limit. So there's always this additional parliamentary process that is required. So even if there's a Clear maturity for any measure in the budget You still have to problem that you might run up against the debt limit and unless the debt limit is Increased there could be a government shutdown that might even be a technical default on on debt so using the combination of the Staggered adoption of the overnight repo facility by money funds and this arguably exorgeness stress event we provide evidence that the public provision of safe assets Indeed consistent with the model which uses financial fragility both at the level of the money funds and at the level of the ultimate borough So let me so I'm an applied theorist. So let me talk a little bit about the model. So we're going to Construct a global games model of investor redemptions So the the money fund has Made some investments Investors receive a noisy private signal about the performance of these investments And as a result of that signal did they decide whether or not to redeem their shares? I mean very much as kind of in the bank one Setting and we're building on a particular paper by Jen Goldstein and in Chiang and in the JFE who have looked at the dimensions from mutual funds But only with one asset So we're going to look at a different market looking at money market mutual funds And we introduce asset heterogeneity So the fund holds a portfolio of whiskey and liquid assets the least risky asset in our application is this lending to the ultimate boroughs via asset-backed commercial paper CDs and so on which is This is kind of fairly illiquid investment. So that's very costly to liquidate At the same time these money funds hold some liquid assets in particularly overnight reverse repo and treasuries so the The overnight reverse repo is the safest asset around It's backed by collateral safe collateral And the counterparty is the federal reserve so arguably No default risk plus it matures the next day. So this is the safest possible asset and usually Treasuries government debt is is safe as well. So usually these two would basically be interchangeable But not in October 2013 in October 2013. We had this debt limit episode And and treasuries Became costly to liquidate I'm going to show you a chart in a moment that is you have nobody thought that the US government would default So if you hold your treasuries until maturity you get back both principal and interest however, if you face large redemptions and you have to liquidate then there was a cost and At least in this episode we see positive liquidation cost for treasuries and I'm going to show you This slide shortly So the difference between control and treated funds is that treated funds have access to this perfectly liquid asset To overnight reverse repo while control funds only have Government debt which in normal times would be would be fine, but in this kind of particular crisis episode was Exposing them to to risk in particular liquidation cost So we we know from this literature of once on mutual funds that these redemptions can impose a cost on Non-redeeming investors. This is what sometimes is called the strategic complementarity Property and these costs may arise because selling involves some transaction costs or some market illiquidity such as these Costs of liquidation of of treasuries and the key aspect here is that this cost is not fully borne by redeeming investors So if I choose to redeem Unfortunately, I impose a negative externality on all of you if you don't redeem from the same given fund Note that of course in 2016 there was a there was a reform and there was it was a change and We can discuss well, you know, we can discuss the efficacy of that reform in any case our data is from 2013 well before that reform so So we are we are kind of in a period where it's very plausible that we have this externality in the strategic complementarity Interestingly once we introduce a portfolio choice or we introduce kind of liquid assets We we need not have strategic complementarity. We could also have strategic substitutability When there are only very few redemptions Then I can use the liquid assets to meet these redemptions and then actually The the pie tomorrow, which is going to be fairly big will be shared among fewer investors So actually if you know the liquidation cost is small and a few redemptions I actually prefer you to redeem just means this amazing great pie is shared among fewer Investors this is not true in a standard bank one because in banking we have a deposit We have deposits which is a deadline claim while in money market mutual funds We have an equity like claim so we share among all of the equity holders at the end So this this makes it theoretically very interesting because we move from kind of strategic substitutability to strategic complementarity Fortunately, there's the seminal paper by Goldstein and Pausen and the Journal of Finance in 2005 Which in 30 pages of appendix give us the methodology to deal with what's called one-sided strategic complementarity We apply this to our setup and generate a Unique equilibrium, which we can then use to Divide comparative statics and especially for the empirical part of the paper testable implications our first Implication is that money funds with access to liquid assets the ONRP will be less fragile in Empirical terms treated funds experience smaller outflows in response to at-risk treasuries during the dead limit episode I do that That's been a change in the monitor Okay So this is kind of the headline result we did funds are less fragile we then slice and dice the sample By showing this is particularly so for money funds with a lot of active investors because active investors respond To news and then the degree of strategic complementarity is particularly high While funds with a lot of passive investors the channel is It's less strong also We've been recently asked to clarify whether this is really kind of a one-like Mechanism at play or there's just some general visco version So we we have provided new and additional evidence in the in the most recent version of the paper Where we show the result really comes from the least liquid funds So this is in response to the shock those at-risk treasuries Those funds that are particularly illiquid Respond most to the availability of ONRP. So that is consistent with our one like behavior but inconsistent with the general risk aversion story and Then the second headline result is that because money funds with access to liquid assets are more stable they Liquidate less of their portfolio ie. They continue funding the real economy more. So let me let me show you a Bit on the empirics If I can this is not responding or now it's responding Okay, so let me talk a little bit about the ONRP. So this is a facility the Fed Started kind of more than 10 years ago and the key fact here to know this is for monetary policy So that at this point interest rates were really low in the US They knew they would start increasing interest rates over time and they wanted to have a firm Crip on on on kind of the interest rate and on kind of monetary policy transmission So they ink introduced this additional tool. So this was purely for monetary policy reasons they if you also look at the the Statements by the FOMC. There was no talk about financial stability considerations so the Eligibility meant that you needed to have assets under management of at least five billion dollars and in terms of identification What's really helpful for us is that in September 2012 you had to apply for this program But then you could only use it in September 2013 And we're going to look at funds that in September 2013. So one year after the application deadline They look virtually identical. So in terms of observables, they're very similar However, the treated funds did not miss the deadline. They actually applied in September 2012 and they were definitely had access to the facility While in the control group those funds missed the deadline. They had not applied although there would be eligible in 2013 if If the Fed would have a new application deadline, but the Fed did not have have it They only had it in 2014 and indeed in 2014 all of those funds with one exception But all n-1 funds did apply to become eligible So we have this this really nice period between September 13 and November 14 where we have Observationally equivalent funds some of them have access to the or an RP and the other stones. So we have treatment and control and All we need now is before November 14. We need a shock So fortunately for us there was the debt crisis in in the US. So that gave us a plausibly exorcist shock event and Here we see so this is October 2013. We see that treasuries Bill yields of these treasuries went up. So in normal times Treasuries are super liquid super safe. So they yield is like one basis point two basis points But here's going up to something between 30 and 50 basis points So that is if you if you have to sell treasuries at during the debt limit episode then you're incurring a loss and So we're combining these two things this Exorgeness stress coming from at risk treasuries in that period so we're going to calculate the share of treasuries that are at risk and interact that with the the kind of staggered adoption of O and RP and we can compute a treated versus control Money funds so the data is Mostly publicly available with the exception of O and RP take up Which is a proprietary is data. You only have access to add in New York Fed and This is our first main the question we want to look at floors, which is the change of assets under management responding to a bunch of controls the Whether we under in the crisis period, which is the debt limit episode and the exposure to At risk Treasuries so the share of at of treasuries that are at risk that are in this period Before the debt limit episode was resolved So this is our first main table. So if we look at sample three, for example, we see that In in the In the first column crisis times at risk leads to a significant coefficient that is negative So if a money fund has a larger share of its treasuries in terms of at risk treasuries Then there will be will be outflows and Then the second column is the treatment versus control effect where we compare money funds of similar characteristics just the treated funds had access to the own RP The completely liquid and safe asset that is out there and that Fully undoes the effect of of the outflows. So we see the magnitude of basically one for one So the having access to a liquid assets Stabilizes money funds in crisis periods So that's our first main financial stability Implication and we're going to have two more. I mean many more but two more tables in the paper where we show This is mostly driven by money funds with a fairly large share of active investors And it's mostly driven by fairly illiquid money funds in the sense that the The average time to maturity is fairly long so they cannot just sit it out if if my assets mature in a day or two Then that's fine. I just wait for a day or two and then I'm using the maturing funds for redemption consistent with this bank one or the dimension Type a mechanism as our model suggests So we've talked about the first leg the first leg was the investors in In money funds and a potential one on the money funds And now we want to talk about kind of the real effects Side where we see how this affects the lending of of money funds to the real economy So we are constructing this measure called prime risk. So this is lending to the To the real economy kind of anything that is a bit risky. So this could be commercial paper at tier two and Foreign bank obligations kind of anything that is not super safe And then again if you look at sample three for example We again first have the effect that during the debt limit period during the crisis period having more exposure to at risk Treacheries which uses your lending to the real economy. So there are actually many many nice stars there So that's a fairly strong and economically large effect but then again in blue if you're a treated money fund if you have access to The overnight reverse repo you have access to the liquid asset then this basically fully undoes the negative effect This insulates you from the shock So our our conclusion is that with the publicly provided liquid asset that Improves the stability of the money fund and allows it to Continue lending to the real economy even in terms of crisis. We do many robustness tests One of them is one concern could be that this is driven by manager skill That somehow kind of an alpha story that some managers are good at avoiding certain risks You find no evidence for that in our sample it could be that take up of overnight reverse repo is Kind of correlated with with Observables we find no evidence of debt We look at the at the possibility that overnight reverse repo actually doesn't help you at all So it's not about that you're not incurring liquidation cost It's just about if you have access to that you kind of get a stamp of approval What's sometimes called the imprimatur effect, but we find no evidence of that There are also some concerns where we find evidence in the opposite direction We're actually if anything our results get stronger and then In terms of I should also say that we have fairly few funds So in terms of constructing treatment and control group we want to be very careful that they look very similar The consequence of that is we actually down to 20 funds So do you you see from those confidence intervals so they're they're just positive But they are they are fairly wide no so the we do find That they you know the pre-existing there no pre-existing trends, but significant trends afterwards But the confidence bands are fairly wide and we do a few other tests in the paper But all of them such also all of these additional channels are not significant Suggesting that we can conclude that the public provision of liquid assets in this particular case By the Federal Reserve in in 2013 by the overnight reverse repo program Delivers two separate financial stability benefits The first benefit is the relationship between the investors and money funds So this is the funding side of money funds making money funds less fragile And the second benefit is that this law of fragility Allows money funds to keep lending in some setups even increase their lending To the real economy even in in times of stress and you know some of that is second tier Asset-backed commercial paper. These are foreign assets. So this is really that's kind of some Mistaking involved. This is really Not just investing in safe things So as a result some concerns that the provision of liquid assets could lead to kind of a flight to safety Channel in times of crisis is not born out in this particular episode and in our data set. Thank you Thank you very much Tony for a very clear presentation the Discussing of the paper is Puriya Abassi who is joining us virtually Puriya, can you can you hear us? Yes Excellent Lights are being uploaded now. We see them if you make them full screen Excellent the floor is yours. You have 10 to 15 minutes. Thank you so much. Thank you Tony and good day everyone I'd like to thank the organisers for inviting me to discuss this very important interesting paper Needless to say that the usual disclaimer applies Well liquid assets are crucial for many reasons right on the one-hand day at the investor level They very very crucial for hedging diversification motifs They actually serve as pricing instruments benchmarking tools They're playing people to role in regulation into provision but also for the transmission of monetary policy and by their effects on Supply and demand and their interplay they also can actually shape financial markets real economic Outcome variable and thereby also affect financial This is one of the reasons or maybe the key reasons Why the provision of safe assets to the financial system has been a core central bank function? During normal times, but especially also during financial crisis as they Help providing an impetus for sustainable recovery Now liquidity provision already prior to the financial crisis of 09 and 08 has been Provided through various facilities the land of last resort facility the interest rate on excess reserves Open market operations, but even more so in the aftermath of the financial crisis So a challenging question could be why is this topic then on the agenda today, right? And it is on the agenda because the provision of liquid assets runs through the balance sheets of depository institutions, right? Which are in the in the pole position to provide it within the banking sector among financial firms, but also to the real economy Also over the last decades. There's been an increasing trend in Participation of non-bank entities with increased footprints in financial markets Also, it's been shown that frictions may actually arise even in the context of ample equity When there is heterogeneous bargaining power or access To central bank facilities slash the lack thereof I think they showed this for the US. I can I did this for the year air Now Tony and Marco actually focused on the latter two points, right? So they look into the increased participation of non-bank entities and try to understand What an access to central bank facilities can actually do in overcoming short-term fluctuations? And they studied this in the context of the Fed reserves overnight reserve purchase repurchase program, which started in 2013 and it basically grants money market funds access to an altered version of the I OER Facility or the deposit facility in the year air and the idea behind this is money market funds lent out funds overnight to the Fed Receive Treasury securities as collateral in the form of tripartite repo and receive an overnight RRP rate and again, this is very similar to the concept that we all know from the I OER facility for depository institutions and Tony and Marco focus on primary MMS simply because those MMS have more holdings invested in Non-governmental securities and they try to find out whether there is a differential effect in these In the way these MMS perform with and without access to this facility and they do this over the exogenous event In 2013 when the depth ceiling standoff in the US actually When I started basically preparing this discussion, I ran across the the paper by a colleague at the board Which I found pretty pretty interesting. So this is the period Surrounding the 2013 depth ceiling Standoff and in five business days ending on October 16th there's a huge net redemption from prime MMS totaled to 15 billion dollars and a Substantial reduction in CP outstanding declines to 20 billion So this set up alone raises the natural question. Does access to a Sort of Tental bank reserve facility help to reduce outflows during such a standoff period, right? And that happens to be the question of the paper and The long hypothesis of the paper is MMFs with more cash-like assets or You may also think of this as those with zero average cost of liquidation by means of holding marketable assets would they can We sell easily a better equipped to internalize shorter fluctuations for instance by meeting withdrawals and redemptions And there are two key insights that the paper generates and Tony forgive me for Simplifying the vector of results that you that you guys produce in the paper So on the theoretic front What is a major takeaway is that zero average cost of liquidation helps MMFs to internalize losses and withdrawals But on the other hand high in marginal cost of liquidation Induces fire sale type dynamics, right? And this is what Tony discussed as complementary one-sided complementary strategies on the empirical front there is Results provided on the differential outflow for affected MMFs depending on whether they had access During the period to this on RRP facility or not This is a super interesting and important paper and topic that it actually Tackles it fits nicely into the recent literature Emphasizing benefit of the public provision of safe short-term aspects and enhancing financial stability as Tony has stressed by Displacing private money-like assets that are prone to Jeremy has a paper on this among many others. It's very carefully executed polished And you can see that it has already gone through the major revision that Tony talked about My comments are more some sort of a melting pot of ideas that I had that I tried to cluster within these buckets and In the interest of time, I'm just gonna go through the first and the second the third bucket and leave the Rest as additional comments that Tony you and I we can discuss over a coffee Um, so when I started thinking about the paper it actually reminded me of How models like the diamond dip big type of model can consider and talk about financial fragility Where financial institution actually must choose between short-term safe assets like investments in the overnight RRP and A long-term technology and an exogenous increase in the return of such in safe asset Which would be consistent with an increase in the over overnight RRP rate Could actually have two effects one Substitution effect because the tendency to increase investments in the safe asset as its relative return makes it more desirable On the other end also to have an income effect because there is now a tendency to reduce investments as one can earn the same income with lower quantity of assets basically the same asset and Exactly, it's not very clear to me which one of those effects will Dominates and I think there is a lot to be learned from whatever the outcome of this rational is Now what I thought about the Theoretical part of the paper and tried to link this with the empirical part I somehow left puzzled and maybe this is because I simply didn't get it on the theoretical front it Felt to me like it depends on the availability of liquid assets with zero cost of liquidation This basically means it focuses on the market liquidity. I use the ease with which one can sell an asset to somehow overcome a trigger which has Initial which has been initiated on the liability side and as at the end of the day depending on the share of marketable assets That each MMF holds this might actually have an impact on the asset and the liability But then when I went on to the empirical part I learned that what it actually does is uses the availability of liquid assets hence cash To borrow treasury securities and the overnight facility Which felt to me like focusing on funding liquidity the ease with which one can ease Use an asset for refinancing and there the trigger in the analysis is basically on the asset side So funds that actually with an ex ante highest share of effective securities go into a troubled Episodes and somehow the question becomes how they actually managed to keep the or maintain the same level of assets on the management this felt to me like we're talking about this swap on the asset side And I'm not even sure you need the literature to motivate for the dynamic that you're interested in This is at least my takeaway from this because the liquidity spiral basically just reminded me of this or even the fixed in the patient fixed income investor Rationale that Hansen and I actually brought forward among many others, right? Okay, so this is on the I have obviously also some specific comments, which as I said I put in the additional comment section at the end of Now when I basically just jumped to the last part, which is what can we learn because Tony you Dressed the fact that it actually helps us to talk about and think about financial stability What is the big picture lesson that we can take from this, right? So ultimately, can you say something on the aggregate effects? The commercial papers the results that we saw at least for the control MMS Raise the natural question can borrow us from control MMS compensate the reduction that you show across other entities intermediaries markets and What happens to the aggregate level of short-term funding because as you correctly said We're talking about 12 funds here 12 versus 8 funds and I understand There's a trade-off by you know having a larger sample versus identification as you mentioned But it would be neat to understand. What does it tell us about the aggregates? was The outflows that you that you see and show for MMS for the control group Is this associated with inflows out elsewhere again feeding it to the question. What is the effect on the aggregate? also the paper seems to have a positive connotation on Lending to the real economy Which is also an emphasis on private money like asset creation and my question is is more lending always better This I think goes to the question of capital allocation slash capital misallocation and where is it allocated and so on and so forth At the end of the day Even if we take everything at face value the question becomes the graceful What about the mechanics of the graceful exits, right? Because the facility has been introduced as a temporary facility and the question becomes once there is a phasing out to be planned What are potential symmetric slash asymmetric effects which in the literature has already been raised? last comments I Can see why there is a beneficial effect That you highlight in the paper, but here's another thought Elastically supplied risk-free assets like an investment in the O and R RRP facility amplify run dynamics or even other flight to quality dynamics 2008 2011 2013 there were episodes were flight to quality episodes all the destinations of safe haven flows, right? Cash that in absence of such a facility might have moved quickly to liquid deposits at banks could go to this deposit Sorry to this facility through government MMFs in results leaving prime MMS Experiencing larger outflows and the reduction in the availability of short-term funding in like repos and CCPs, which would be Very very Importance and interesting for financial stability considerations, right? And all I'm trying to say is it would be neat to have a more balanced discussion about not only the Benefits but also the shortcomings of such a setup and what it's actually does With all the different challenges that we have ahead of us potential Qt restrictive monetary policy episodes and so on and so forth So in conclusion It's a very very very interesting Topic it's a super neat paper. It's already polished my comment basically try to give some hints as to whether The link that has been established in the vapor between the theoretical framework and the empirical part is necessary in the way it is Maybe some additional robustness tests that I didn't talk about today, but I have Part of additional slides Maybe the paper could say something on the big picture lesson and provide a more balanced discussion Of the identified effect from a system-wide perspective and obviously its implications the financial Thank you so much Thank you very much Puriya for an excellent discussion I suggest we collect a few questions from the floor and then give the word back to Tony to Address it all at the end The floor is open for questions for Tony your question Yes, thank you very excellent study and a very nice to have this empirical evidence in the US my question is Would you extend this? Research also to the European situation because we have money market funds here Also, and I would be interested whether this would help the European situation as well Can I in the meantime also add a question? So Tony you highlighted the benefits of public provision of liquid assets through the rp facility of the Fed Any of you what frictions prevent private provision of safe and liquid assets for example through the repo market, okay? Let me let me try to Answer some of these. Thank you. Thank you very much Puriya for an excellent discussion awesome Great job. So I'll definitely take you up on on the coffee and want to want to learn more about the additional robustness checks that That we can do and also It seems that we overlooked some very important contributions in the literature. We're very happy to to cite them and in the new version Yeah it's it's I like how you think about the theoretical model and you know talking about how changes in In the return on liquids like safe assets could have income and in substitutions effects There is so in the context of a bank one model there's there's actually a theoretical analysis by by Philip Koenig at the Bundesbank doing doing exactly This and as you suggest finding in the vigorous a relationship, but what we have done here is actually we we took the the short and sweet way out by Assuming that these liquidation costs only happen when you have to liquidate So there's this kind of difference in the benefit without the cost So in in the environment you suggested there's in there's you know an ex-underport for your choice Where I hold a low yielding asset that is very liquid and helps me meet the dimensions But there's also a risky and high yielding assets and then I have to think about benefits and costs and in order to circumvent this interesting but complicated analysis we make the assumption that There's just this this liquidation cost So we are so we fully agree with this concern, which is why we try to work around it You you talked about market versus funding liquidity, and I want to you know think a bit bit more about that. I I'm not too unhappy about it So if at the end of it you feel like the model is no longer needed because all the effects Intuitive so that's that's not the worst place to to be but but hopefully you know the model itself might also offer some some insights and In terms of the big big trick questions, I think you you make some some great suggestions I think it would be interesting to think about spill overs GE effects welfare effects so far we thought that's kind of beyond the scope of this particular paper But maybe maybe there's some some follow-up work That they could be done One concern is sometimes when you Try to identify something kind of Hopefully cleanly and we already have to problem that we are kind of down to 20 funds Then going back and try to aggregate everything and talk about the aggregate effect. Sometimes this is kind of difficult. I mean I have some Empirical papers with a model where the my identification is not so clean and then actually can talk about the aggregate effects very easily And then I get the opposite usually I get the opposite concern that that concerns about the identification So it's It's it's can be challenging One thing I think we should actually Clarify is indeed implicitly more lending is good. I mean that's a bit of an implicit assumption here Kind of positive NPV Project with some risk, maybe we should We should at least qualify a bit because we indeed don't talk at all About potential costs of lending So we we just happened to there was a question about us versus Europe, so we just happened to Have access to the European data and we we found this lucky coincidence About staggered adoption together with a crisis which seemed to have worked for the empirics But I'm definitely you know, it's very important that we and some people in the room have done what that we look at the European market as well and and Understand these things. I mean What I like about the project the question itself, you know the public provision of liquidity and what are the financial stability Implications in principle. We can use any country and any jurisdiction to study that I mean sometimes us data is actually not so good I know some of the colleagues at the Bank of Canada sometimes benefit from from high quality microdata So and then you can wherever good data quality is available We can use that to try to make progress for these for these questions We have not one of the other things that we have not done and maybe Should do is think a bit more about the private The private provision, I guess it could be that I mean usual information frictions might might might kind of prevent a purely private arrangement to work, but it's probably something we have to think a bit more Going forward. Thank you very much. Thank you very much, Tony and put it out. Thank you