 Good afternoon. It's a great pleasure for me to introduce for third year in a row the winners of the ESLB research prize. As the Interlector has mentioned, this is an annual prize that was established in 2014, first awarded in 2015, in the memory of Eke Banderbu, a longstanding member of European Parliament, that was also a member of the first ASC at the ESLB, the first advisory scientific committee as a, as a Euro-parliamentary, she was also defending the idea of having something like an advisory scientific committee in an institution like this. The ASC runs this prize in collaboration with the ESLB secretariat, reading and assessing the papers submitted in response to an open call. So those of you who have contact with young scholars producing research in the field of macro-pru, you should encourage them to be aware of the next call, which will come in a few months' time. The idea is to award the prize to young scholars, young in the broad sense under the age of 35, if I'm not wrong, with an auto-standing research paper on one topic or several topics related to the ASLB mission. This year, the ASC was confronted with a difficult choice. The papers were very good. We have plenty of good papers in the short list. So we deliberated. We had a couple of voting rounds with a tie. So eventually, the ASC decided to award the prize ex-equal to two winners. So you will not see one winner today, but two winners. So we will have to compress the presentations in the remaining of the time to see each of the winners. These two papers actually share been excellent pieces of empirical research on topics that, as you will discover, are clearly connected to macroprudential discussions. And in each paper, you will find an element of great originality. So I'm not going to reveal my own wording of what that element of originality is, but you will discover in the presentation of the solo authors who are winning this prize this year. So without further delay, I'm going to proceed to announce the names of the winners. I'm going to do it in a strict alphabetical order, because there will be a first and a second, but this doesn't mean that we are ranking the winners. This is ex-equal. And novelty this year is that there is a small object that is going to symbolize the prize. And I'm going to also deliver to the winner. So let me say who the first winner is in alphabetical order. This is André Silva. I think in Portuguese national. Currently an economist at the division of research and statistics of the Federal Reserve Board. And he's a winner for the submitted paper that is called strategic liquidity mismatch and financial sector stability, a paper that is now being accepted for publication in the review of finance studies. So André, please come to the podium. And I will give this prize to you with great pleasure. Our second winner, again, in alphabetical order is Guillaume Vilme, currently an associate professor of finance at ASC Paris. And his submitted paper actually had a longer title than it has. So I'm going to announce the current title of the winning paper, which is the value of central clearing. This paper is, and I think this sounds that the committee has good taste. It's now accepted for publication at the Journal of Finance. Guillaume, please come to the stage. Congratulations. Thank you, André. Thank you very much. Please take your seat. OK, so now I think that the rest of the time is for the winners to solve their achievements in these very good papers. So let's start. Let's start with André, who is yours. OK, so first of all, I'd like to thank the ESRB, Javier, and the rest of the scientific committee for the invitation to be here today. It's a great honor to receive this award and to be part of this conference. So this was my job market paper of my PhD at Cass Business School in London. But because I'm now at the Federal Reserve Board, the usual disclaimer applies. So everything I will say here are my own views and not necessarily representative views of the Federal Reserve. OK, so let me start by giving you some motivation into this paper. So as we know, since at least the seminal work by Diamond Eddick Vig, banks have a unique ability to create liquidity by financing illiquid or if you want long-maternity assets such as corporate loans with liquid or if you want short-term liabilities such as demand deposits. And crucially, this combination of lending on one side of the balance sheet and deposit taking on the other allows banks to protect firms and households against both idiosyncratic and systematic liquidity shocks. And there is also pretty evidence showing that it also helps promoting economic growth. However, there are always two sides of the same coin. And the bad news here is that due to their fundamental liquidity provision role, banks are also intrinsically fragile. So as the financial crisis a decade ago clearly shows, excessive liquidity means much can lead to bank runs, the breakdown of funds in all sales markets, as well as these three-sided sales that threaten not only the solvency of individual banks but the financial system more broadly. Now we kind of knew all these before the crisis. What is new and one of the main insights that came after the crisis from the theoretical literature in economics and finance on this issue is that this relationship between excessively good transformation activity and financial visibility can be further exacerbated when banks collectively as a group engage in strategic risk-taking behavior in the form of common portfolio choices. So I think it's important to take a step back here and try to understand, well, why would banks have an incentive to engage in collective risk-taking in the first place? And there are at least three very reasonable explanations to why this might be the case. The first are bailout guarantees, in which a bank might have an incentive to take on more risk. So for instance, by financing more and more long maturity, for instance, corporate loans with more and more short-term liabilities, if this bank observes that their competitors are doing exactly the same thing, and why is that? Because if there is a shock and one bank gets into trouble, well, probably all the other banks will get into trouble as well, because they were doing exactly the same type of risky activities. So the lender of last resort will have no alternatives, then bailing them out altogether in case of generalized distress. So this is the classical too many to fail story. This collective risk-taking can also be explained by contractual features in the compensation of bank managers. In fact, there's two recent theoretical papers show relying on relative performance valuation in compensation packages incentivizes banks to choose investments that are increasingly correlated with their peers, thereby increasing systemic risk. And importantly, while public guarantees would magnify this mechanism, RPE and associated correlated portfolio choices can still generate systemic risk, even in the absence of a lender of last resort. And finally, this collective risk-taking behavior can be driven simply by learning. Or if you want free riding and information acquisition in which a bank particularly fits small, might mimic those that things that have greater expertise. So if this channel is at play, we will observe at least in principle, small banks mimicking large banks, but not vice versa. It will not make too much sense. The key point I want to make here is that despite this extensive theoretical literature on this issue, this collective risk-taking strategies among banks have not been yet empirically tested in a convincing manner. So in case I lose you towards the middle of the presentation, this is the main thing I want you to take home. So in this paper, what I do is to show empirically that first of all, commercial banks strategically incorporate their competitors' liquidity mismatch policies when determining their own. And second, that these collective risk-taking decisions have a negative impact on financial sector stability. Okay. Why is this important? Given where I am, I don't think I need to spend too much time on this, but let me just state it for the record. So commonality in portfolio exposures as well as a reasonable high levels of liquidity transformation activity obviously increases the likelihood that banks fail altogether, which ultimately can sow the seeds for costly crisis as we all know. And this issue is particularly relevant after the crisis with both academics and policymakers questioning the efficacy of recent liquidity regulation reforms. So Javier has done some work on this. The ESRB is talking about these issues since at least 2014. The ECB has published a report in October last year, the ECB Task Force on Systemic Liquidity that again talks about these issues and the importance of regulating liquidity risk from a macro-prudential perspective. Okay. So before I tell you in a bit more detail the results in this paper, I think it's important to explain you why it's so difficult to identify these peer effects from an empirical perspective. And here, just consider a simple linear image model in which the liquidity transformation activity of a given domestic bank I that operates in country J at IMT is modeled as a function of the average liquidity transformation activity of its peers, of its competitors in the same country, as well as a bunch of controls, okay? And here the peer effects of interest will be captured by the coefficient beta basically that basically captures the influence of competitors liquidity in much positions on those of bank I. So there are at least two well-known issues that one should bear in mind. The most famous, the most important one perhaps would be that the strategic reactions are intrinsically simultaneous. So this is a classical reflection problem of Charles Muskie and here is very simple to understand. If the liquidity mismatch position of bank I is affected by that of its competitors, then the liquidity mismatch position of its competitors is also affected by that of bank I, okay? And the problem here, so in other words is that each bank affects and is affected by all the other banks in the system. So one cannot disentangle if bank I's decision is the cause or the effect of its peers' choices. The other potential issue is correlated or if you want common group effects in which banks in the same local network, in the same peer group are subject to common but unobserved shocks for the researcher that lead them to choose similar policies, okay? So here we could observe similar behavior among banks not because of any type of strategic behavior but simply because they were subject to similar shocks. So due to the absence of an experiment or even a quasi-natural experiment that I could exploit in this setting, I instead use a so-called instrumental variables approach. So basically a variable that induces variation in the dodgeness variable of interest, so the liquidity mismatch position of peer banks but that does not affect directly the outcome variable. So the liquidity mismatch position of bank I. So in a bit more detail what I do is to explore systematic differences in peer group composition. So different peer groups for different banks in the sample operating in the same country. And having this structure of connections with partially overlapping peer groups allows me to use the liquidity mismatch position of a peer as an instrument, as an IV. The question here is how? So how do banks of similar size and similar business model operating in the same country have different peer groups? So the key feature I exploit is that large cross-border banking groups tend to manage liquidity on a global scale as well as coordinate their risk management policies within the group. So it might be reasonable to assume and that this is exactly what I do in this paper that in addition to the liquidity choices of its direct competitors that operate in the same country, a foreign home subsidiary also takes into consideration the overall liquidity mismatch position of its bank holding group when determining their own. So in this bank holding group operates in a completely different country. Let me give you just very quickly this example just to illustrate. So here all these banks have similar size, similar business model, they all operate in the same country. The only difference between them is that bank A is a foreign home subsidiary and the banks C1, C2, C3 and C4 are the competitors of bank A that are purely domestic banks. So now we introduce this bank X, this bank holding company that is based in a different country but that owns the foreign home subsidiary A and now you start seeing this destruction of connections with partially overlapping peer groups. Why is that? Because any domestic bank here, so for instance bank C1 only has four peers which are the banks that are operating in the same country while a foreign home subsidiary, the bank A has both the four peers, so the four banks that operate in the same country but also considers the liquidity mismatch positions of its bank holding group that is based abroad, okay? So what this means in practice is basically that one can use this liquidity mismatch position of the bank holding group, the bank X that is based in a different country as an instrument for the liquidity mismatch position of the peers of banks C1, C2, C3 and C4, okay? The identifying assumption here is simply that a domestic bank, so for instance bank C1 should have little incentive to mimic directly the liquidity policies of a bank holding group that is based on another country and in this setting this seems to be a plausible assumption. So going back to the three main explanations that I gave you in the beginning of the presentation, so first of all within country banks have higher incentives to mimic their peers since they share the same lender of last resort. If one thinks about relative performance evaluation there's also plenty of evidence out there showing that firm select peers quite narrowly so for instance other firms in the same country and industry when setting RPE because the objective is to filter out common shocks to performance and also learning is also more likely to occur within countries where information for bank managers of small banks is more accessible and where banks share a similar regulatory and economic environment. So in terms of the other characteristics that I use to define peer groups it's a business model implicitly because I only have commercial banks in the sample as well as bank size. So I don't want to waste too much time on this because I try to be as least controversial as possible in terms of the liquidity mismatch indicators I use. So I basically use the three main measures that are out there in the literature. So the Berger and Baumol liquidity creation measure the liquidity mismatch index as well as a proxy for the net stable funding ratio. And again the sample everything is pretty standard here so I use both the cross country sample of banks operating in the OECD countries using bank level data from banks scope and an alternative sample more granular, more high frequency from the US that allows me crucially to capture liquidity created off the balance sheet that for instance in the US accounts for almost half of all liquidity created. Okay, so as promised let me go to the results. So first of all what this paper shows is that commercial banks follow the liquidity mismatch policies on their respective competitors when determining their own. The economic impact is large and consistent with the coordinated behavior where each bank constantly adjusts to each other's positions, okay. In terms of cross-session electrogeneity what I show in the paper is that these peer effects are concentrated in exalted riskier banks that have lower capital ratios, lower profit stability and lower distance to default or if you want higher default risk. And I also show in the paper that this collective risk-taking behavior is purely driven by liquidity created on the asset side of the balance sheet of course where of each lending is a key component. And finally in this first part of the results while the objective of this paper is not to tell you what is driving this collective risk-taking behavior is simply to demo, it's to be able to show you hopefully in a convincing manner that this behavior exists and the consequence of it but I still show that in my setting that small banks only follow other small banks and large banks only follow other large banks. What this means in practice is that learning or if you want free riding and information acquisition might not play a big role in this setting because as I mentioned in the beginning if this channel is at play we would observe small banks be making large banks but not vice versa. All right, so in the second part of the paper I show that strategic complementarity in banks with mismatch decision deteriorates the stability of the financial system. So first of all to analyze the direction in which these peer effects operate I first show that the response of individual banks to their peer choices is a symmetric. So MIMIC only occurs when the competitors are taking on more risk that really suggests that this behavior is indeed strategic. And then I show explicitly that these peer effects are associated with statistically and economically significant increases in default risk of individual institutions as well as systemic risk, okay? So I have 30 seconds left so let me conclude. So what do I do in this paper? I show that liquidity mismatch choices of competitors do matter for liquidity mismatch decisions of individual banks. I show that this effect is concentrated on the asset side of the balance sheet and this asymmetric. And then in terms of the consequences of all these I show that the strategic liquidity risk management decisions increase both individual banks default risk and systemic risk. And in my last 15 seconds I will just tell you what this means in terms of policy. So first of all the results this results clearly highlight the importance of regulating systemic liquidity risk from a macro-prudential perspective by for instance introducing a time-varying NSFR that operates in the contextical manner. This is nothing controversial the ESRV and the ECB already discussed this issue at length and I think this could be of value. And obviously and finally this is my last point because I run out of time the move from bailouts to credible bailings is clearly an important step to mitigate the incentives for collective risk-taking behavior even if and allow me the self-advertising the self-promotion here even if this is associated with short-term but limited negative effects to the real economy as I show in a follow-up paper this time a co-order paper exploiting in this case a quasi-natural experiment which corresponds to the failure and subsequent resolution through a veiling of a major bank in Portugal. This is all from me. Thank you. Thank you very much. Andres, thank you very much also for you exceeded the time because you were explaining the policy implications which is important in a house like this. If we have time at the end we may come back to some of them. The second you refer to self-promoting subsequent job because actually my doubt was that I couldn't see direct evidence for the second argument that let's proceed with the second winner. Guillaume, please, the floor is yours. So thank you, Ravier, and thanks a lot everyone for giving this price. So you just had a very clever paper. I'm just gonna have a very simple story but I think a very interesting story which is going to be the following. So as you know, one of the main reforms following the financial crisis is to make central clearing of derivatives mandatory in a world, as you know, central clearing is just essentially having CCP central clearing counterparties guaranteeing trades in markets for derivatives. The story I'm gonna tell you is a very interesting story which is the story of the very first derivatives clearing house in history and this CCP was created in France in 1882 in the market for coffee futures. And so what I just did is to go to some archives in Le Havre and in many other places and to try to understand why people at the time without any incentive given by regulators created the CCP's and what kind of effect it had in particular, what kind of real effects it had. And I will show you that it had very large impacts. In particular, it changed dramatically the geography of trade flows throughout Europe and after traders in coffee in Le Havre could hedge futures or could hedge inventories without counterparty risk, a lot more of the coffee trade throughout Europe was going through Le Havre and Le Havre became a much bigger harbor before, within a few years, central clearing could spread throughout Europe. Okay, so I'm just gonna tell you essentially the story in a second. So of course what you should have in mind here is all this region of the north of Europe, what the people in the geography called the Northern Range, you know with a lot of large harbors at the time, London, Liverpool, Le Havre, Antwerp, Rotterdam, Hamburg, all of these are large harbors at the time, it's actually the most active trade area worldwide at the time. Free trade is quite general, steam boats are replacing sailboats, so trade is booming, and the key feature of the coffee market which I'm studying, which is very important is the following, is that at the time a very large part of the coffee in the world is produced in one country which is Brazil, and so it all comes within a few weeks during the year. So you need dealers to hold large inventories of coffee for the full year and to cater slowly to consumption for one year, and so of course as any dealer in the market that has to hold big inventories, those inventories are exposed to some price risk and so there is some demand for hedging those inventories, okay. And so what had happened for centuries is that you have had forwards and then futures that people were using to hedge those inventories. And the key thing is that before these events, before CCP's, how people were trading those forwards or those futures? Typically there was very little collateral, so how could I trust Ravier or André if I enter a future with them? Well typically Ravier would be in a family which would have been in the coffee business for one century, so his name would be very well known in some harbor, and so I would know that most likely he's not gonna burst tomorrow because he has been around for many, many years, many decades, sometimes several centuries. And so reputation was a substitute for collateral, there was very little collateral. And so what happens, and which gave rise to central clearing is that in 1880, there is a very big crisis worldwide starting in the US and spreading through Europe and several of the main dealers in the US failed very almost simultaneously within a few weeks and so all this system based on reputation collapses. Suddenly I don't know if I can trust Ravier anymore because other big dealers have failed. And so people everywhere start to look for ways to restart trade and so traders in Le Havre come up with this institution which would become central clearing counterparties. And what is quite fascinating is that the way it worked at the time is almost exactly the way it works today. So before I had a bilateral contract with Ravier, now we have a CCP becoming the direct counterparty to me and to Ravier, this is exactly how it works today. More interestingly, how would the counterparty guarantee or ensure each of us against counterparty risk? It would collect margins, both initial margins and the variation margins exactly as they do today, okay? And if Ravier or me fail on paying a margin call, then the CCP would liquidate the position exactly as they do today. And one difference, if the CCP liquidating Ravier on my position creates a loss, all of the loss would be borne by the equity holders of the CCP. Today typically we would have more complex loss sharing mechanisms, the so-called default waterfall, okay? And so this CCP starts operating in 1882. What is very interesting is that it's a completely private initiative. It's really the traders and the dealers in Le Havre coming together at the Chamber of Commerce and coming up with this, there was zero regulatory incentive at the time. And so the key innovation is really this idea that there is one institution that becomes the direct counterparty to both traders of coffee futures. Okay, so what I could find is first, so I went to probably 20 different archives. So a lot of institutional data, minutes from a general assembly, a lot of futures market data, so all of the prices every day in this market, some information of the identity of the traders. And most importantly, what I do is that I go to the customs archives of eight or nine European countries. And for each, within each of them, I reconstruct trade flows in coffee and in many other commodities, harbor by harbor and between pairs of countries. Okay, and I have the following coverage of countries, Belgium, France, and so on. Okay, and so what I'm gonna do, I don't need to look at this. What I'm doing is extremely simple. So it's what, you know, econometricians call triple difference in differences. I'm simply comparing how much coffee goes to Le Havre with how much coffee goes to other harbors where there is no central clearing house. And within Le Havre, I'm going to compare coffee which benefits from this insulation against counter-party risk with other coffee like sugar, tea, vanilla, whatever, tobacco, where you don't have any of this central clearing, okay? And so the control group which I'm gonna use to compare coffee with is what you see written at the time as colonial commodity. So these are, as I said, sugar, cocoa, and so on. And I will look at trade flows six years before and six years after around this event, okay? So I start at the within France and I have the data from 22 different customs in France. So most of them are harbors, but not all of them are harbors where coffee can enter France. And I'm gonna compare, well, how does that change? What is the share of coffee entering France within Le Havre before and after 1882 relative to all other harbors and to all other commodities? And so the basic number to look at is 0.111 top left which is essentially telling you after central clearing comes in place, there is an increase by 11 percentage points in the share of all the coffee imports of France that comes through Le Havre relative to other harbors. So France imports not more coffee in volume, but a bigger share of it comes through the harbors where now dealers can hedge their inventories without counter-party risk. And what is interesting is that it's not that suddenly people in Le Havre start to drink more coffee, not at all. I actually have statistics on stocks, so they store more coffees and they actually re-export more of these coffees. So it's not that it's really that they become a hub for holding inventories because now they can hedge them without inventory risk. It's not that they consume more coffee. Then I do, what is interesting is that actually they also introduce it for cotton. So you also have the same results basically for cotton. And then I do the same regression, Europe wide. So I'm gonna go there. So I have, how many? Seven regressions for seven different European countries. And what I'm looking at is exactly the same thing. What is the share of their coffee imports in Belgium, Germany, and so on that comes from France relative to other countries? And this share again increases significantly after central clearing is in place in France. So this shows us that we really have real effects in terms of re-shuffling completely, trade flows throughout Europe. I also have a lot of narrative evidence that this was going on. And what is perhaps more interesting is that of course within a few years other countries realized that Le Havre was really taking over in terms of the coffee trade. And so they wanted to imitate and to also introduce central clearing. And within a period of 10 years we have about 10 other harbors in Europe that also introduce central clearing including Hamburg, which was the main competitor of Le Havre at the time, but many others in other commodities. And actually it's a very interesting historical period because out of those CCPs not all of them succeeded. Some failed really quickly. So there is a lot of interesting events in terms of the design of the CCP which differed across all of the CCPs at the time. So that's about it. So then I have a number of alternative tests which I'm not gonna show but which I'm gonna discuss very briefly for the time, which is to say, well, what is really the mechanism? What really did the CCP do? And so if I come back to my example with Ravier, the problem is that before, you know, there is reputation, Ravier has reputation, André has no reputation. So I want to trade with those who have the best reputation, so I want to trade with Ravier. But then there is this crisis and I don't really know anymore who is a good quality counterparty, who is a bad quality counterparty. So there is what we could call in economics a pulling equilibrium. I cannot distinguish anymore who is good and who is bad. And so the argument is what the CCP really does is by requiring members to post high margins and the margin requirements are actually quite high. It will allow us to separate the types, to separate again the good types from the bad types. And only those who will have sufficient financial resources to pay those margins will be able to enter the CCP. And from this situation where I cannot distinguish anymore who is good and who is bad, I will be able to separate again only the one with sufficiently many resources will be able to join the CCP and to trade in the CCP. And there is actually a lot of evidence that lower quality counterparties are complaining a lot about those margins because they were finding very hard to continue trading in the market afterwards. And the last thing I want to show is that another impact which was probably completely unintended is that it made markets a lot more competitive. Why is that? Well, before if I trade only based on reputation establishing the reputation of having been in the coffee business for one century takes a lot of time. It takes maybe one century to build. And of course, if Andre has never been in the coffee business he will not really be able to enter and to build this reputation. Even if he is very good. Now that he just needs to post sufficient margin he doesn't need any pre-existing reputation to be able to enter the market. He just needs sufficient financial resources to pay the margin cost. And so what we see afterwards is a quite significant entry into the market because of new traders suggesting that now that trading becomes more anonymous in a sense just based on margins. Entry is actually a lot easier in this market. So just to conclude a bit ahead of time because we are late as I understand. So what I show is essentially that central clearing reshaped trade flows Europe wide at the time. So it had very significant real effect and one of the main mechanism is because it reduced adverse selection and mitigated some information problems. It made it easier to separate again good and bad types. That said, most of you are policy makers. I don't think there are any implications unfortunately of this for the assessment of current reforms. The main reason is that clearing at the time was voluntary, now it's mandatory. And so of course the incentive structure may be very different. Both are of course very interesting to study but I don't want to push it too far. I think it's interesting enough like this. Thank you very much. Thank you very much, John. So I think your focus on the coffee market is sort of a perfect introduction for our coffee break. And we were taking a little bit of extra time to start. I'm tempted to, well, let the audience judge on the originality grounds but I was mentioning that yes, if any of you has been sleepy before our coffee. In the first paper there is this exploitation of the overlap of ownership of group banks and then introduction different domestic markets to create this original instrumental variable for identification of otherwise very tricky to identify network effects that I think was very clear in the presentation. In Guillaume's paper there is this huge investment in historical sources and this originality bonus in trying to look at the CCP's of the old days and it's quite amazing to discover the essence of the economics behind is to be so similar to the ones that somehow you started mentioning the crisis of 1880 and they move into compulsory clearing. Central clearing was a reaction to another financial crisis. When the repo and other derivatives market went into crisis we reacted somehow like this little market in France well not so little market in France reacted at much more innovative historical times. If you any of you had something. There was a debate in the French parliament on mandatory clearing in 1900. There was a crisis in the wool market and there is a big report written by the French parliament a century ago, about 120 years ago, discussing mandatory central clearing. Interesting, so it sounds like... It didn't go through at the time but there is really a big... Because of course what is missing from your particular example is the sort of system-wide motivation. I guess the creators of the CCP were essentially self-interested and with great success. When we in Europe introduced mandatory clearing the feeling was that the system was gonna work better. Why the agents were not converging to the private solution by themselves? The answer is the externality. Spill overs and the first paper topic. Unfortunately no more time for the discussion. So thanks again, thank you very much. You are established a very high bar for the next years. Thank you.