 Morning. The session now is about derivatives and systemic risk, which is clearly one of the main general concerns of any systemic risk manager. And the papers we have today, I think, are a good review of the main issues about the area and the research questions. Now, after the crisis, there have been sort of two initiatives in terms of derivatives. The first one was to concentrate derivatives clearing within CCPs. And that was a move that was inspired by the fact that when the large broker dealers had balanced books of derivative transactions within their balance sheets, they were running a business effectively running a central counterparty business. However, they were running a central counterparty business alongside a number of other businesses, some of which were quite risky. And therefore, the central counterparty business was vulnerable to the riskiness of the other businesses that the large broker dealers were running. So the idea of the reform was to concentrate the running of the central counterparty business within specialized institutions, which were set up only for that purpose and certainly not affected by other kinds of risks in the financial system. And that was, I think, in my opinion, a certainly very wise move, which, however, raised some questions, which I'm sure that Mr. McLaughlin will discuss about, and he was mentioned also before, the systemic role of central counterparties in this new world. And that's certainly a question that we will have to talk about. The other fundamental pillar of the reforms after the crisis was the creation of the trade repositories. Why is that a fundamental pillar? It is because one feature of the financial system that the crisis has brought in broad light is the fact that authorities and systemic risk managers, which existed before the creation of the SRB and there were central banks before the SRB, turned out to be at a huge informational disadvantage relative to actors in the financial system and the financial system as a whole. And that is not what a systemic risk manager should be. The role of a systemic risk manager is to have an informational advantage relative to everybody out there playing in the financial market. And so the trade repositories were created. The papers that the other papers that we will hear about today, in part, use data coming from those trade repositories, but in general, provide analysis of the structure and the working of derivative markets, of different derivative markets. In one case, foreign exchange, in the other case, credit. Now, they are, I think, very good examples of the state of knowledge of, let's say, a systemic risk manager, because this work has been done within the SRB, of the state of knowledge of the derivatives markets, which I would deem is still pretty preliminary at this point. We are able to map things in a very important way, but we still have a number of questions that need to be addressed and a number of really important investments both in research and in systems that we need to carry out, because in order to handle the information that we gain from the very significant transactions that every day occur in the financial systems in derivatives, we just need to set up a lot of know-how and a lot of computers to deal with that know-how and so on. In particular, one of the problems that are out there is that despite this important reform of trade repositories, trade repository data is now far from being easily usable by systemic risk managers. So we need to really fill a gap that has not been filled. And I believe that some of the presentations today may touch upon this issue. I stop here with my initial remarks and I give the floor to Professor Hao with his own presentation. Thank you very much. I just want to say one thing about your presentation, which is important for everybody to pay attention to. It not only talks about the one derivatives market from the perspective of risk, but also talks about it from the perspective of efficiency. And that's a very important sort of line to take, which I think people should take notice of. Good. Thank you, Alberto, for this introduction. So the title is Stop a Go, the Reform Agenda and OTC Derivative Markets. And my role will be to speak about what the new data that EMEA and European regulation provides us and what light this new data can share on the reform agenda in derivative markets. The results are somewhat preliminary, but I think already this preliminary data has some, I think, very pertinent insights that I want to share with you. The work is joined with Peter Hofmann and Sam Langfield and the two of them have really done a tremendous and terrific work on making this data accessible, which is no small task given the amount and complexity of the issue. So let me hop back with a little reminder of where this reform agenda in derivative markets came from. It's certainly true that derivative markets were not the origin of the 2008, 2009 crisis. That was structured products. But at the same time, it's fair to say that derivative markets played a very important role in the transmission of the crisis. People talk a lot about the Lehman disaster in September 2008, but the real elephant in the room was AIG. And if AIG hadn't been saved, the derivative exposure of virtually all the big banks would have brought them down. We would have had to nationalize every single large bank in the global financial system. And AIG kind of preempted this. And in a way, the AIG nationalization was forced upon by derivative exposure. So it's the elephant in the room. It's a very important issue, derivative reform, if you talk about systematic risk. So the commitment was taken in Pittsburgh in 2009 to actually progress on derivative reforms in three different dimensions. The first is the trade reporting that Alberto mentioned, that the authorities should be put in a position where they can actually make informed database judgments on the systematic risk that emanuates from the derivative market. The second was the clearing mandate to move a large proportion of the clearable trades to clearing houses. And the third is to change the market structure so that these derivative contracts could be traded more efficiently and would be then cheaper also to resolve and to aid with the too big to fail problem and the resuppability of institutions. And maybe to make also the market more efficient. So where we are in this reform agenda of Pittsburgh seven years later? Well, the results are very, very mixed, as you can see from this traffic light. We can distinguish reform and interest derivative, credit derivatives, FX derivatives, and we can put on the three different domains, the reporting obligations. There we have made the biggest progress. Central clearing, the results are very mixed in an exchange rate trading. I think we haven't really made much progress, or at least in Europe, not as much as in the US. In this talk, I will actually focus on a particular sector, the FX derivative markets, where the progress has been least pronounced, where we are essentially on a red light the latest recommendations that emanate from the Paris-based ESMA Board actually suggest that we should give up altogether on pre-trade reporting and so I think going to exchange rate trading is a very far away, and so we have a very important implementation shortfall with respect to the Pittsburgh agenda. And so my question is now based on this new EMEA data, so we do in fact have the reporting, now what does this new data, what light can we shed on the costs, the economic costs of this implementation shortfall? And here we can distinguish between two types of costs. The first cost is continuing to running with the current OTC market structures. There are costs to this. There's some people that pay conservatively more. We might have higher costs to risk management. These are costs that are borne possibly by the real economy. But these issues of market efficiency are probably less relevant from the systematic kinds of risk concerns that we are talking about here. So there's a second dimension to the question of implementation shortfall. What does the implementation shortfall really mean for how we think about systemic risk? And here the question is given that we stick with this OTC structure, what does it mean for the global risk allocation? Are the risks really borne by the institutions that should be at them? Generally we view important macroeconomic risks should be borne by the best capitalized institutions. Those are most capable of bearing these risks and what kind of consequences does the structure have for this macroeconomic risk allocation? The second issue that has to do with resolution and to make to fail, obviously, we have a lot of non-standardized trades that are on the books of these large banks. You kind of give up sort of implicitly on to big to fail and their costs to that as well. Del Duffy always makes this point in his books that the resolvability of institutions and the nature of the contracts are obviously highly interrelated if you have a lot of OTC contracts that augments resolution costs tremendously and kind of forces public authorities into bailouts. Let me talk about both issues. The first one is the more transactional perspective. The second one is sort of the more global macroefficiency perspective on this implementation shortfall in derivative markets. The good news is that we are now in a position, I think to take a more informed judgment on the costs of this implementation shortfall. Let me talk quickly about the data. So the data comes from these trade repositories that are mandated by the new EMEA standards. It's very hard to work with this data. It takes a lot of technical capacity. It's big data. We have 20 billion observations of 100 variables. You can imagine the problems there. But I think the European Systematic Risk Board did a tremendous effort here thanks to people like Sam and Peter and others that worked on making this data accessible and exploitable for scientific analysis. So in my talk, I want to focus here on a relatively short sample of a month but still a tenth of thousands of observations that spends FX derivatives. And we are actually looking at only at one cross here, the dollar euro rate, the most liquid rate in a way, for a one month period that involves roughly 170 banks and roughly 3000 firms. We are focusing on the transactions between banks and so-called clients. These are corporates in our sample. So the whole market, as you know, the OTC market has a core periphery structure. We have in the center of the trading network of the bilateral trading network 16 big banks and then in the periphery, more to the periphery other banks and at the very periphery, the clients. And in this market where you don't have a central trading platform, their considerable search costs, which raises the opportunity of having informational rents because obviously someone who is in the center of the network trades a lot, gets much more information as opposed to just skimming off this information from a trading scheme stream where everyone is at an equal footing. So the OTC structure per se creates a potential for considerable information rents. So what are the main findings here? What we find in the data is from a transactional point of view, three things. First of all, we find a high dispersion of spreads across clients. I mean, these are really enormous differences in the fairness of the market to different market participants. And secondly, particularly unsophisticated clients, those that trade relatively little and that are dealing only with one major dealer, and that are small, these are the ones that pay enormous spreads in this market. And maybe the more controversial point is the last one. We have average spreads that are overall high, I would say. So the first two are pretty undisputable. The second one is more a matter of appreciation. How do you compare the transaction costs to other markets? Now on the issue of overall macroeconomic efficiency, we find enormous concentration of trading revenue in a very small number of banks, dealer banks. So 85% of the profits or revenue go to a relatively small number of banks. Moreover, this market is inherently, I guess monopolistic in the sense that the more you trade, the more profitable you become per trade. So I'll show you some evidence on this. And finally, I think that's probably the most pertinent remark for the issue of macroeconomic risk allocation. It's optimality is these dealer banks are pretty much all SIFIs and they're all highly leveraged institutions. From the macroeconomic distributional point of risk, if you believe that at least a proportion of the risk is not passed on to ultimate clients, while we discovered this in structured products where we believed all the risk was passed on to other market makers that could actually take the risk, we don't really know how much of this risk is passed on, but to the extent that some of the counterparty risk, in particular, unclear to risk, stays with these banks, we have kind of the worst possible allocation of risk from a macroeconomic stability point of view. I come to this point at the end. So let me skip this, spread costs come to the dispersion of spreads. So it's a very simple analytical framework, they're not the deep conceptual issues like with monetary policy, you just take the spread that everyone pays and you request it on a fixed effect for each client when you get the distribution of client costs and you have this huge right tail that some clients here on the right end of the tail pay very high spreads on average. By contrast, if you have fixed effects for banks, there's much less dispersion of bank revenue across banks. Now, if you look at more detail, what are these clients that actually pay huge spreads in this market? It correlates highly and convexly with the variable we can call inexperience or lack of client sophistication and that's basically a composite of two things, how small you are and how less diversified your interactions are with dealers. So basically clients that don't shop around that in principle always deal with the same dealer, get a very, very bad deal in this market. And there has been a lot of talk in OTC market that we need them because of customization is important. If customization is important, we would expect it to particularly important for clients that have a low credit rating because if you are already a fragile corporate, then tailoring your derivative to your particular maturity that you need matters a lot. But surprisingly, we see none of it. We don't see that client risk is an important determinant of transaction costs, which is very surprising giving this tailoring argument. Tailoring, I mean, in the maturity dimension. We can probably do a lot more to discard this argument of tailoring, which we haven't done yet, but client risk not being related to trading costs is a very boring issue and sort of makes you think where there's any substance to the tailoring argument whatsoever in this market. Now, coming to client sophistication, we can do quantile regression. We can show more precisely how trading costs actually increases with your size, namely inversely, the smaller you are, the bigger the more you pay and the concentration of your counterparty. So the more you deal with one counterparty with a single dealer, the more you pay. Of course, that's endogenous. It's sort of indicative of trading costs, but those of search costs, those that have high search costs, they get very disadvantaged in this market. Now, let me come to the third issue of the transactional aspect. How big are the overall costs to this? That's a bit more touchy issue. Depends a little bit how you normalize trading costs. If you take the average spread of maybe 25 basis points here in this market, how does this compare to other markets? Let's say the equity market where spreads and absolute terms are maybe similar or somewhat smaller. Well, if you think that the notional that is actually transferred is only 5% of the total value transferred, you would conclude that this market is by a factor of 25 bigger, the trading cost 25 times bigger than in equity trading. Well, that might not be the right normalization if you want to normalize with carry risk, for example. Still, you get relatively excessive costs, carry risk in FX is smaller than in equity, and still you have higher absolute costs here. So it depends on the normalization. There is a study here by the Deutsche Börse that we put in here that roughly eight estimates that the trading cost to this OTC structure for a client might be up to eight times what it could be in OTC. I don't wanna push the factor of eight or any particular factor, but I think the academic research shows that moving to OTC structures can substantially decrease trading costs. Already, pre-trade transparencies does a lot. We have this experience from the municipal bond market and the experience in corporate bonds where trace and other pre-trade transparency mechanisms led to a substantial decrease in spreads. So I think from a scientific point of view and our experience in other markets suggested, this is a relatively expensive and inefficient market for clients. Now, the total, we can also calculate what's the total cost of revenue that is generated here for the dealer banks. We come up with a 3.2 billion number. Remind that this is sort of concentrated on export firms. So these are the ones that actually have to pay these costs. And so it's kind of like a substantial trade tax on those firms that have to hedge their risk. It might even impede some type of corporate risk hedging if you believe that your costs are exorbitant, you might not do it. So maybe liquidity is actually lower in this market as it should be from the demand side because of the cost. But we have no concrete evidence on this yet. I think it is worthwhile looking more into the real effects of this OTC structure, which we haven't done yet. On the other hand, on the dealer bank side, revenue is highly concentrated. So the distribution shows very sort of right tilted and given that this is in one sector that makes these banks bigger and certainly reinforces the too big to fail problem stare. Here's the picture that plots trading revenue per trade per unit of trade as a function of how big you are. And you see a convex line here relative to the 45 degree line. So this market, the bigger you are, the more profitable you are in a way per unit of trading. So there's a natural kind of monopolistic or oligopolistic structure to this market from an IO point of view. And we wouldn't believe that this market therefore would evolve towards less concentration because the economies of scale here, this is just the revenue side. There's also a cost side, which probably reinforces this further. If you are bigger, you might have cost savings, but already from the revenue side, you have to have economies of scale in this market. Okay, so let me come to the last point of the issue of, how do we look at this market from the systemic point of view? What is the vulnerability? I think yesterday it was said we have to focus, I mean, John Cunliffe emphasized this notion of resilience, resilience is important. Do we have a resilient structure here? Well, think about the overlap. We have here the, I think these are just 13 global systematically important banks, CIFIs, and they share 75% of the market share in this market. And these are also the same institution that have the highest leverage. So I guess from a systemic point of view, we would say this is sort of the powder kek of the financial system. So if one of these sort of is hit, I think we have an explosion. It's like the Royal Navy in World War I going out to see into the battle with the door to the ammunition chamber open. And first hit, the whole thing blows up. So I think from a distributional point of view, this is the worst possible risk allocation that we can have. One interesting question is of course, the issue of contestability. Why isn't it that other people come in and contest this market? Of course, given that you have such a high concentration of profit, this industry is probably also very organized in terms of lobbying and defending these rents. And I think there's also considerable coordination problems on the part of the buy side. These are dispersed group of 3,000 corporate firms. They would have to coordinate to all come to the market to a new exchange at the same point. So it has been tried apparently repeatedly by some exchanges to actually dislodge this market. But given the structure of it, it might be very hard without I would call a regulatory notch that you kind of employ some force in transferring this to a more efficient market structure. So, good. Of course, the important issues that we don't really know and that goes to the presentation that will follow by Thomas. We don't have a clear view yet and this research doesn't do it. How much compression is going on there? How much transfer of the risk to ultimate institution that should carry these risks is going on? And so these questions are very fundamental for judging how explosive the situation actually is. Okay, that's a bit outside the scope of what we did so far. So let me come to a conclusion here. What does this new AMIA data tell us about this market? Well, it's a very costly market. It's a costly market for the expert sector in particular for unsophisticated firms. So I guess from a efficiency transactional point of view, there's a point to be made here for reforming this market. I think so far most of the academic research that has been done on similar reforms, OTC market reform in municipal bonds and bond markets show that substantial cost reductions available here for client. It's also clear that it entrenched too big to fail by preventing non-banks from entering. In equity, for example, a lot of the intermediation is done by the retail sector today. I mean, if you look at equities, that's a very stable and very reliable source of liquidity, which doesn't play a role at all. And if we would transfer the OTC structure into an exchange rate structure, that's what we would get. We would get a lot of other participants into the market and so get at a very much better risk distribution of the among intermediaries this way. Okay, finally, as I said, we have a very undesirable macroeconomic allocation here of risk in the very institutions that are least capable of taking over this risk. I think that's pretty much what I wanted to say. So, last picture, yeah, so I think there's also, I mean, the question then comes, of course, why didn't, why did we switch to red light here? I mean, that's a very good, maybe someone from Esma or the European Commission can answer this question, why did they stop derivative reform in this market, which is such an obvious candidate for moving to exchange rate market. That's a very good question. I think something that the systematic risk board should probably take up and I can only speculate since I'm not very familiar with the particular process there. I would, however, stress that it poses an important credibility issue. I think if we plow the big fruits here from derivative reform and the substantial improvement of systemic stability that we get out of this by not doing this, I think we have a considerable loss of credibility, if we stay with store. So, in a way, my talk here is also a plea for reconsidering the Esma and Commission decision on putting a red light on derivative reform here, because otherwise we face tremendous costs of credibility here and that's captured by this cartoon. Let me finish here, thanks. It seems to me that the natural answer to your question about the FX market is that FX forwards are not very different from an FX spot market. It's just that the settlement is delayed and people will say that the FX market has been working like this OTC for decades, more than decades, very well. So, this is the standard industry response, but of course, you're raising important question that need to be thought about. Now, in terms of the concentration, again, it's a fact of life of derivatives dealership that these markets are always concentrated. And again, it's an interesting regularity we need to understand a bit more, maybe from an industrial organization perspective, why is that so? But it would be if we move to another market structure, that's sort of the point. It probably wouldn't be, and the question is how to do it, but we now move to a different set of questions. Mr. Mellachling, on CCP. Whatever you prefer, when I sit down, sit down, it's fine. We need the hope, come on, there are more issues here. Yeah, yeah, yeah, yeah, yeah. This is some hope here, I'm going to make this. Thank you. I'm going to try and talk a little bit about probably where the biggest culprit in the debate. LCH is probably the biggest global derivative CCP out there right now. And there are just a few statistics. We have about 90% of the interest rate swaps market, which uncompressed is over $500 trillion in-house. Compressed, it's in the low $200s, and obviously the business keeps coming in. We are also in the repo market in Europe. We have about two thirds of cleared repos. We are also in the FX market, FX forwards, non-deliverable forwards, although that is not mandated as yet to be cleared. We're also in the CVS market. We have probably about 20% and growing in that market. And then we're in, they were all pure derivatives, if you view repo as a derivative. We are also in the more traditional, physically settled markets linked to exchanges like equities and commodities and bonds, yeah. Interest rate, futures, things like that. So we have a mix of mostly derivatives, I would say, but also the more traditional, physically settled products. And just a few words before I launch into the evolution of the markets, but traditionally the CCP is there to act as a shock absorber against the impact of a member default. So we, in the CCP, we collect margin to ensure that once a defaulted member, once a member defaults, we can step in, we have enough margin to ensure that the VM obligations to the non-defaulted members are met. Then we take the defaulted members portfolio. We actually assume ownership of that portfolio and we close out that portfolio in the market. If it's a traditional exchange traded instrument, we use brokers to access the liquidity on the exchange. If it's a more derivatives-like position, we have to auction that off and use the membership to spread out the positions in that auction, or we can go to people outside the membership, which tend to be large banks, to more buy-side institutions to try and move those portfolios. The process follows a very, very strict protocol that's pre-agreed with the members upfront. This is what we're going to do. This is the steps we're going to take. The oversight of that is obviously the CCP is looking at it, but also members give up traders to sit on what's called the default management group to help administer the closing out of the defaulted portfolio. This is the theory anyway. It serves to limit the potential contagion in the default event. That's how a CCP traditionally works. That's how we've traditionally worked. I will talk a little bit more about what needs to happen now because that's, I would say, under strain right now from various things that are happening. The key systemic risk tools that we use to try and, as I said, the numbers are quite staggering. So we have to be quite careful about, first of all, who do we let into the CCP? As a member, an active member, we have to have very sharp minimum credit standards. So for example, we do not allow subprime members to join the CCP. You cannot join the CCP if you're a subprime equivalent. That's not universal, but that's our standard because it's hard to justify the contagion of the CCP pool if you like, the mutualized pot with a subprime member. Obviously, there's the cover two standard, which puts a floor if you like on the kind of financial resources we must hold to ensure that there's enough financial resources to withstand the instantaneous default of the largest two members. But if you think about the trading mandates, the clearing mandates that have come into place, it's quite hard to manage the concentration risk. All we can do, because we can't say no, they have to clear, all we can do is continue to charge more and more financial resources, exponentially increasing the financial resources until such a time as the trading, if we feel it's too big, it's starting to be discouraged. So that's really difficult to do. We have a process for doing it, and the way we do it is we demand a fairness principle that no one member can take up more than half of the default fund because that's the mutualized pot, so one member should not overuse that if you like or take advantage of the other members. And the extent to which they do go over that 50%, we charge additional margin to bring that total resources required for the default to cover the default of the largest person, the share of the mutual. It was my best talk to everybody. Each individual member is required to stay below that limit if you like, 50% of the default fund. The other thing we have to look very carefully at is prosyclicality, which I know is a huge issue for systemic risk management. Here we have our own standard, which we felt that the ESMA constraints, the ESMA standards were not strong enough. And what we've done is we've looked at the credit cycle, so to speak, probably going back 10 years. And we have a standard in place that demands that if we were to evolve the past into the future over the next 10 or so years, there should not be any one day. The model should be built, so there would be no one day in that future where the margins for a member would jump more than 25% in the over the holding period. That's a very interesting standard. And we run our models to that standard. The only exceptions that we have noticed in that standard are due to central bank or government actions, which is a very interesting finding from our perspective where the source of systemic risk comes from, okay? The last part is on actual default management, we of course are managing to a cover two standard. What happens, what's the safeguard if you like, or what's the fire escape if more than two people default? Well, then we have assessment powers built into our rule books, where we can actually call a number of new assessments from members if we need those resources. We have variation margin gains haircutting tools, and we have, and I put partial in brackets here, contract tear ups. The full partial contract tear up is a very, very difficult issue for us because we would love to be able to, in derivatives markets anyway, it's unclear exactly how you objectively upfront specified the trades that you would tear up. It's very, very difficult, especially in a derivatives market where you have so many trades offsetting so many others, it's very hard to identify exactly those trades you would tear up. That's much easier to do in more traditional physically settled markets where the culprit's very easy to identify. Now, one of the big topics right now for discussion is these powers have obviously systemic consequences. Tearing up a member has all sorts of knock-on effects, or could have knock-on effects, and therefore we have to be very, very careful about exactly under what circumstances we do this, and this is part of a resolution and recovery workstream that's going on right now, run by the FSB to really nail this down. So that's sort of a, that's where we are today and kind of what we're doing. I think things are moving on rapidly because as I mentioned, the size of the numbers are quite, quite large. The, also, there's a couple of things that have merged as new developments here. The first one is the capacity of banks to provide clearing of members to provide clearing services outside banks is declining. So anybody who wants to see this can look at the number of FCMs there have been and how that population has shrank dramatically in the last couple of years. Due to things like SLR constraints, regulatory constraints, capital constraints, et cetera, are making it more and more unprofitable. And that means that third-party users are outside the core banking sector. Well, they don't right now have access to the CCP. They're finding it very, very difficult to get clearing services from the core banking structure. So that's a very, very interesting development because now we're under enormous pressure to try and extend our membership beyond core banks, if you like, to insurance companies, pension funds, investment managers, all these kind of institutions. That creates enormous headaches for us because obviously they don't have the same link into the financial banking system and the European system, for example, that banks would have. So it's a bit of a headache and we haven't figured it out yet. But the other surprising development that's happening is that central banks are now looking and have actually joined the CCP as a fully fledged member. Now, if you think of the implications of that, a central bank could be borrowing enormous sums of money or lending enormous sums of money. That's the kind of activity we're seeing today and trying to parse out what the risks are that they're bringing into the clearinghouse. And if you like polluting the default fund, the mutualized fund is quite interesting. Some of them, huge debate about whether they should be paying into that mutualized fund or not. So this is not easy. Probably the biggest challenge, I think, that I face in the CCP is, and this was unseen at the time, obviously that in Pittsburgh, when it was all put into place, what exactly would actually grow up at the end state when you've done all this and then you see what has grown up? Well, I think we're the poster child for what's grown up and so we begin to see constraints that we had no idea that were out there. And the one I really feel for is the, it creates enormous problems for me is the fact that we have to take margins off our members, obviously, because that's how we protect ourselves, but those margins have to be stored somewhere. And under the conflicting regulations by various different bodies, they're all in conflict because on the one hand, and they all make very good sense on an individual basis, but when you put them together, they don't. So for example, Amir says that no more than 5% on average of your margins should be invested unsecured, and I totally see that, that's a really good policy. On the other hand, so you're not allowed to use money market funds either. Okay, so what do you do then? Well, you would say would put it in the central bank. That'd be a good place if you had access to the central bank for every currency that you're dealing in, but we don't. We have 17 currencies globally, at least 70, and we have central bank access for deposits in some of these relevant currencies, but not in others. So then what that forces you to do in a stress event is that you have an enormous inflow of variation margin, sorry, VM calls coming into you, full of cash, and you have nowhere to put the cash. So you're kind of defaulting, if you like, and the fire escape is the only thing left is to leave it with a commercial bank, and that's certainly not a good place to be in a stress event. These days. Yeah, so this is one of the biggest challenges I see where each individual regulation makes a lot of sense, but the overall consequence of that is not making sense at all. Other things that we're seeing, I mentioned that traditional focus clearing members totally changing. The focus, if you like, on the old mission of the clearing house for managing clearing member defaults is out of date, totally out of date, because if you think about the structure that's been set up is we have a rule book and rule book powers which totally allocate the losses from a clearing member default. They totally allocate it, although you might be worried about the systemic risk powers which we can get to later, but they totally allocate the losses due to a clearing member default, but there can be other losses, many, many other losses, and there's a paper that I wrote on this a couple of months ago that came out which there are probably eight or nine categories of losses of which this is one, and for the other categories of losses, it's not so clear what you would do because many of them cannot be allocated back to members, in which case the only thing you can try and do is absorb through the CCP capital, absorb the losses that way. So each one of those losses have to be addressed individually and you, if you like, rules putting in place for how you would actually either allocate it back to the members or absorb it or deal with it through some kind of new rule book power. And there are many examples and this is an ongoing debate, probably the biggest debate right now going on in the resolution and recovery work stream for the FSB. One consequence of this is that these losses that cannot be covered by member margins and traditional rule books, they tend to, a lot of them tend to arise in operational risk kind of events. And hence the operational risk discipline, which was only a few, a paragraph or so in a mirror, is greatly expanded now. We've had to work very, very hard to find out, well, what are all the kind of things that can happen, the operational risk events that can happen at the CCP and what controls do we have in place? That's a very big issue for us. The other really big issue is that the key CCP relationships in the financial ecosystem are not well understood. I just gave you an example of the interaction with the repo markets. Very, very good sense to control the repo markets from one aspect, shadow banks, et cetera. But it's a very unintended consequence of giving the CCP a difficulty in actually storing the margins in a safe manner. So you have two conflicting things going on there. As I said, we have interactions with central banks, which are becoming very, very complex to say the least. And we're not envisioned at the time at the Pittsburgh Summit. There are other people looking to join as well. There's a whole list of them. Obviously hedge funds are at one extreme, central banks that are another, but there's a whole spectrum in between. I would say that the other interesting thing about central banks is that when an action takes place, for example, quantitative easing, announcement of quantitative easing, that has an enormous impact on the CCP because momentarily we're unbalanced. Our margins suddenly are completely off. And so if a member was to default at exactly that time, we would not have enough margins. So there's interesting, if you like, and central bankers are becoming aware of this obviously, but I would say more slowly rather than more quickly. And I would just say one more thing is that there's a huge difference between asking for a deposit account at a central bank for the reasons I laid out, versus asking for a repo facility at a central bank. And that's not always understood. In the public mind, it's been equated with a bailout, even if you just need the deposit facility. And that's absolutely no credit risk to the CCP there. And there's a very good reason for asking for that. And then the CCP recovery and resolution just to bring it to an end. Some of the big open issues are who is the resolution authority for a CCP? I only found out just before Christmas that in the US it was the FDIC. That was a surprise to a lot of people. Well, it was a surprise. There was a room full of banks and CCP, we were shocked because they hadn't been in the discussion up until that point. And the other one is what is the relationship to the CCP regulator? Because in some cases, they're the same. The regulator and the resolution authority are the same. In other cases, they're not. In the US again? Yeah, not in the US. And the thing about that is the concern across when you cross borders internationally, it's quite a problem. Because we are systemically important in many, many countries. Well, at least five. Not in one where you have a domestic view if you like and there are transactions taking place on a venue in that country. You can have regulations in the infrastructure and the ecosystem in that country to look after that. But when you're systemically important in several countries, then it becomes a question of, well, exactly how does this work? So I won't go into that just to say that it's creating a bit of a headache. And the final point is, when should the CCP be put into resolution? That is not at all clear. If it happens too early, if it's mandated to happen too early, you might skew the incentives and not drive people to try and explore all the recovery avenues that they can before you put them in resolution. So it can be a little bit strange. So that's probably the biggest open question that we have right now. And then finally, I think there's a role for the ESRB in all this because this issue I raised about everybody in their own little silo acting to legislate, legislation that makes sense. But overall, when you put them together, they have very, very strange consequences. That I think is a role for a systemic body like the ESRB to try and ensure that we're all singing off the same hymn sheet. One thing I would say though, when activities do take place to try and address these issues, that we remember that the CCP itself is, right now it's a private entity because a taxpayer-funded solution is off the table. So if it's a private entity, it has to make it cost of equity. And it's not a bottomless pit. So more and more drives to increase capital will degrade that return on equity. And eventually, private sector funding won't flow into the CCP unless the cost of clearing increases dramatically. So I'll just leave you with that thought. Thank you very much. Just a clarifying question. You mentioned an increased reluctance of banks to perform a clearing member role vis-à-vis clients. But isn't there an issue, a concern on their part of being disintermediated if you allow direct access by other actors to the CCP? Have you not observed that? Absolutely, right. Yeah, and there, obviously, banks have to defend that, the role of intermediary. But on the other hand, the buy side are finding increasingly difficult to actually complete these transactions. So there's a trade-off. It's okay. All right, thank you very much. I appreciate that. We move to Tuomas now. Okay, ladies and gentlemen, I'm Tuomas Pelden, the deputy head of the ESRB Secretariat. Very delighted to speak to this distinguished audience. Let me talk about our ongoing work on analyzing credit derivatives markets, the flow of risk, notional access, and portfolio compression. Let me thank here my colleagues, particularly Marko Derrick, Tariq Rognia, and Sam Langfield, who have helped me to prepare this presentation. And finally, as a disclaimer, I'm talking here about my own views and not necessarily those of the ESRB and its member institutions. We already heard some facts about the role of the global financial crisis of revealing the size and opacity of the OTC derivatives markets, and also the G20 reforms, the Bitsberg Summit, that basically aimed at reforming the OTC derivatives markets, making them safer and more transparent. What we learned also, that we have very large notional amounts in place, depending on the different statistics you use, you get a bit different numbers, but here we use the BIS 2015 numbers, that the interest rate derivatives are roughly 384 trillion, the CDS-12 and so forth. What we also know is that there is quite some heterogeneity, and in some derivative cases, also complexity of the instruments. But the key point that Harad also already referred is that there is a very large fraction of intra-financial exposures. So there has been considerable efforts at the EU to increase transparency and order in the derivatives markets. The EMI regulation is a founding building block in this respect, and here at the ESRP, what we have tried to do is to exploit the enhanced reporting of derivatives transactions to the trade repositories. So the EMI requires all EU counterparties to report their derivative exposures to the trade repositories. We have a double reporting, and we include all derivatives, so also the extreme-straight derivatives are reported. And what our work at the ESRP Secretariat has been focused on is to basically develop a credible data infrastructure, also to work with our colleagues at the European Commission and ESMA to contribute to improve the data quality, and then conduct policy-relevant analysis so Harad was showcasing you one ongoing work, and I'm doing a showcase for a couple of others. By doing this work, what we have done basically, we have looked at the sky and tried to get some motivation for the work. So I'm showing you here the NASA dark matter map of the galaxy cluster, ABL, and why this is relevant. It's actually has quite some many connections to the derivatives markets. First of all, as Harad was mentioning, the EMI data is expanding every business day. So it's like space, which is sort of expanding in its way. According to one theory. According to one theory, exactly. But I think it's a mainstream theory for this respect. We also... We go back to the previous part. There is a second factor, which is similar to the space science, is that you pretty much need to be a rocket scientist to be able to work on this data set. So it's a complex data set. It's a huge data set. You need to basically be able to put in order in that before you can do analysis. And finally, like the dark matter here, the EMI data is like that basically. So it's sucking you in and you can't get out. So once you start working on it, you basically have to continue. So actually yesterday, what we did was that we published in our website an occasional paper number 11, which is getting some inspiration from this dark matter and other academic research, actually. And it's titled Shedding Light on Dark Markets. And this work actually is our first publication regarding the works of... Recording EMI data. It focuses on three market segments of derivatives at the interest rate markets, credit derivatives, and FX derivatives. And we already can conclude from that work that the EMI data, despite all the debate, can already provide useful insights. Clearly, there are data quality issues that should be addressed and can be improved, but already now we can already get some useful insights for the policy. What also Harald referred to, we have observed that there is indeed a very high level of intra-financial exposures, especially the intra-dealer exposure to G16 dealer exposures that were referred. Moreover, what we have also observed in the EMI data and it's highlighted here in the occasional paper that the market structure and the network of trades and exposures is already reflecting these key regulatory and other changes, so the central clearing applications and compression and so forth. So I recommend all you to go and look this paper. Today I will focus on the two concepts, the flow of risk. So I will explain you the concept, which is basically the method to track transfer of risks. I will show you one chart of what we have done using this methodology to map the global CDS network using a global DTCC data. And I will also show you some example of ongoing analysis using the EMI data, so the European part of the derivatives markets on geography of risk flows as well as potential ways to look for instance, wrong way risk. The second concept relates to the portfolio compression. I will explain you what is it all about. It's basically an post-trade technique aiming at reducing the cross-notional amounts, which are very large, as we know, while keeping the notional exposures unchanged. And then opening basically some further research for this implication of the portfolio compression. So just to give you a quick overview of the OTC credit derivatives markets, and here focusing on the single name CDSs, and this is basically an excerpt from the occasional paper, a table using data as of November 2015. And what you can see from this table is clearly that the OTC credit derivatives markets are mainly between the G16 dealers, between banks, and other financial institutions and insurance corporation and pensions funds. But what is also interesting here to look is that the role of non-financial firms in hedging basically credit exposures seems to be relatively small in this total amounts of trades. I will explain you the flow of risk concept, so you will see here a bow tie chart which illustrates how risks are transferred in the CDS market. So you can see basically in this chart going from the left to the right that the underlying credit risk which is being traded in the CDS is ultimately transferred from the ultimate risk sellers on the left in green through the strongly connected component of dealers to the ultimate risk buyers. So the ultimate risk buyers are actually those who sell the CDS protection and the ultimate risk sellers are those who buy the protection. And what you observe here is that there is a closed intermediation chain in between transferring this risk. Of course, this flow of underlying credit risk introduces counterparty risks which move from right to the left so the opposite direction. So what are the examples of relevant policy questions that can be analyzed using this flow of risk approach? So for instance, we can use it to map the geography of risk flows. For instance, cross-border exposures. Exposures to different type of entities and so forth. We can also, when combining the emir data with other data sources, we can try to also make, to understand actually why these derivative exposures are used. So why certain counterparties enter into this, whether they're hedging, gaining synthetic exposures or so forth. And we can also understand an important element which played a role also in the past crisis is the concept of wrong-way risk where basically the probability of default of the CDS protection seller is correlated highly with the underlying credit risk. As promised, this chart maps you the global CDS network. It's a one part. So this highlights really the complexity of the derivatives markets. I'm showing you here one flow of risk chart in one day at the time in March 2011 for one major sovereign underlying CDS. Naturally, the markets are very complex. So in the CDS market alone, as mentioned, this market continues to trade every business day of the week and it trades off hundreds of different underlying entities. And of course, when one wants to really map the whole network, one has to put this into the multi-layer network context and then do it in dynamic fashion. Here I'm showing you one snapshot of this global network. So it actually includes all derivatives that are in this major sovereign in the world. The DTCC is the main trade repository globally for the CDS market and accounts more than 90, I think 99% of the trades. So here what we can see clearly that there is a change of the flow of risk happens here. So the credit risk is being intermediated in the center. It goes from different types of entities on the left to the right or different entities. What we can observe here is that the ultimate risk buyers, so those two who are the sellers of the CDS on the right-hand side chart are asset managers, banks, and hedge funds. And as I mentioned, the counterparty risk goes to the other direction. What is interesting here is that when we look, not the absolute core of the centerpiece, but look a bit the outer circle, we also see some non-bank intermediaries. And of course, from risk analysis perspective, it's important to understand how they are being supervised and what are their connections to the other financial institutions. The sales of the circles mentioned the sales of the circles. They are basically the exposures. That's important. Yeah, sure, sorry. So here, now turning from the global CDS market to the European market, we are again analyzing one particular sovereign. Clearly I'm not going to tell you which sovereign this is and who are the institutions in this map. But what I want to just highlight you is that here we can basically look, again, in this one snapshot, who are at the end of the day the ultimate risk buyers and sellers and who are intermediating in the center. What type of institutions they are. And what is actually further use of this is to look not only the type of the counterparties, but also where this counterparty resides. And then mapping, basically, whether the underlying credit risk is correlated with the residents of the counterparty who is selling the protection. So in this chart, you would see that basically the red dots on the right-hand side are actually in the same location, which is basically the underlying credit risk that is being hedged. So there is potentially a presence of wrong-way risk. Whereas in this chart, which is another sovereign, we don't see this happening. So what are the, let's say, the open questions and further research avenues. Clearly, we are working hard on this, but as I mentioned, you need a lot of rocket scientists around to be able to work on this. Resources are constrained and so forth. But what we are really trying to do is to map, basically, the portfolio and counterparty overlaps of different market participants between different types of derivatives and underlying entities. So this is one important factor for risk monitoring. Then understanding the dynamics of this network and develop different monitoring tools such that Francesca and I basically can just open the tool and see, okay, how the risk is going. So I know that Alberto likes to use the word of risk manager. So this would be actually this type of tools that would allow the risk manager to easily observe developments in the markets because they are very complex. You cannot open a 20 million billion observations. You have to have a visualization tools, other tools. We also want to understand the role of the CCPs and how the network structure is changing because of the central clearing. Understanding also the motives of different entities using the markets, mapping this with other data sources. And then one very important factor still to be analyzed further is actually the economic role. It was a bit referred in Harald's presentation of these very large, intra-financial positions. And also if issue which is very pertinent and important is the collateral flows and the reuse of collateral. Turning to the second topic of the talk here which relates to portfolio compression, you will see basically the paragraph from the MIFI regulation explaining what basically is portfolio compression. But just to summarize it in a nutshell, it's a post-trade operation that reduces the cross amount in the market basically without affecting the market participants net positions. This portfolio compression is actually a significant issue. Multiple derivatives are being compressed interest rates, swaps, cleared and non-cleared, CDSE single names and index products, also FX commodity swaps, inflation source, all kinds of derivatives. According to TRIoptim, which is one large private entity providing multilateral portfolio compression, reports that they have compressed 840 trillion of OTC derivatives. The notional has been eliminated in this trade compressions under 2016. I actually don't know when this statistic starts. According to ISDA, 214 trillion was eliminated between 2007 and 2012. So we are talking about very big numbers that are being and also Dennis was referring to this. There are many reasons why this portfolio compression takes place. One clear motivation is actually that this is a required by Article 14 of the Commission Delegated Regulation 149 slash 2013, which basically says that you need to have a valid explanation to the relevant competent authority for concluding that portfolio compression exercise is not appropriate. So you are encouraged by the regulation to do so. There are also private sector or individual incentives to do that as clearly you can reduce counterparty risk, operational risk management of all kinds of collateral settlements and so forth. But there is also important factor which is related to the incentives to shrink the balance sheet for regulatory requirements. The bilateral compression can take place between mutual agreements between two counterparties and the multilateral compression is basically done through our external service provider. Okay. The portfolio compression, just to show you what it's all about, you see the two charts here. On the left-hand side you see the before compression situation and on the right-hand side you see after the compression exercise has taken place. The key here in this multilateral compression is that this basically happens in this closed intermediation chain. The bilateral compression can take place also beyond of this A, B and C. It could also happen between C and D for instance. But here when we are looking at the multilateral compression and there you basically, you're required to look this intermediation chain. Quickly introducing the concept of notional access. So basically this intermediation between the dealers leads to notional access. It's a concept by a few of my colleagues who are working on a theoretical paper referred here where basically they show that this access is part of the cross-notional that can be eliminated without changing the net positions. So at the end of today the compression is a network operation that reconfigures the web of outstanding trades so that basically there is a lower access in the market. So the work by Marco De Rico and Tariq Rogni basically identify different classes of compression. They also look different necessary and sufficient conditions when compression can be applied. What are different methods to do compression and also what are basically the resulting changes for the network structure. Just very briefly give you some example of how we can basically look this compression using emir data. So depending basically the level we aggregate trades and the algorithm that you use to do compression, we find that roughly 20 to 50 percent of single name CDS notional can be reduced. Naturally if you include prouder set of reference entities that you sort of for instance a little bit different timings of the notionals when they expire, you can enhance this level of notion that can be reduced. Key point here is that again the ultimate risk sellers and buyers on the left and right are not affected while as the compression is focusing on the excess between the dealers. So in these charts we show basically that the notional excess, so the amount that can be basically in theory reduced, was reduced with certain algorithm and certain assumptions by 24 percent. So one can clearly reduce this here. So what are the implications? Clearly the numbers mentioned earlier show that compression is significant and is reshaping together with the central clearing the OTC derivatives at the moment. Generally we are very much under researched in this area of importance. So again I would encourage people to really focus on their efforts of this topic which is indeed quite important. And one factor which is also clear to us when looking the EMEAR data, it's actually not trivial to identify in the current EMEAR trade reporting framework the history of portfolio compression. The private entities that are making the portfolio compression are keeping track of the trades and they have the data but using the public reporting it's not very easy to construct basically the history of compression. So what are the implications? Overall the cross reduction of notional amounts is positive and it may reduce opacity in the time of stress. So the G20 objectives are being addressed here. There is also clear positive impact on the reduction of the payments due at the times of stress and also the liquidity needed when it's really desired to have. So this is a positive factor. However there are also some dimensions that would need to be further researched. And clearly one factor is what are the implications for the counterparty risk at times of stress because we are actually reducing the number of counterparties as we are changing this exposure structure. We are potentially concentrating more counterparty risk to certain parties. Even potentially more important factor is that we have no yet results whether basically the compression can enhance or reduce network fragilities by altering the network structure. And this is very important to highlight that the compression that is taking place is between private entities who do not know the whole network of the markets. Actually there are very few instances who know the global markets. I mentioned DTCC for instance has a data globally. But currently the regulators or the ESRP for instance is limited with its access to the European derivatives markets. The parties making portfolio compression have even more limited information. They are relying on the information that the counterparties who want to make the compression provide to them. So basically the compression is done with in a local information and not globally using global information. And this can have implications from the network theory perspective. Would there be any instance that it increases risk because they do not know the whole picture? So this is the question that I'm asking. So basically when you are operating in a non-full set of information you might end up in a situation which is basically leading to the network structure is actually less robust than when you would use global information set. So this is one factor which needs to be analyzed more in depth. Finally, also things that needs to be looked at is basically the macro-pronential dimension of the compression for capital and collateral I mentioned. There are private incentives to do compression for instance to reduce the level of the balance sheet. So we need to understand basically more what could be these type of implications. One final fact that I want to emphasize is that there is clearly important further need for regulators to have more transparency of the portfolio compression methodologies. And also as I mentioned on the emitter reporting so that the history of trades can be better constructed. Thank you very much. Thank you, Tuomas. Very interesting. We much further ado Professor Paditone, if you want. My turn. Point to you. Thank you very much for inviting me to discuss these three, let's say, talks. And thank you for sending me the slides yesterday. So, you know, my, a lot of time. Yes, good morning. Yes, good morning. But, you know, I'm very happy to discuss this topic that is related a lot on what, on part of my research. So I'm starting with the first talk and he's on stop and go, the reform agenda of the OTC derivative market. And, you know, clearly this, this, I find this analysis very interesting. The key point that is, it has been stressed is this OTC market structure is inefficient pretty much and there is an inefficient risk allocation. You know, my role is just to start to ask questions on this fantastic, let's say, data because I think that we shouldn't just stop and look to what the data on this market are telling us but we need to consider also that this is one of the market through which, let's say, firms can hedge their foreign exposure. So, you know, I have several questions and that I've already discussed with Aral a little bit. So the first thing is, why are you surprised that in an OTC market there is concentration? Pretty much this is something that we find everywhere and is largely due to this information, let's say, so this economy of scale coming mostly for getting information. So this kind of structure is the same that you find in the CDS market, is the same that you find in the OTC market for corporate bonds, sovereign, and so on. So clearly I think that we need to take it as the result of the market as soon as you have this kind of economy of scale. So the question is, is this is good from the systemic point of view or not? And we decide that it's not how we can eventually try to change this structure. Then I think that on the other side, we need also to seriously ask to us because this is not again related only to the forex market but is related to a lot of other OTC market. So why is this structure of the market inefficient? So which is the perspective from which we are saying that it is inefficient? What are the alternatives first? And, you know, and why if there is some alternative that has not been implemented? Pretty much what it has been already stressed maybe there is this, you know, incumbent that are already there and it's very difficult, you know, even if there is a huge amount of money that you can earn it's very difficult to come in and, you know, maybe change this structure in a different way. But pretty much so far what we observe is that in many cases there are many players I have long chat with Deutsche Börse about this for example that they try a lot to set up in several different OTC markets starting from the Bund market for example. Bund OTC market in exchange rate market and they were never able to do it for several reasons. So I think that we need really to try to analyze deeply why this is impossible. And on the other side from the inefficient risk allocation where I think that we need to be seriously conscious that if we are really preventing banks to be at the center of this OTC market we need to figure out who will be the alternative. So they can be potentially edge funds, insurances, pension funds and the question is is this really the solution that we want? Do they have the capacity? Do they have the ability to do this? Are we eventually creating other systemically important institutions? You know, talking about edge funds maybe is a good thing having them doing this business but we know that they can create also some problem. In the past we have some examples. So I think that we need to be pretty much when we are saying it is too risky to have these 16 let's say big players in the market to play this role we need to seriously also address the issue on what is the alternative? And then I'm just asking what regulators should do because pretty much what the experience is telling us is that you can impose that all trades are on the exchanges but the question is, you know we have already exchanges on the forex market. You have the exchange trade market for this let's say the spot and you have the future market. So why we still have an OTC market for the forward contracts? And in some case also for the spot contracts. Either people that are going to trade these firms that are trading in the OTC market have no access to the exchange trade market. So the question is maybe we are putting too much bias on these exchange trade markets because you know otherwise there is it's not that we are talking about the market where there are no other possibility like the corporate bond market. We're talking about the market where there are alternatives. So why these companies are not going to trade in the alternative markets? And they are the exchange trade. The only way that you can think to implement to make eventually this market more efficient. The other alternative that I think regulators should also consider is to have more transparency as it has been done in the corporate bond market with trace and maybe this will be enough. This is clearly a question. I don't know if this is true or not but clearly there are several steps that you can do for these over the counter markets to improve their let's say efficiency. And one thing before to move directly to imposing an exchange trade market that maybe you are just killing this market is to think to have more transparency. And again, I'm asking how peculiar is really the forex market because you know how different is this market microstructure with respect to the exchange forex market, maybe even there you have that you have these 16 large players you know that are the market makers of this market. So even in that case you will have that they're also having all the rents coming from this market. Is this a big issue? Is this a problem? And again, how different is the distribution of the profit in the future exchange market with respect to this one? It would be nice to have data and to analyze this. And then how is costly again to get access to the future forex market? Because if it is the cost larger than the one paid on the spread for the forward forex market well maybe this is a good reason of not going in this other market seems so efficient. So as you see I have several questions that I think that will be nice to start to address. And then the other thing is that why technology you know didn't reduce this kind of barriers or is not really improving the inefficiency both in terms of you know having these large players playing this big role and also in terms of you know of distributing better risk among the different player in the market. And do we really should care about client and experience? You know it's not we are talking about householders that maybe they don't have a lot of experience that we should protect. They are companies. And you know is this really so relevant for systemic risk that we are caring about this client that have inexperience? You know the forex market is not such a complicated market. So why do you have these clients that are having this inexperience? Are they really an expert or maybe either they don't have alternatives or there is some other reason of why at the end you know they're paying this large spread. Now I move to the second talk that I think is very fascinating and is related to cleaning CCP and cleaning house and to be honest I'm largely working also on this topic. I just pick up two things that I think are really really important even if I think all the points rise are you know well placed. And one is related to without access to the central bank in the relative currency this result in increased unsecured deposit at commercial banks during a stress event can be you know very dangerous. So I think that the big question that you need to ask about the access of the CCP or cleaning house to the central bank is you know what is their legal status? If you're looking around even in Europe but also around the world all these different central clearing have different legal status. The reason why in the US FDAC is the regulator of CCP and ICE pretty much and the other is because they are legally banks. If you know in some other states this kind of let's say institution are not banks they are not you know subject to that kind of regulation but to something else. So I think that the first question we need really to ask about this institution is what should be their optimal legal status? And then you know the second question is how they should be regulated and how they should be supervised? Because as it has been already stressed you know this kind of institution are subject to completely different you know type of regulation from the macro side to the macro perspective. But really you know if you are for example consider a bank you will be under the supervision of all the other banks but on top of this you have ISMA that is overlooking over you because clearly you are playing a big role in the market. So clearly I think that we need really to seriously think on how to build up a serious regulation on supervision of this kind of legal of these institutions. And then I would like just to stress another point. More than serious uniform you mean? Sorry? More than serious uniform. Yes, exactly. Okay. And in terms of CCP recovery and resolution clearly this is also another important problem and there are the issue you know who is the resolution authority and when should be this resolution. Regarding to this I would like to stress that I just wrote a white paper with Jan Kranen on predatory margin in the regulation on supervision of central counterpart cleaning house and you know we are really addressing exactly this point on top of the other issue about how to really regulate this central clearing and so on. And you know one of the thing that we stress in this paper is that in line with this robust but fragile property because I think that the really CCP reflect exactly this position. A CCP triggers a systemic risk event with small but clearly possible probability. We cannot whatever is the kind of requirement that you are asking prevent this possibility and clearly in this case you know there is no other way that the government will rescue this. So the answer to the question who are going to pay we know there will be a bailout. There is no other way. And what we stress is that that is not pointed too much is that we need also to figure out how is the optimal market structure of the CCP? You know so far we have a strong competition that is going on among at least in Europe among 16 different central, different CCP and clearly we know that when you are competing there will be of course the risk of under-margining by aggressive CCP in order to have the large fraction of the market. And this is clearly having an impact on the level of systemic risk that this kind of institution may eventually create. And then the second thing is also related to transparency. How much transparent is their individual exposure and if you have a lot of them clearly you have again a problem about under-margining. So this is also another thing that we think should be stress and analyze. So you need really to have an efficient design on one side we need to figure out how this competition can affect the level of margin that they are going to charge. And then clearly we need to have a proper regulation because if you are allowing them to compete you need to regulate in order to have this under-margining effect. And clearly the best solution is that you should have a supervisory practice that is not as it is now mostly at the national level that is just related to where they are located. And clearly this supervisory standard should be uniformly applied without regard to the local champion. So even at the national level you shouldn't have big difference between the larger than small one. And then clearly in terms of the regulation and supervision it should be centralized in one agency as there is also the hope that this happens in Europe and including all the national let's say economies in the CCP counterparty that are clearly domiciled and the set of country that will clearly ultimately face the bailout should be clearly agree on this. And remember that if we are then agreeing that we have a single supervisor and we are maybe treating them as banks because they are getting access to the central bank well the bailing rule that we have in Europe is not working for CCP. So we need to think to something else because clearly this total loss of serving capacity for sure will be not enough and clearly this robustly fragile nature will clearly prevent the possibility of having a bailing that will be enough and you need to have a bailout. So clearly we need to make sure that the bailout will be not just at the nation where this CCP is located but we need that all the different let's say countries that are involved where the CCP are having clients should be clearly considered for this and we know that the bailout clearly is giving a strong incentive for adverse election and moral hazard. This is why we need to have a consolidated supervisor in Europe. Then I'm going to the last topic that is analyzing credit derivative market, flow of risk, national access and portfolio compression. And the first thing is that clearly this is a very great, nice database. Every researcher would like to work with this data and few months ago I was in the same room here at the statistical conference and I was stressing there as I'm stressing now that this data should made available also not only to the supervisor but also to academicians because the improvement and the kind of a nice that you can do with this data will improve a lot our knowledge on how these markets are working and so on. Even if we know that there are some limit on the data that ESRBR is having because it's just only if the reference entity is a European reference entity or one of the two traders is located in Europe you have this data so clearly it's very partial and it's so surprising that we cannot get a global database about this kind of transaction. So the focus is on the flow of risk, wrong way risk and portfolio compression and regarding flow of risk what they find is that CDS market is highly concentrated on few central dealers, nothing new you know and clearly this is the same result that Getmansky, Girardin, Lewis find and they publish this in the journal Alternative Investment using DTCC data on CDS transaction, well you know here as the one shown by Peton and Tuomas you can see that pretty much you have these 10 big players that are at the center and you have the rest, you know just buying and selling is more fraction. The other thing is that concentration of ultimate risk buyer are hedge funds and asset managers so clearly I think that we need also to ask, so what? It is okay, this is what we want or this is something that we would like to avoid. So clearly it will be nice also to have a view from the systemic point of view if this is really what we want as a final result of this, let's say transfer of risk. The other thing is this wrong way risk what they find is that pretty much they just show to us two cases but it doesn't seem that this large wrong way risk for sovereign risk but clearly the question is what about financial references? I think that there is really where we have wrong way risk and again from the paper of Getmansky, Gerard and Lewis what you get is this table for example where you can see that really there is a huge amount of wrong way risk among the different, these 10 institutions so pretty much one fifth of the total amount is really there. And about portfolio compression I think that is very interesting and important we are thinking about portfolio compression how to eliminate or reduce this counterpart risk and so on but clearly if we are going on the direction of having OTC contract clear well directly you will have this portfolio compression because this will be really what at the end the clearing house are doing. So what they get is that you can really improve a lot the level of let's say the amount of risk that it has been compressed by looking to the data that they have but you know it will be pretty much the same or even less you can have more by clearance and you know what I'm going to present you is really what you can get by using in this case this is a work that I'm doing with SCC with Giulio Gerardi and Mario Bellia that is one of my PhD student here. This is a work that I'm doing on DTC data on CDS. So the question is to clear or not to clear and unfortunately I can't show to you only one slide about this but pretty much what we find is that even if it is not compulsory in the US that you have to clear your transaction at the end you know they do. So more than 80% of the transaction that can be cleared because they'll say the dealer is member of the clearing house or because the reference entity is accepted by the clearing house or because the characteristic of the contract is really in line. So this means that you don't even need to impose it. As soon as you give the possibility it seems that now the market is really reacting in a good shape and the decision to clear balance the cost of the CCP margin against the additional capital requirement that clearly you will have because you are not clearing and we do find that you know really the decision to clear do reacts to the incentive that the regulator are providing. So you know you have that the one that are clear are the safer and more liquid. They tend to flatter exposure with the CCP in order so they clear only really they have a reduction in the margin requirements and usually you know you are clearing because the counterpart is more risky. But what is coming out is that pretty much less than half of the dealer to dealer CDS trade national value qualify for clearing. So even if you will impose to clear at the end you will have that only a small fraction of them can be clear because we have this issue about the non-standard contract. So the issue is either we need to try to accept more non-standard contract and this is a question for LCH how really to improve the amount of content that can be clear and you know because it seems really that the market is already reacting on this direction. Thank you very much. Thank you. Now I remind that we are already running out of time. It's gonna be lunch. So every minute that we take now is taken away from lunch, right? Authors please. Just a very brief. I don't wanna comment the many interesting questions that Lajana posed but just two remarks. First is I happen to read Angus Deaton's Nobel Prize speech in Sweden last I think November it happened and he made a very simple point that all good policy starts with good data or with better data. So I think what he said about development economics equally applies to financial economics. We have the unique chance here with the ME data to elevate I think the policy debate to a new level and we should make as much use of it as possible and have a more informed policy debate on many of these regular issues. Secondly and I think Lajana is very, very right on this. We have to make sure that we bring in as much human capital into this debate as possible meaning a lot of finance researcher that can very productively use this data. And so there must be some found some way of broadening the participation in the research agenda here by giving access to data like the Bundesbank does. And I think that's also very, very important. Just react to one. It's gonna come, it will come. Okay, I'll just react to one question about the level of margins and the, if you like the motivation to lower margin, sort of a predatory margining regime between different CCPs. We had an example, big example a few years ago where the obviously the interest rate environments have fallen, we've heard enough about that today. But the impact on CCPs was pretty dramatic because the old relative return way of calculating margins, the distribution of margins using relative returns was giving margins that were far too low. And many, many people in the market were just as upset about it as I was. And we moved to an absolute return distribution way of calculating margins and resetting them. The consequences were that the margins on the swap clear service, which is the big one that I was talking about, doubled. And the interesting thing was from a commercial perspective you would have expected that would have been basically suicide, but in actual fact our market share increased dramatically from then on. So the members were also just as concerned about the relative levels of margining between them as we were. So I think that effect shouldn't be missed. As I also like eating, I will be very brief. Let me thank Lorena for her comments. Actually the key difference between our analysis with the global disease data is that it's a global and not the U.S. So we, but it's confirming the same type of results. Recording the, I like actually the attitude of so what? So this is the way that researchers should operate so they should ask so what? So basically regarding the ultimate risk buyers being hedge funds or asset managers, so what? Actually I think risk sharing is very important and it's fundamental piece and it was related to, but also you mentioned to Harald that it's not necessarily good that every, all risk is concentrated in the G16 so having a risk sharing is very important. Why we do this is we need to basically map who is exposed to what risk, when basically the risk materializes and therefore it's also important to know what are the entities exposed to this and it's also important to know that these entities who are exposed to risk are prudentially supervised. So this is one important factor. The second point is that on the wrong way risk for banks this is actually something that we are working on, clearly a very good suggestion and regarding the compression, as I mentioned there is clearly positive aspects, less notional, less opacity and so forth, less liquidity needed. However, we don't know yet all potential issues and that's what I was raising so we need to work more on potential risks rising from this as well. Thank you very much. Thank you. We do not have time for questions from the audience so if anybody has questions through the authors you could approach them during the break but unfortunately we need to cut the time for that. I'm very sorry for this. We have a very dense program which will continue in the afternoon. As to the location of the lunch it is exactly in the same place where it was the dinner yesterday. For those who are not there we have to go back to the entrance and then turn to the left and then when you will arrive at the security doors my colleagues will help you. To pass the security doors you need the badge and then we will have to go to the end front of the Grossmarthalle.