 But you are the first. So, please. Thank you. Three quick comments. One, on Stephen and transparency and over-the-counter derivatives markets, I think I'm a great fan of transparency, except when I'm not. And I've been convinced over the past year to, by the work of Bank Holstron, that transparency can sometimes have a significant negative impact on liquidity in certain kinds of markets. And I think we have to be very careful about when we release data and the kind of data that are released. And second, following up on the data point and Lex, I mean, I think, Lex, we had five papers presented at the general board yesterday of ESRB using the eMir data, right? That I think demonstrate very clearly and conclusively that those data are useful, not just for academic research. I say just. I, of course, think academic research highest priority, but not just for academic. But that these data can give you conclusions of great policy relevance. And so I'm a little bit worried about demoting the data. And then finally, Vitor had stressed synthetic leverage. And then Gabriel talked about insurers and hedging. And I spent about close to an hour on the phone with a senior person in AXA the other day talking about trying to learn about how synthetic leverage is used by insurers. And I came away somewhat reassured. I went into the conversation, this is dangerous stuff. And let's put it this way, I still have to be convinced fully, but I see your point very strongly. Yeah. The problem is that there are no perfect edges. And I know of examples in the crisis of institutions that have collapsed because of the edges, which worked in the opposite direction they expected. But OK, now I will give the floor to four answers. Yeah, just now. Yeah, so I thought that Stephen was the first to be addressed by the transparency question. And then I'll give the floor to you, yeah, Stephen. No, just to emphasize, I fully agree with the remark, Richard, is that the process that you're now seeing by bringing things on exchange, it is done in a very measured way. And so only very specific bonds where we have, this is under method two, where it will be first measured how many transactions are there already happening. Because precisely, in some cases, if there is not sufficient liquidity already or not sufficient transactions already, if you then oblige transparency, it might actually kill the liquidity. So it is a general trend to also bring other instruments than equity to the lit markets. But of course, always very well calibrated and fully understanding that probably transparency in the Rifted markets and bond markets will never get to the same level as in equity markets because typically these are about much smaller issuance, et cetera. And so also under the incoming regulation, which bonds will be under transparency obligations will be very carefully calibrated, also taking the actual trading into account. OK, very good. Lex, please. First of all, to underline it, I also very much agree with Richard's first remark. That's why I, in my intervention, said transparency, yes, to the extent possible and wise. I think you elaborated on that. I fully agree. On the data collection, I think my point is not that data collection is too expensive and we can do it in a cheaper way. Now my feeling is, even after other remarks on how useful, for instance, we can now predict systemic risk. We can predict the previous crisis. It doesn't tell anything about predicting the next crisis. And I still feel that we are talking about data collection without a good connection to our ultimate objective or at least the objective that is mentioned, often systemic risk. I don't see that you can make the connection. You will ever be able to do it. And you lose time in really responding to actual threats to financial stability. Now, please. Much of the panel focused on reducing leverage, either through minimum haircuts or explicitly or the synthetic leverage, and clearly leverage as we know can cause problems with fire sales. But if the real aim is the resilience that several of you talked about, then the other theme of the panel, namely diversity, becomes potentially all the more important. And leverage, perhaps paradoxically, can be a source of that diversity through relative value trades. A minimum leverage in one place, 10 times leverage on trading leverage loans is not the same as 10 times leverage on trading bonds against OATs. And so how can we be sure that we don't end up inadvertently reducing the diversity? Unfortunately, I can't share the same vision of the diverse ecosystem that Stephen described. As I look at investors in the market today, what I see, and this is where the interaction with monetary policy unfortunately does come back in, what I see is people giving up on relative value trading, giving up on single name stock selection, putting their money into the ETF, which is the single best performing thing in the market. And part of the reason for that is also that the constraints on leverage, which would have allowed hedge funds and others previously to do relative value trades, again, they're just not doing the same thing. They're all forced into the same trade and that trade is basically rich field. Yeah, I have two more, at least two more, one back there first and then you, yeah. And so I collect now three. Back there, please, the mic there and then here. Just to respect the order of the requests for the floor. Yes, please. Many thanks, that's a question to Lex Sogdwin. I'm all of it from the ESOB second Harriet. And this was on the comments you made on prosyclicality and you mentioned you had the anti-prosyclicality and tools. Now understand that, and those of course work, say in the downturn of the cycle when asset prices fall. But what do you do in the upswing of the cycle when asset prices rise and your margin and haircut models tell you that you actually should lower those margins and haircuts? How do you behave in a anti-syclical way, in a sense there? And then the second question, and they're a bit of self-advertisement for the work of some of my colleagues who've just published a working paper in the ESOB working paper series which shows that client clearing is very important in Europe. So there my question, the second part of the question would be how do you ensure that your prosyclicality limiting tools all the way carried through to the end user, i.e. to the client, namely when a client doesn't clear, well, if somebody doesn't clear directly as a CCP member but as a client to a member and how do you make sure that your prosyclicality limiting tools actually extend all the way through to the end user rather than stop at the clearing member who then may behave in a very prosyclical way towards the client? Thanks. Thank you, and to finish this group of three questions. Thank you, yeah. I wanted to come back to the data question very quickly that Lex and others have touched upon. I mean, for me it's not just about collecting more data about sort of specific institutions, but when I look at the shadow banking monitor of the ESOB, the two numbers that struck me most is that we have 40 trillion of shadow banking assets approximately, half of which are in OFIs, other financial institutions, and it's not exactly clear what these are still. And second, half of those, not necessarily the same half, are in Luxembourg and in the UK. And so clearly the assets of shadow banks are very different depending on where they're located, very different on what role they play in the system. And it would be very good to know more about what these assets are actually are and what they're actually doing. And so in a sense, I wouldn't really mind having some of Ufundstein's bond funds less regulated or less documented because these are harmless animals, we understand them, they're counted somehow in the shadow banking system. But for macro proof, we really want to understand what's happening with those large amounts of assets that are fairly opaque and basically located somewhere where they certainly don't belong in terms of the real activity. And so that would be much more interesting in terms of data. And I think there we have quite a way to go still, unfortunately, in particular if I think about who is collecting those data. They are very often still collected by national authorities. So we are still fragmented in terms of data collection in Europe, whereas this is something that the ESOB or the ECB should be doing according to standardized guidelines, according to standardized procedures, according to standardized data categories, et cetera. And we are currently just hoping that the Luxembourgers get it right or that the UK's are getting right or whatever. I mean, I'm exaggerating a bit here, but here is I think an enormous amount of data that we still in fact need, not just because we want to go further and further in academic studies, but because you want to understand what are the channels in the macro system when it comes to monetary policy or asset inequality or whatever? Yes, if the financial system would not get ever more and more sophisticated, we would not need more and more data. Yeah, right. So, can I ask how many still want to ask a question? Yeah, so if it is just one and it's a test goal, I'll give you the floor to then to join to the three questions that we already have. Thank you. I wanted to pass a message, which in fact I already referred to this morning when I was intervening, but was very early in the day. In Europe, we are using the data. It's very important to say this, not only, but we are doing a very important investment in terms of IT. This is really big numbers. We have created a technology based on disk, which is a machine which is able to elaborate the 50 million data entries which we get every day. We are looking at possible indicators to identify vulnerabilities. We are creating an infrastructure to examine this type of thing. So it's very important that we pass the message that in Europe, these data are being used, will continue to be used. There is a very strong consensus of all the institutions to do it. And this is not something which is so simply planned, but this is work on which I am having one or two meetings per week. Okay, very good. So we have had questions that at least involve Lex and Steven, but of course any other member of the panel should feel free also to chip in and comment on the questions that were raised. But I start with you, Lex. Two questions were asked to me. Let me take one of them and then see what others want to say again on the data. Perhaps trying to make my point again, slightly different way. The financial system, even before it became more sophisticated, has always been a complex system. And on a complex system, by definition, it's impossible, literally impossible, to collect all the data that in the end drive systemic risk, drive the decisions being taken if only because the main data are mental data, expectations of people on which they act. So you can, whatever you do, you will never get a full picture of the state of the financial system at any point in time. You cannot take a photograph of the financial system, let alone make a video or a movie of it. So you have to select at what you look and you cannot just go to collect everything that you want to collect. But because that's the reflex that you always see. We have missed something, so we start to collect data on that. The point is that without a conceptual framework, without a theory that links systemic risk to driving factors and interconnectedness is one. And we have discussed that that is still missing in the framework. Without a clear conceptual framework, the data are silent. They don't tell you a lot, it's just data. It's bits and bytes. It has no meaning. It only gets meaning within a conceptual framework in a theory where you link it to what you want to achieve and the factors that are driving them. And I think systemic risk is simply beyond measurement. There was a question too, you want proselyticality. Yeah, proselyticality. That's a more technical question. So if the markets are in the uprise, let's say the margins are driven not so much by the increase or the decrease in price, but are driven by volatility. So if you have increasing volatilities, volatility, your margins go up according to the models and of course in line with all the regulations. What we want to avoid is that we have to sharply to raise margins. The flip side of that is that at the moment that volatility comes down, so you would reduce margins. That you then do not simply reduce the margins all the way but keep them a little bit higher than you would otherwise do, so that if volatility increases again, margins do not need to, don't have to be increased that sharply. The otherwise would have done. Stephen, leverage, proselyticality and so on. I will be, many issues were raised and I will limit it to two issues and on the last point on the data issue, I think on this point and also considering where it's Friday late at the day, I violently disagree with Lex on this idea is that we are in the situation where we are asking for the last little element of the financial system. We are at least for financial markets in this situation. We're still very fundamental questions. We cannot answer them. If you now ask me what is the current leverage ratio of the EU hedge fund industry, I cannot give it to you. We don't have reliable data at this stage on the leverage ratio in the hedge fund industry. We are now getting closer to information, complete information on counterparties in the EU but it is working hard as an authority to get the quality of the data better and reliable and so at least for financial markets and I cannot judge it. I'm less able to judge it for banking and insurance. We are still in a very starting phase of getting data. I can remember in 2008-2009 if we had securities regulators would get together and we would have a risk discussion. We would have on the table the reports from the banking sector, their own reports, the commercial banks providing us the information on their risk analysis. So on this point I really need to make the perspective and give the perspective. This is ongoing work. I know it can become a never-ending story but we're clearly not in that part. Final very little comment on the point on the ETF, et cetera. Obviously I am all for fundamental analysis by professionals, et cetera. I was questioning whether for an end consumer that has a couple of savings would be the most logical to go into active management. So that was the context of the remark. Very quickly on two subjects on data. I couldn't agree more of what Stephen was saying. Actually I think the biggest value that we see now already for example in insurance is the fact that we have harmonized the data. Because like this we can have an idea and much better information about how the system as a whole behaves in Europe. In the past it was very scattered in country A or country B or country C. So the value of having this harmonization I think it's fantastic. We should go in that direction and we are definitely far from going on all the details, et cetera. But for example on the asset data we are now I think in a much better position because we have harmonized that. On the other point on prosyclicality and I couldn't agree more that this is something that we will need to continue to look from a point of view of the regimes and to try to embed in the regimes these kind of anti-cyclical elements. What we have done for example in Solvency 2 on the equity side we have already embedded in Solvency 2 what we call an equity dampener so that the risk charges on equity they are increased when the markets are really booming on an uprise and they are decreased when the markets are going down precisely to mitigate this point of our fire sales and to give the I would say the right incentives and to try to overcome prosyclicality. So everything that we can do I think on the regimes in this case a micro potential regime to deal with this element of prosyclicality I think it's good from a macro from a macro potential perspective. Mario. Maybe just one more because I must say the more Lex talks about data the more you get it wrong, I'm sorry. I mean the point is certainly not the last data but actually if you take the share of the data we are collecting with respect to the data we are produced we even go down in share because of course there is a huge amount of data that are produced and what we are trying to extract is the trend what we are trying to extract is exactly the harmonization that Gabriel was saying so sorry, relax and collect data. You. You want to add something? Well I have to weigh in on the data I commend the use of data maybe I'm more practical and more proportionate the new MIFID directive is a big bank to collect even more data because in the US it took trace at least 8 years to be implemented maybe there is a bid ask here between the pacing and proportionality so that the quality of the data can be very useful for a policymaker. Well I thank you very much and we will stop here but not without me also adding something to the data to the data issue just to reinforce now the point of view of say the economists and one of the things that is at stake in our discussion here is precisely non-banking and what initially was called shadow banking and shadow banking is not a set of institutions it's a set of activities and instruments because shadow banking was done also by banks so the proper concept of shadow banking it's a longer story to explain and the way it was then misused but the point is why these word shadow banking came it came because precisely what happened before the crisis was the attempt to create private safer assets in view of huge demand for safer assets and scarcity of safe assets from the official side and the attempt was done by using OTC derivatives secretizations and repos these were the three instruments that created inside liquidity short term quasi-money instruments and all sorts of edging to make safer some of the instruments that were being used and the point is that these things are not in the statistics after the second world war all countries made a big effort to have national accounts to have flow of funds to have monetary statistics they don't include any of this let me give the example the forms of short term liabilities that act in the financial system as quasi-money are not in the monetary statistics not in the monetary aggregates the exposures that come in the flow of funds are totally changed by the derivatives activities so what is there in the balance sheet and goes into the flow of funds in many cases means very little because the exposure has been transferred via derivatives and it's not there in the flow of funds numbers and also these means indeed that we need and there is a very good book by economists Bruno Meyer, Gary Gorton, Krishna Muti and many others called risk topography where they demonstrate and appeal to a big effort from the advanced economists to go in the overall of these basic apparatus of macro statistics that was developed many decades ago and no longer reflects what is going on in our financial system and with that appeal to last word to data to reinforce what you said which was also very good and thank you very much for all panelists and we conclude here