 Good afternoon. I'm very pleased to be here. And I'd like to start by saying a few words and first to thank Bob DeYoung and Luke Levin and Phillip Hartman for including me in this anniversary celebration. I guess I should have known, given my own history, that this was the 50th year of the JMCB. But it came as information to me when the email invitation arrived. And Bob, who sent it to me, said that they would like to include me in this. Anniversaries are often a time for some, a little tiny bit of reflection. And so before I introduce these esteemed panelists, let me say a few words. The vice president this morning noted a number of issues about the history of the journal Money, Credit, and Banking. He noted how it was started by Carl Bruner in 1969. Interestingly, several years later, Carl Bruner left Ohio State University for Rochester. And he attempted to take the journal with him. But instead, we got the Journal of Monetary Economics. And so that was interesting. And I think what's really interesting for me is that they serve totally different purposes going forward. In I was privileged to become the editor in 1992 with the late Paul Evans. We were colleagues at Ohio State. And Paul and I made a number of changes, or I guess maybe I should refer to them as adjustments if I was being more politically correct. And we took, I think, a critical decision that we didn't understand at the time how beneficial it was going to be for the journal and for the profession at large. And that was to include, to invite people from outside of Ohio State to become editors, real editors, not just associate editors, but co-editors with us of the journal. And this led in short order to the creation of what Bob referred to this morning as a banking editor. And I think that was really important. I just want to emphasize that it wasn't our foresight. It was more like an accident and more like the realization that we were getting lots and lots of papers that were coming in as submissions that we did not feel that we were competent to handle as editors. Now, when you're editing, a number of you have edited journals. You know that you always go past the limit, probably, of where you should in retrospect go. But this was so far outside of what we did that we figured that we needed help and we got help. And some of the people that helped, like Mark and early on and Alan, are here today. So it's hard to recover how you felt about these things 25 years later. But at the time, I think we can all agree that at the time in 25 years ago when we started this, that banking wasn't really a field of anything. And in fact, finance wasn't really a field. It was like a subfield of economics. People would get degrees in economics and they would study finance. They'd go find somebody who knew something about finance and they would study finance with them. And I think that just to put emphasis on this, a few years later, when I was asked to be the director of research at the Federal Reserve Bank of New York, we obviously needed people. And this was several years later to do research on banking. I don't know if Don Morgan is here yet. He will be here. He was there then. But the interesting thing was that in order to hire people to do banking research in 1997, I looked for labor economists. And the reason I looked for labor economists was that they had the tools and they knew the methods associated with doing banking research. They just didn't have the data, but we had the data. And so that was sort of interesting. And I think that the way the important thing about the reason this is related to the JMCB is that the JMCB has existed since the beginning almost as a forum for discussion of topics that related exactly to these things in its name. And that included not just monetary and macroeconomics, but also the economics of banking and the economics of intermediation more generally. A second point I want to make is related to the fact that the conference is here at the European Central Bank. In the early 1990s, there was genuinely, I think, very little interaction between policymakers and academics. Those of you that were already in the profession at the time remember, I'm sure, that you were off doing your stuff. But the amount of time that you got to actually interact with the real policymakers was pretty limited because they weren't that interested. And I think that, again, the JMCB took something of a lead in this. And what we're seeing today is the follow-through of that. In the JMCB, what we did was we started by convincing the Federal Reserve Bank of Cleveland to sponsor a conference that would be published as a special issue of the journal. And this happened in the mid-1990s. And this has persisted. And today, this has grown to the point where this morning we heard the new chair of the Financial Stability Board stand here and emphasize the importance of interactions between standard setting, policymaking, and academics. And so I think that now I'm not sitting here taking credit for this. I'm just relating the fact that there was a vehicle for all of this. And that this celebration in many ways, I think, first of all, to cover primarily banking topics in this conference. And secondly, for it to occur inside of a central bank are really exactly the right way for us to celebrate this. So I hope that everyone here still publishes in the JMCB going forward. I'll tell you the last two times I tried, I did not succeed. So it's always good to be, you reap what you sow, as I say. So if you try and help create a selective journal, then when you're on the short end of the selectivity, you can't complain. So I don't complain. I like to my papers, but that's the way it goes. So I think that I hope that the current editors, and I expect that the current editors, it will be hard work going forward, but they'll be able to maintain the quality and the diversity that's critical to the success of the journal. And I hope that those of you that continue to participate from the perspective of the official community also continue to support this journal and other journals that make that cross-fertilization and that relationship so important. So thank you all very much for being here. And with that, let me turn to the four excellent speakers that we have. And I'll introduce them all now in the order in which they're going to speak. So first on my far left is Mark Flannery. Mark is the Bank of America Eminence Scholar in Finance at the University of Florida. And I won't say how long you've been there, even though it is on your CV. And it's been a long time. Mark did serve, as I mentioned, as a co-editor of the JMCB. He did that from 2000 to 2005. He was also the chief economist and director of the Division of Economic and Research Analysis at the SEC, another thing that is more in keeping with the recent movement back and forth of people between the official sector and academia that 20 or 30 years ago was extremely rare. Mark has written extensively about banking. And I just want to mention very quickly that he has a 1998 paper in the JMCB. And I want to say that if anybody asked me what it is that I remember about what you've written, maybe this isn't what you want me to remember, but it was a very important paper at the time because what it did was it established the importance and the complementarity of market information and supervisory information in the evaluation of the creditworthiness or the health of a bank. And I thought that was pretty important and that is a JMCB paper. Second, on my left, immediate left is Anil Kashyap. Anil is the Edward Eagle Brown Professor of Economics and Finance at Chicago's Booth School of Business and an external member of the Bank of England's Financial Policy Committee. Anil is well known for his work on the importance of bank lending, where he was an early contributor to helping us understand the role of banks in economic activity generally. And he's also done a lot of work and research on banking and other issues associated with Japan. So much so that he's received the order of the rising sun, golden rays with neck ribbon, an honor which, by the way, is also shared by Clint Eastwood. So I think that this is really, I think, really an unusual thing. Anil published a paper in 2003 in the JMCB summarizing the findings of the first euro system research network that examined the monetary policy transmission mechanism. And that, again, is a, I think that research networks has served as a model for what's going forward in central bank, not just euro system, but more generally central bank cooperation. I can tell you that I tried to get some people at the BIS to imitate it in other parts of the world because it's such a successful model. But Anil was involved in that and that is published in the JMCB. On my far right is Till Sherman. Till's currently a partner and co-head of risk and public policy practice for the Americas at Oliver Wyman. If I didn't know Till so well, I would think that he came from the dark side. But it's not true. Till is really a bridge between the practitioners and the, and academics in the best way. He does advise, he does advise public sector clients on things like stress testing, capital planning, enterprise risk management, and corporate governance. And Till published a paper on computing credit losses in the 2006 JMCB. So he's a contributor and that's of course just one paper. So, and finally, immediately to my right is Isabel Schnabel. Isabel is the professor of financial economics at the University of Bonn and a member of the German Council of Economic Advisors. She's also a research fellow at CEPR in SESIFO as well as a research affiliate at the Max Planck Institute. Isabel has published a paper on bailout expectations and sovereign bond spreads in the 2006 JMCB. Now I just chose one paper each, okay? So this is not, I don't want people to think that by only mentioning one paper in the JMCB that that means there is only one because that's not true for most of these people, okay? So thank you all for coming and joining us today. Each of you has 15 minutes. Mark is gonna speak first on securities regulation and bank regulation, then Anil on the Bank of England's Financial Policy Committee's policy framework, Till on stress testing, and finally Isabel on resolution. So with that, thank you. Well, good afternoon. I won't give you any history of the JMCB, but I would like to thank the organizers for asking me to be part of this program. It's been really nice to be here. One way to motivate, I try to figure out a way to motivate what I wanted to say and what I wanted to, the distinctions I wanted to draw. And this is what I came up with, that banks are the only firms permitted to sell their debt obligations to the general public without SEC disclosure requirements. So a bank is a firm that can sell obligations to the general public without going through the SEC. And that reflects, I think, some very important things about the way these two different regulatory environments interact. So up here, I've got the Fed, the SSM, the European National Prudential Regulators, the FSB, and I'm gonna contrast that with the Securities Exchange Commission, ESMA, and IOSCO, which is the International Organization of Security Commissioners. Basically, what the banking guys up top do is they protect depositors and they try to protect financial stability by assuring stable institutions. That's not the end of it, but that's very, very high up in their list of preferences is stable institutions that don't fail. That makes the deposit safe and that gives people a safe place to go, either people or corporations. The securities folks take a much different view of how you protect depositors or how you protect investors. And that is they disclose information. And they spend a lot of time saying either we have a structured way you need to disclose information or we're reading your prospectuses to make sure that you're not glossing over risks and you're accurately representing the kinds of securities you are offering for sale. And after that, the investors can fend for themselves is the idea. There's no judgment made about the quality of the assets. There's a long history in the United States with so-called blue sky laws or state level securities issuance laws. And some of those states did have opinions about whether a security was too risky but not at the national level in the US or I think in these other organizations. There are however some limits to what can be sold to different kinds of investors and I'll come back to that. So here's a list of five things that need to be in a regulatory system and I'll talk about a few of them. One is consumer protection and by consumer I don't mean just the man in the street or the retail investor. I mean people who buy services from these firms. The Fed is very into prudential oversight as is the ECB. We're gonna make deposits safe. We're gonna restrict asset investment opportunities. We're gonna make sure if you concentrate your portfolio it's not an over-concentration relative to capital. And we wanna make the firm safe and that'll make deposits safe so people don't have to worry. The SEC by contrast as I said doesn't restrict very much what can be sold but whatever is sold they make sure that there is adequate disclosure of the risks. Now the fourth one down here is systemic stability and the banking regulators worry a lot about that particularly after 2006 but as I was reading up on some of the history for my preparation here I kept coming across concerns about systemic stability back in the early 2000s so it wasn't invented during the crisis. And so the Fed, the bank regulators worry about fire sales and I think the SEC is a little bit like Alfred E. Newman in this regard, what me worry? They have an oversight apparatus that's based on the notion that securities are priced efficiently, that securities markets clear and that their primary responsibility is to make sure that those markets function well. And then failure resolution, we find quite a difference as well which is as you know the bank regulators are paying a lot of attention to orderly resolution. There's a lot of supervisory discretion in the mix and firm survival is an important concern. By contrast, the SEC is more than happy to worry about customers that is the people whose securities are being held by a broker but they don't worry very much beyond that and they certainly don't worry about the survivability of the firms per se as much as the banks do. So what are the financial products we're talking about here? One is liquidity transformation and a way to think about that for banks is sort of a diamond divvig model that banks are going to fund themselves short and lend long because of some sort of insurance idea. I think people have embellished that model not technically but interpretationally an awful lot but there's this notion that there is liquidity transformation and it's potentially dangerous. At the same time broker dealers or investment banks have long used repo agreements, short-term repo agreements to finance their portfolios and as long as their portfolios were made up of government bonds, this didn't cause any trouble at all but when they tried to push the envelope a little bit into their changing asset concentrations into investments in more and more illiquid securities then the liquidity transformation became problematic. And then liquidity transformation outside the regulated sector takes the place at least in two forms that is special purpose vehicles for securitization and maybe mutual funds and the FSB just this year changed their annual monitoring report from saying that they were looking at shadow banks which sort of implied that anything that looks like a bank ought to be regulated like a bank and this year they called it non-bank financial institutions so they might be backing off a little bit but there's a lot of pressure from the FSB in particular to on mutual funds, certain kinds of mutual funds and on special purpose entities which are nominally under the SEC's regulatory umbrella. So we've also got credit risk transformation. One of the interesting things about that is if you think of securitization what you may not know is that the way those things get financed those special purpose vehicles is through privately issued debt. So if you're gonna bring a security to market and you wanna sell it to everybody in the street you have to register it and that means reveal information and disclose adequately. If you're only going to approach so-called accredited investors or various kinds of specialized investors then you can sell that debt privately and the SEC basically gets a memo that says hey we're gonna sell some debt and that's the end of it. And to the extent that there is or could be maturity intermediation, liquidity transformation going on in there I think that's a really important thing that nobody's really watching because the data don't exist. All right so here are some mission statements. The SEC has three investor protection maintain fair and orderly markets efficient markets and support access to capital for the real sector. But this is a fellow writing from the inside the government the Congressional Research Service. The SEC is not primarily concerned with ensuring the safety and soundness of the firms it regulates but rather by protecting investors and so forth. So that's a really important distinction between the securities market regulators and the bank regulators. The Fed says and I don't know if... The Fed says and I don't know if this is actually their mission statement but their mission is to foster stability, integrity and efficiency in the monetary, financial and payment systems so as to promote optimal macroeconomic performance. It sort of sounds like Alan Greenspan wrote it. It's trying to cover all the bases without being too specific. But I stumbled on an OCC document where after the OCC Office of Comptroller of the Currency it said in quotes ensuring a safe and sound federal banking system for all Americans which seems to go back to the idea that the regulatory goal is to make those banks safe and sound. The FSB has an interesting mission statement. They'll address vulnerabilities affecting financial systems in the interest of global financial stability and by talking about vulnerabilities or potential vulnerabilities they have essentially left open a huge range of possibilities if they decide that they want to worry about them. Bank depositors of all sizes are protected by this prudential regulation. And bank depositors I think have self-selected into banks in large measure because they don't like the possibility of default. So they're kind of lexicographic on getting paid back and that's why when a bad thing happens to a bank the people who have chosen to be very averse to default they run quickly and I'm not sure it should be extrapolated to people out in the markets who are looking for risk exposures. So the people who are market investors are often looking for risk exposures and they don't worry about repayment certainty. I made the point about registered securities and private securities. So after I did the dual sort of dueling regulatory systems I asked myself, okay, so what? What difference does it make? And I think that it makes this difference. Since Gramleach-Bliley in 1999 and always previously with Universal Banks in Europe commercial banks in the US have gotten closer to broker dealers. And broker dealers have gotten closer to commercial banks too by taking on less and less liquid assets onto their balance sheets. And as they've moved together of course it's an old observation that you can have the same product produced in two different ways under two sets of regulations and that's where a lot of the conflict or potential conflict comes up between these two regulatory systems. And then in terms of the shadow banks so the non-bank financial institutions I just observed that it's perfectly possible particularly with liquidity requirements for the banks that an awful lot of liquidity transformation is gonna happen on the books of those special purpose vehicles. And again we don't have any window to speak of into the kind of maturity transformation those folks are doing in the aggregate. Now the SEC has clashed with the banking regulators in the past. In 1998 and then again in 2016 but more famously in 1998 the SEC looked at SunTrust and said you've been running up an awful lot of loan loss allowances lately. And SunTrust said yes we have. And the Fed said well that's good that's capital that's a good thing. And the SEC said it looks like income smoothing to us. It looks like you're not presenting good accounting information. And that fight went on for a couple of years until Congress essentially said why don't you guys play nice and SEC if you're gonna talk about this again at least give the Fed a heads up that you're gonna bring the subject up. So there the notion of transparency and the notion of what the accounting means is very important. Money funds. There was the prime reserve fund of course in 2008. There was a reform of the money market funds in 2010 that I think the SEC was pretty happy to do and I wasn't there at the time. They shortened up the permissible maturity differences between assets and liabilities and they tightened the credit quality requirements. Now neither of those things would have addressed the reserve fund problem because that was a highly rated commercial paper that went into these funds. That problem was a lack of diversification which isn't anywhere addressed there. In 2014 there was, and I was there for this one, in 2014 there was an effort to address the first mover advantage in mutual funds. So the idea being that there are strategic complementarities among the depositors at mutual funds or the investors at mutual funds and if assets are illiquid then there might be times when the asset value falls the net asset value doesn't fall appropriately and informed people will run and extract value from the people who stayed behind. So in 2014 the SEC eliminated dollar rounding or half penny rounding for the money market funds and required that the net asset value of the money funds be expressed to four decimal places. If you were a prime fund but not if you were a fund that owned only government bonds and so what they did and I think we were surprised when this happened was we basically killed the institutional prime money market business. They all changed to government bonds and they all took with them therefore they reduced the amount of liquidity transformation we have. In the recent FSB review of off bank of non bank institutions they've got a comment which you'll see frequently it says a key structural vulnerability from asset management activities is that open ended funds often offer short term liquidity to their investors not with standing that the liquidity of fund investments varies across different open ended funds. So they were getting all excited in particular about corporate bond funds and about leverage loan funds and they were pressing the SEC to regulate these things got to have some liquidity requirements and the SEC went out with some actual liquidity requirements and all of the mutual funds wrote back and said gosh that would be hard we don't need to do that and so the final rule said why don't you figure out what your policy is and write it down and make sure you tell your investors what your policy is toward liquidity. So it was as if the FSB the Fed wanted to force the SEC to do something but they were so ingrained in the sort of laissez faire nature of what they really wanted to do that they couldn't quite get there. The same things happening again with stress testing so I last year I spent some time talking to some people at the board about whether stress testing results should be more or less widely distributed than they are and I made the point that I thought that if you knew the stress test result was going to be publicly available as much of it is already then that would contribute to supervisors overcoming the temptation to tolerate bad situations. It would put a little bit of more pressure on the supervisors and the SEC would be very much in favor of that but the bank regulators were much less excited about it because their view is if I share bad information that reduces the amount of time that the bank and the regulators have together to make sure that the institution remains solvent. So a couple of points I've made here. There we go. Evolving institutional boundaries have complicated the interaction and increased the interaction between bank and securities regulators in Europe and in the States. Disclosure's effectiveness relies on efficient markets and doesn't allow very much for imperfect liquidity. So that the notion that simply knowing what the underlying risks are unless you know a lot about the liquidity of the markets you haven't really understood the risks and the Fed is I think concerned about that. I do think the place where I worry the thing that I would track if I could is the special purpose vehicles and what is the maturity structure of the liabilities they're selling in order to finance relatively illiquid assets that they hold. So that's something that I think is potentially very big or it will get very big again and I don't think we know anything about it. Thank you very much. Thank you. Let's see, do we have, I have to figure this out now. Go up and close this one, I think there's a new one. There it is, cash in. Okay, while they're getting started let me say a few introductory things. So I also want to thank the organizers for letting me participate in this conference. GMCB has been a great important journal in our field and it's an honor to be on the panel. What I'm going to talk about today are my and the most important word up there is my reflections on the FPC strategy. So I'm an external member of the financial policy committee of the Bank of England. I'm guessing half of you don't have any clue what it is so I'm going to try and tell you a little bit about it but I'm not speaking for the committee right now so these are my own reactions. There's a full speech that's posted on the Bank of England's website and if there's anybody from the media here or anybody with a Twitter account I'd ask you just to quote from the speech not from the shoot from the hip version I'm going to give here because I won't have time to go through all of what's in the speech. So I'm going to customize what I'm going to talk about to the audience to try to reflect the fact that I've got a rare chance to speak in front of a bunch of very smart economists who may be able to help us with some of what we're up to. So I'm going to cover five things. I'll start with a little bit of description of what the FPC is and then I'll give you a description of how we try to make this operational. That's going to include my kind of theory of the case as to what we're up to. Then I'll describe some of the tools that we've got and I'm going to specialize beyond or down below what's in the speech to just focus on a few of the tools that you may be more surprised by, especially if you're an American. And then I'll close with some lessons and open questions. Okay, so what is this FPC? Probably you've heard of the Monetary Policy Committee of the Bank of England. That's the inflation targeting policy body that has been copied in many, many other countries. The FPC takes its structure in a bit from the form of the MPC, but with some important differences. So like the MPC, there are insiders and external members. There's five external members. I'm one of them, Don Cohn, who many of you would know is another. There's three other British private sector people that are on our committee as well. The CEO of the conduct regulator is there and we have a member of the Treasury that attends our meetings. We only meet four times a year and twice a year we write a financial stability report and we try to operate by consensus. We can make recommendations and they've done a lot of that. Before I was on the committee, there was a furious round of recommending and advising to other bodies as they were setting up the capital framework. So there've been 47 recommendations and unlike the stuff that you see in the US with the FSOC, every single one of these has been agreed. There haven't been any cases where there's been a recommendation that's been ignored. So that's the background. Now, what's our remit? So this is the copied from the act that sets up the FPC. The tagline that I think we use, that's the version of the OCC one, is we wanna be able to make sure that the financial system will be able to serve UK households and businesses in bad times as well as good. So that's how I'd explain what I'm doing to somebody that's not an economist. Now, the specific systemic risks that were charged with attending to are listed there in the bullets. The first two probably wouldn't surprise you. I mean, obviously you'd be looking at connections between financial institutions and probably you'd have to do something about markets. I'm gonna focus most of my discussion today with you on that third one, which is unsustainable levels of leveraged debt or credit growth. So that's written into the law and it informs a lot of what we actually do. Okay, so how do we actually make this operational? I wish I had equations and a worked out dynamic general equilibrium model. We don't, but we have to carry on. So this picture is my implicit model of how I think we're proceeding, okay? And it's built on three premises. The first is that you need savers and borrowers working together to channel funds to support the real economy, okay? The second is that the savers can put their money through two pipes to get to the borrowers, either banks or capital markets. I say a little bit about shadow banks in the speech, but let's simplify to this. And then third thing is that there's feedback from the real economy, both to the borrowers and the banks and the markets. So the way I think of what we're doing is this is kind of the way we think the world works and we're trying to watch all parts of this. So we're trying to make sure that the credit supply function doesn't get disrupted either because there's problems in the banking system or problems with capital markets or because there's problems with the borrowers. And the borrowers bit is the part that I think may be less familiar and less obvious. There's nothing like this in the United States. There's nobody that's in charge of worrying about borrow resilience and the role that that might play in financial stability. So this is pretty out there. And I would say, I think it's a fair assessment to say that FPC is probably the most muscular financial stability regulator in the world. We've got more tools and more authority than any regulator I know of, at least in, you know, let's say the G8 countries. So if we're successful, maybe we'll turn out to be a model for what other countries will evolve towards. So it'd be great if anybody here wants to try to work out a model that embeds all of this. I think that that will come at some point. For now, we're gonna have to do with the pictures. Okay, so here's an example of some of our tools. And I've sorted them into the three prongs that I said underlined the model. So there's the market aspects. There's the bank resilience aspects. And then I've put in bold and yellow the borrow resilience ones. I'll talk about those in a second. This is the first time I'm gonna talk in public in a while where I haven't had to say anything about Brexit there. We spend an immense amount of time on Brexit, but that's hopefully one-off and not something we'll be remembered for. So what we do though on borrow resilience, I think you might find interesting. So we have taken two decisions that affect the housing market in a pretty profound way. Banks as they issue new loans have a 15% limit on the flow of new mortgages they can give to people that have a loan to income ratio above four and a half. Okay, I'm gonna repeat this in a minute, but there's a restriction on the flow limit. There's also a stress test for the borrowers. So a borrower taking out the mortgage has to be checked to see what would happen to their ability to pay where interest rates to rise by a certain amount. Right now the rule says 300 basis points. Now that's a pretty profound intervention in the housing market. I mean, I'd say in the United States there'd be pitchforks out if you said you're gonna do this kind of stuff. In the UK this hasn't been controversial. Perhaps because there have been so many financial cycles in the UK where housing was in the middle of it. So we've done this, it's been reaffirmed. You can, I have a footnote in the speech that cites some of the op-eds saying, hey, good for the Bank of England for tackling this, but it's really not something that's standard in many other places. So in the remarks that are posted, I go through some of the other things we've done, but let me just focus on the borrow resilience part because I think that's kind of interesting. So here's one of the, I would say, features not bugs of this policy. So this shows you the proportion of new mortgage lending by loan-to-income rates and the loan-to-income rate of four and a half, which is what's in the law, you can see it's been pretty flat. So you can have up to 15%, the banks aren't kind of close to that, but what has happened is a lot of bunching below four and a half. And I'll explain why I think the theory of the case is, that's just fine. That was the intended consequence of this policy is to try to stop people from getting over their skis. Okay, and the second thing that you can look at is what I think of as the state variable that underlies what it is we're trying to do with this policy. So I think of one of the lessons from the last crisis was overly indebted borrowers may cut back on their spending. The individual bank that's making a loan or the borrower that's taking it out will recognize that that's gonna be a feature of what happens and they wanna check that the borrower is not gonna default. Now in the UK, mortgages come with recourse. So people don't actually default. Credit losses for UK mortgages were small. However, the individual bank and the individual borrower don't take into account any of the macroeconomic effects of a bunch of people becoming over indebted and cutting their spending. So imagine that you have the distribution of debt service ratios. So here I have in mind how much is your mortgage payment relative to your income. And I'm worried about the right tail of that distribution. And I'm worried about what the right tail is and how many people are there and could wind up there. So I think of our policies as saying if too many people wind up in the right tail then that's something that's gonna lead to this deleveraging and it's gonna spill back into the macroeconomy and thwart our ability to serve the financial sector serving in good times as well as bad. So that's the logic for what we're doing. There's nothing that says you could only think about this in terms of the household sector. You could do a corporate sector version two perhaps. But this is what we've done. Now, how do we take such decisions? Again, not speaking for the full committee but my attempt to kind of rationalize what we do. I think the first thing is you've gotta figure out what the externality is that you're trying to fix. You don't wanna just regulate for the sake of regulating. So for instance, in this case, why are we doing this? Well, we think there's this aggregate demand effect and the individual borrowers and lenders won't attend to it. I think the second thing that we spent a lot of time doing is cost benefit. So my colleagues talk all the time about we could have the stability of the graveyard. If you didn't want any amplification coming out of housing, you could just say you can't get a loan where you're loaned income ratios above four. Pick whatever number you wanted. You could do that but the consequences would be dire. There's lots of people that probably should be able to borrow that and pose no risk and so on. So you have to use cost benefit. And then finally, you wanna pick a tool that's kind of appropriate for what you're going to do. Here, we often struggle with whether or not we're into having a time varying tool or something that's more structural. The FPC does not see itself in the business of trying to eliminate all cycles. That's again, that's not our job. We're not trying to outlaw recessions or do anything like that. We're trying instead to do this cost benefit and think about things that often can be done structurally. So a lot of our capital regime is a structural stuff but then we have other time varying tools. The counter cyclical capital buffer is one that every quarter we have to vote and decide on whether there should be an add-on to the level of capital that depends on what the situation is in the banking system as we go. So the housing tools are kind of structural in the sense that I think of the number of people we want to have out in that right tail of the debt service ratio, that's something we take as kind of a fixed objective. Whether or not adding 300 basis points to the level of interest rates is the right way to think about it as something that you might reconsider periodically. So you can kind of think of the jump variable that would change the number of people in the right-hand tail of the debt service ratio could come from either a fall in house prices, okay, that that would prevent somebody from perhaps being able to refinance and the interest rate goes up, or the interest rate could go up just because people are on adjustable rates and the policy rate goes up, or you could lose your job and see your income plummet. So think of trying to model all of those things and having a fixed objective as to the fatness of that tail but maybe having to adjust your policies as time goes. Okay, so I'm gonna close with two things. First, some of the lessons that we've learned and then some open questions. So in terms of the lessons, I think number one, first and foremost, we should be pretty humble. We've been at this now seven years and we had a two-year interim committee. I think it would be super premature to declare victory. Who knows whether the current period is indicative of what we're gonna face. There's still a lot of recent memory of the crisis and it took a long time for the economy to dig out. So we haven't probably seen how this committee and the structure would function in a world where risk had built up again and things had become frothy. So that's one lesson. Second lesson we spend a lot of time trying to balance is on the one hand, trying to be predictable and transparent. It's pro-financial stability if markets understand what it is that you're trying to accomplish. So an example of this is our stress tests become more severe as the economic cycle progresses. So if we go another year and growing even at trend we'll have the drops in the levels of certain variables be bigger. The fact that the banks know that that's the way that the stress test operates gives them an incentive to have a prudent dividend policy and a prudent payout stock repurchase policy because they know that they're gonna face a test like that. So there's an example of trying to be predictable where it's at the same time, you need to be forward-looking and sometimes forward-looking is gonna be taking judgments that people may not be able to completely understand. So you gotta balance that. Third thing is the committee needs a portfolio of skills. You would not want five academic economists being the external members of our committee. It would be a way narrower perspective than you get having, for instance, somebody who was the CEO of a major global financial institution in the room. You get way more cross-fertilization in a way that I think lots of people have anchored in on the way the Monetary Policy Committee works and for that having specialized knowledge in economics is a huge, huge advantage. Here, I think you need a diverse set. Then finally, when this thing was set up, there was all kinds of questions about whether it could be a model for committees elsewhere. There were concerns that there'd be fights across the committees, maybe the Monetary Policy Committee be stepping on the gas and you're stepping on the brakes or maybe you wouldn't share information. Maybe the public would resent your actions. None of that stuff has happened. So for people that say you can't make this work, I would say in that sense, maybe we have learned some pretty good lessons. A big part of it is having so many overlapping people on the policy. There's four deputy governors of the Bank of England and all four of them are on, well, there's at least four members of the BOE. They're on each of the three main committees, the Micro, Pru, the Monetary One, and the Financial Policy One. But that part's worked well. Then let me close with just three challenges. First, it's much harder to tell the public what you're up to if you're trying to deliver financial stability. You're talking about tail risk stuff. Most of the time, things aren't going on. It's not like inflation where you can see it coming up and going down and everybody kind of understands where it is. People probably have no idea exactly how much financial risk is present in any given moment and you're trying to tell them, don't worry, we've got this covered. So that's hard. The reason I asked Vice Chairman Quarles, the question I did earlier, is what's the tolerance ratio that a committee like this should set? Imagine you thought that the native underlying risk in the economy would generate a crisis once every 20 years. We have a secondary objective that says we're supposed to try to support economic growth but it doesn't tell us where to land on that trade off so should we be trying to move to one in 50? If we were successful, how would you know it? I mean, you do Monte Carlo's of how long it would take you to convince somebody if they were just a Bayesian to figure out that you actually change this probability. And then one thing that I'm particularly keen on is coming up with a summary statistic for financial stability. The joke I make is, imagine you were doing inflation targeting but you hadn't agreed on a deflator. Okay, well we're making financial stability policy but there's no summary statistic of what financial stability is. There's sometimes financial conditions index but those are not exactly translated into what you wanna worry about. There's a lot of buzz around GDP at risk as an idea that's being studied, people at the bank are working on that and I think we'll see if that leads us anywhere but I think those are three questions that'll have to be wrestled with if we're gonna progress this agenda. Thank you. Thank you very much. Till? So I'm living my own private stress test here which is to follow as speaker, Anil Kashyap, especially after I learned that he's in the same pantheon as Clint Eastwood. That's gonna be a tough act to follow. So first let me thank the organizers and the Journal of Money Criterion Banking and the ECB for inviting me to this conference. To me it feels like it's been sort of stress testing all the time, all day, certainly since 2008, the summer of 2008 when I was at the New York Fed working on liquidity stress testing of the then four remaining investment banks. So let me tell you what I'm gonna talk to you about today. I'm gonna start with a quick description of how we thought about capital adequacy before the crisis and then the role that the crisis played in sort of evoking stress testing as a way to get out. I wanna spend a bit of time talking about what makes a good stress scenario. Anil already gave us a little hint about what we might think about in designing a stress scenario and then how stress testing has influenced about how we think about capital adequacy today. And then I'll wrap up with some thoughts about where we might go from here. So at this stage, banks from economies representing about half of world GDP are subject to regular supervisory stress testing. Stress testing has become important and for some regulatory regimes actually the most important tool for assessing capital adequacy. So look, stress testing is not new. It's really one of the oldest forms of risk management but how it's done in the role that it plays is actually quite novel. So pre-crisis and therefore pre-bosal, I think we're on three now, pre-bosal three, the approach to thinking about capital adequacy was the sort of idea of a single stage calibration really to a default point. So think about this as a value at risk approach and to do this what you, if you tell me the desired probability of survival or one minus the desired probability of default, you then tell me what the evaluation or assessment horizon is and you give me the portfolio characteristics, I can go and figure out how much capital do you need to hold. The focus really was very much on the balance sheet really mostly on the assets and it was mostly static. So it was really kind of a one hit approach to capital adequacy. The bespoke version of this was called economic capital and the regulatory version of this was called regulatory capital through the Basel Accord and both result in a risk waiting essentially, a risk waiting just of the assets though they might differ in the parameterization of this approach. So note there are big and legitimate sources of disagreement on many of the parameters especially about the correlations. So think about this as the risk dependent structure but also about parameters like probability of default and loss given default. So the disagreement on correlations really matters and it matters not just because we clearly got them wrong in a financial crisis. We essentially thought that they were lower than they turned out to be but also because correlations are essentially reduced form representations of both a joint distribution of the state space and also the sensitivities of assets like loans to that state space. Now the core idea of tying assets, asset value dynamics to default risk we owe to Robert Merton and this still forms the foundation of modern credit risk analysis today. And credit is and remains the dominant risk in banks. So the next step of taking it from asset to a portfolio we owe to old Rich Vasecek and then Michael Gordy helped build the bridge to what essentially became the Vossel II Accord. These ideas are also behind my own research in this area and which was sparked actually by the Asia Financial Crisis in 97 and then the Russian Bond Default in LTCM in 1998. In a series of papers with Hash and Pesseron and various co-authors we essentially developed a vector error correction model of the state space and then bolted it onto a credit portfolio model. And the idea was to look at the response to shock. So understanding resilience to things like GDP, inflation, the interest rate and equity prices. We cared a lot about capturing the interconnectedness of business cycles and financial markets across the globe. The potential for and the limits of diversification was one of the things that we really wanted to pin down. The paper appeared in the GMCB in 2006 which is really my excuse for talking about it here. So let's fast forward to 2007 and 2008. So banks that looked, well, they looked really well capitalized by the existing metrics. They were starting to experience runs and some of them needed government bailouts, some of them even failed. The problems weren't limited just to having a poor risk assessment. So the idea of having poor risk weights. But also to weak forms of capital. So not enough common equity. So regulators really needed to do something different. They needed to do something big. And importantly, they needed to show the results and how they got to those results. And to do that, that was just to regain market confidence and also regain confidence in the banks. Now, it's not just regaining confidence in the banks. It was also regaining confidence in the banking regulators and supervisors. So, you know, we were in this awkward situation where by the metrics that the regulators were using, everything was, well, fine. But of course it was not fine at all. And so this is where stress testing really entered the picture. The scenarios had to be easy to understand. They had to be credibly severe. You needed a government backstop in the case private capital markets weren't able to support the raising of required capital for the banks. Not every sovereign had this luxury. Importantly, regulators developed their own models to project losses and profits. Now this in my view is a really important, extremely important part of the stress testing pictures for the supervisory authorities to develop their own perspective because it won't surprise you that banks will tend to be slightly more optimistic about how they will fare in a crisis situation than the supervisors would. The results actually produced genuinely new information about bank health and asset quality. And you need that some of this new information about banks because a lot of the banks were near that kink that Gary described this morning. And you wanted to find out which banks were to the left of that kink, likely, and which banks were to the right or far to the right. And what would it take to get the banks back to the far, to the right of that kink? So the information in the 2009 exercise in the US told us that 10 of the 19 banks stress tested needed 75 billion in capital. And it was a way of really transforming uncertainty into risk. So why do we still, what was it about stress testing that made it so appealing? Well, I think there's a couple of things I'd like to go through. So the first is it's really tangible, right? If I describe to you an economic downturn, unemployment shooting up, house prices declining, equity prices falling, if I describe this to you, this is something very tangible and concrete. And it's something easy to understand and imagine. It forces important and constructive discussions about among bankers, but also amongst policymakers. So let's go through some of these things. The first and probably most important ones is what actually are my risks and my vulnerabilities? So from the bank's perspective, the risks that come about from the products I sell, the clients I serve, the markets I operate in and so on. And from the regulator, questions that I need to ask myself are, well, if I design a stress test, which kinds of banks do I actually target? Because there's a whole wide variety of banks, especially if I just have one or sometimes two scenarios to play with. What states of the world then would expose these risks and probe these vulnerabilities? And third, how severe should I make this scenario? Now, one of the questions that I need to ask before settling down on this sort of scenario severity and wrapping up the capital picture is, how strong do I wanna make my bank or my banking system to be when it comes out of the financial storm? So these two questions, the scenario severity and the post-stress capital ratio, they speak critically to the risk appetite of the bank and also to the risk appetite of the regulator. And they need to be decided and pinned down together in order to establish how safe you actually want that banking system to be. Now, I'm not gonna comment about what the impacts are on welfare and trade-offs. I don't really have a view on the optimal number of bank failures and the optimal number of financial crises. But I do wanna point out that whatever the answer is, it has to essentially navigate the Silla and Corriptus of the scenario severity and the post-stress capital levels. So what else do we love about stress testing? Well, properly applied, it forces a careful and comprehensive understanding of the bank's financial dynamics. Since the mapping of the scenarios to the bank's business really requires a projection of the balance sheet and the income statement conditional on the chosen scenarios. And it allows us to explore loss dynamics but also revenue dynamics. So loss emergence and loss realization is actually quite different across asset classes. So for instance, credit card loss is manifest really quickly. Within a month or two, you know that something is up. Commercial rule of state lending on the other hand can take a long time to manifest one to three years or even longer. And then in the other extreme trading, those losses you see right away. Let me show you a couple of what the stress scenarios see what they look like. And more importantly, how they've evolved over time. So what I'm showing you here are two variables, the unemployment rate and real GDP growth, which are two variables that have been in the US stress testing program from the very beginning back in 2009. So these paths are jointly determined. So they come from some view of what the joint distribution is going to look like and presumably there's some formal macroeconometric model working in the background. Now there's two things I want you to notice about these pictures. So the first thing is that these look awfully similar year after year after year. So we've gotten really, really good withstanding precisely this pattern of stresses, at least in the US banking systems. But the second thing I would like you to notice is that these are pretty severe shocks, right? So the black line is what actually has happened. That is history, that's the actual realizations. And these projected shocks are indeed extreme. And we also see something that Anil described which is an explicit policy in the FPC, but also as part of the policy statement of the Federal Reserve and its design of the scenarios is that they get progressively more severe, right? So in the unemployment rate, essentially the policy targets a 10% unemployment rate as it gets better and better, you have to hit the shocks harder and harder and the banks know this and have to be prepared. Now, this exposes really one of the deep challenges of scenario design in my view. Macro-econometric models are based on equilibrium relationships. But for stress testing, we're really interested in off-equilibrium relationships. And in fact, we might be interested in very far and increasingly far off-equilibrium relationships. And off-equilibrium, I'm sorry, states of the world. So the distance between severe yet plausible, which is the sort of catchphrase in stress testing and extreme and incoherent, I think is an important point of debate about where, how far out we can go before we fall off the edge into incoherency and extremities. So arguably the really interesting scenarios are not the ones where you push a couple of risk factors out to the extremes. But they are the ones that expose and tilt and shift just so the dependent structure, the joint distribution that underlies the generation of these risk factors. And it's partly because one of the things you'd like to do is to expose what traders call natural hedges. So two examples of this are, and essentially what those are, they're kind of a belief that correlations are gonna persist and persist in your favor. So two examples of this are convergence trades that were put on by LTCN, that essentially counted on a correlation just being very persistent and steady. And the other one is the low correlations that we all thought existed amongst credit assets for securitizations. So I wanna now return to the tale of Ulysses and this problem of scenario of severity and post-dress capital thresholds and describe to you how we think about capital adequacy therefore today. And we've essentially boiled this down to a two-hit problem. So the first, we have to think about the severity of the scenario. So the severity of the scenario describes the first hit and the consumption that comes with that is described is essentially represented by this yellow box. Second is the post-dress capital goal. That is the amount of capital that I need after it's all done, so that's that red box. Now, how do I think about that red box? How do I think about that post-dress capital threshold? Well, I need to have enough to convince creditors after it's all done that I'm still a viable financial institution. So what do I need to convince the market that I am that? And whatever that is, it's the number that I should target. So to make this concrete, in the last year's US stress test, the amount of capital that the three dozen banks that were subject to a stress test held post-stress was more than the amount of capital that was in the entire US banking system year in 2006, so right before the financial crisis. So I'd say, is that enough? I don't know, but I'm a little bit more comfortable knowing about this result. So is Basel five then gonna be a three-hit model? So in other words, I survive one fine, but if I survive two in close succession, then I'm gonna be on razor-thin cushions. Having survived one, the market knows this, how will they react? We'll find out. So let me discuss just two points before wrapping up. So first, as peacetime stress testing matures, there's less and less ambiguity about banks' capital adequacy. Supervisors are making use of concurrent stress testing for evaluating also qualitative aspects of banks' risk management governance and control processes. So I need you to be sort of really put yourself into some of you aren't supervisors shoes, and others of you who aren't try, you're at a terrible informational disadvantage vis-a-vis the banks, and you have fewer resources. So you're gonna have to leverage that one bit of informational advantages that you have, which is the ability to look across banks. And that's part of what the power of these concurrent stress testing regimes are, is the ability to leverage precisely that. So what does that picture look like today? What does the risk picture look like today? What are the risks that worry and the preoccupy the bankers? After years and years of obsessing about market and credit risk, and market and credit losses, what's front of mind really now are the plethora of non-financial risks. And cyber risk is at the top of the top 10 list of pretty much every banker, asset manager, and head of an insurance company that I talk to. Cyber, technology, political risk, and more recently climate risk are essentially those are the topics that come up when I talk to chief risk officers. So we're in financial peacetime now, which means we're in that state of the world before the next crisis. What should we prepare for? What should be the stress scenarios that we should look at today? If you were a supervisor, what would you think about designing? Now in peacetime, the supervisors by virtue of seeing every bank's risk inventory, they are really in a critical position to be able to aggregate across those risk inventories to see patterns, to see some of the common things. Now we have a pretty good idea of how we're all gonna react to GDP shocks and so on, but how would we react? How would the banks react to a coordinated cyber attack? Are there common third parties that most banks use like data vendors or cloud providers? Are there common risk and control approaches that have particular vulnerabilities? Now the supervisors with good intentions have encouraged common approaches to risk management and thus effectively reduced the variety, the diversity of approaches to doing risk management, and might have made the system as a result slightly less resilient to shocks. So after a while of being in financial peacetime, banks should be I think in pretty good shape micro-preventually, and the importance of systemic risk identification becomes especially high. And I just hope that we have the imagination, and perhaps some of you in this room will have the imagination to conjure up severe but plausible not extreme and science fiction-y, but severe but plausible interesting stress scenarios that we should throw at the banks now. On that cheery note. Thank you very much. Let's see, I should try and do, I should be your assistant. Oh, God. Okay. And now, whoop, that's supposed to be the last one. Okay, there we go. Excellent, thank you very much. So first I would like to thank the organizers for having me on this panel. It's a great pleasure and honored to be here. And of course, it's always a difficult task to talk at the end of a long and exhausting day, but I have a very exciting topic, and it's a topic that we all discussed, that we have all discussed for a long time. And I actually think it's as relevant as it was directly after the crisis. So I'm very happy that I can talk about a bank resolvability and I want to ask the questions whether banks are finally resolvable. And so after the massive bank bailouts in the global financial crisis, the policy makers in basically all major economies made a big promise, namely, that this would never happen again. They never again wanted to bail out banks. And so it was one major goal of the post-crisis reforms that these explicit and implicit government guarantees to banks should be phased out. And we all know that such guarantees are quite harmful. First, because they distort risk-taking decisions. Second, because they distort competition. And third, of course, because they impose a high burden on the tax payer. But it's a very tricky issue because we are facing here a severe time inconsistency problem. And so when there is a threat of a systemic crisis, again, there is, of course, a strong temptation to bail out banks. And therefore it's very important that these promises are actually backed by a sound legal and institutional structure. So let's go back to the time after the crisis. So if we look at what happened to bail out expectations directly after the crisis, it's actually quite disappointing. So what I'm showing here in the graph are the so-called support ratings issued by Fitch. And what we see is that directly after the crisis, there was a sharp increase in these Fitch support ratings for the GZEPs, for the global, systemically important banks. And actually this increase occurred in two steps. So first there was a shift at the end of 2008, and this was, of course, due to the large-scale rescue operations. And then there was another one at the end of 2011, which was when the Financial Stability Board introduced the GZEP statutes. And so then basically these Fitch support ratings were at the highest possible level for the GZEPs. If we look at the non-GZEPs, so the smaller, less systemic irrelevant banks, there was a slight decrease after 2010. And I think it should be mentioned that if you compare different regions of the world, I mean you would see that actually there is a pretty big difference between the support ratings of the smaller banks between the US and let's say Europe. In fact, so the bailout expectations in the US were much smaller, and this, of course, consistent with the fact that in the US quite a few banks failed during the crisis, were allowed to fail during the crisis, and the same was not true in Europe. So I also have some good news. So in a paper joined with Alexander Schäfer and Beatriz Vida de Mauro in 2015, we analyzed the early reforms that were taking place at the national level. So here these two graphs, they show the announcement of the Falker Rule in the US and the introduction or the announcement of the German Bank Restructuring Law in Germany. And what we see here is actually something that we should like, namely that these reforms led to a divergence in the CDS spread of the systemic and the non-systemic banks. So this is consistent with the reduction of bailout expectations, and this is, in a sense, good news. So of course, the most important set of reforms that we had after the crisis was the introduction of the bank resolution regimes, and I'm mostly going to talk about what happened in Europe, because I know that much better. I mean, what I can say is that the US, in principle, were already much more advanced at the time than Europe was. So one reason why there were these widespread bailouts in the crisis was that there was a lack of a special bank resolution regime. So one basically had to rely on the general corporate insolvency regimes, and as we know, these are too slow to deal with banks when the banks are facing liquidity problem. The solution has to be found very quickly. And also they fail to take into account systemic repercussions. And therefore, there's a wide agreement that we need special bank resolution procedures, and these have basically two goals. So the first is that there is the possibility of an orderly resolution of banks without destabilizing the remaining financial system. So it's not so much that we want to guarantee the survival of the bank, but we want to make sure that a bank can be resolved without destabilizing the financial system. And the second is that creditors are supposed to be bailed in in order to avoid the need of using taxpayers' money. Of course, if we look at these rules, we see that, I mean, in that sense, the word resolution is a bit misleading. Many resolution procedures actually do not lead to a closure of banks, but very often the bank in some way or the other survives. And even before the new resolution regime was introduced in Europe, there were actually some bail-ins. And in another paper with the same set of co-authors, we analyzed these bail-in events in an event study, and we found something which is similar to what Elu alluded to today. So, I mean, the most important bail-in event before the introduction of the regime was the case of Cyprus. So Cyprus was a big bail-in event. You probably all remember the discussion about it. And so what actually happened is that after the announcement of the bail-in, there was an increase in CDS spreads, which would indicate that there is increased market discipline or a decrease in bail-out expectations. But what we found in that paper, and which I think is very interesting, is that there were sharp differences between the banks in the crisis countries and in the non-crisis countries. And then we showed that the change in CDS spreads could actually be related to the fiscal capacity of countries. So that basically means that the bail-out expectations were reduced mainly for banks in those countries that had a very low fiscal capacity, whereas for a country like, let's say, Germany, the effect was actually much smaller. So then, of course, we had the BRD and the single resolution mechanism, very important achievements. So there was a definition of a liability cascade. Then at the same time, there was an international agreement on the requirements for bail-in and bail-debt. So there's the TLAC, then there's also the MRAIL. In Europe, there's now the resolution fund that can be tapped if there is a bail-in of at least 8% of total liabilities, including equity. And then just at the end of last year, there was also the agreement on the fiscal backstop to the single resolution fund, which is to be introduced by 2023. So interestingly, if you look again at the support ratings, what you see is that Fitch downgraded banks quite sharply in the middle of 2015. And this is, of course, something that we would like to see. What's also interesting is that there was one group of banks that was not downgraded so strongly, which is the state-owned banks. And that is, of course, for example, the German Landesbanking, and I think it's fair to say that Fitch there had rational expectations. If we look at what's happening in Germany with the Landesbanking, we see that the probability of being bailed out is still pretty high. And, but of course, I think what they exaggerated, and I haven't looked at the current ratings. I wanted to do that, but I didn't have access to the data, is I mean, what they actually did is that for the private banks, they reduced their expectation of a bailout very strongly. And I think there maybe they exaggerated a bit because I wouldn't think that a bank like, let's say, Deutsche Bank has a bailout probability of zero. That's very unlikely. So what happened after the introduction of the SRM? So what you see here is on the left-hand side, you see five-year senior CDS spreads. On the right-hand side, you see the difference between subordinated and senior CDS spreads. And I don't know whether you remember, but in the beginning of 2016, there were turbulences in the market. And so what had happened is there was a fear of a global slowdown related to China. There were fears that certain banks would not pay their coupons on the coup coup bonds. And so there was actually, there was quite some uncertainty in the markets. And what we see is that there is actually a sharp widening of the spread differentials. And this is exactly what we would hope, right? So there was a return, somehow a return of market discipline. At the same time, we experienced a strong pushback. And this is that people were arguing, look at this, what the bail-in is doing, it's destabilizing our financial system, right? By the way, you also see the Brexit vote in that graph that's also visible there, which again added some uncertainty. So in 2017, the new regime was tested for the first time. And we had one story that is typically described as a success story, Banco Popular Español, which was resolved not only over the weekend, but overnight. And this didn't have any systemic effects. I mean, there's also some discussion about what was going on, but overall I think one can still say it's a success story. And then there were several cases that you would probably not describe as a success story. There was Montede Pasque di Siena, which received a precautionary recapitalization. And there were the two Venetian banks, which were resolved at national level, thereby somehow circumventing the European rules. I mean, there were some good points to all these cases. I mean, there was a bail-in of junior debt in all cases, which already is some progress. And also, I mean, there was an exit of several problem banks with high levels of NPLs, which is, in a sense, good news. But there was no bail-in of senior debt, even though that would have been required. There was, again, a substantial use of taxpayer money. And as I said, the European rules were somehow circumvented. So if we look more carefully at these different cases, and here I'm only showing you the case of Montede Pasque di Siena, the question is, of course, whether this kind of reduced the credibility of the regime. And so, again, we were looking at the CDS spreads. And only for the case of Montede Pasque di Siena, we found that the senior CDS spreads decreased more than the junior CDS spreads, but only for Italian banks. And this would be consistent with the idea that bailout expectations for senior bank debt actually increased. But what is interesting is that we do not see the same in other European countries. So there was no spillover to other European countries. So in a sense, Italy was considered a special case. And I think this is good news for the SRM because it shows us that the system has some kind of robustness. So, but anyhow, we see that bailing is always difficult. And remember, these cases that I just talked about, these were taking place in calm times, in a time where the economy was doing well, where there was clearly no systemic crisis whatsoever. And so then the question is, would this also work in a systemic crisis? And this is actually what Ellen referred to earlier today. And I have a new paper with Thorsten Beck and Dejan Radeff, where we are asking the question whether bail-in can work in a systemic crisis. And what we are doing here is, so we are looking at different events, systemic events, and non-systemic events. Here what you're seeing is the Lehman bankruptcy, so a systemic event. And we are looking at the change in systemic risk measured by the Delta Cova. And we're comparing two groups of countries, so at least in these pictures, and the regression, it's continuous, those with the relatively advanced resolution regime and those with the less advanced resolution regimes. And we constructed the data based on the FSB's key attributes. And what you are seeing, so the right-hand side shows you the difference of the two graphs on the left-hand side. And what you see is that actually there was a much stronger increase in systemic risk in countries that had the more advanced resolution regime. So that means that if we believe that, that the bank resolution procedures indeed may have a destabilizing effect in crisis times. And this shows us that there is, of course, a tension between market discipline and financial stability. So let me give you some tentative conclusions. So I think it will always be difficult to escape this time inconsistency problem of the bank rescues. And this is, of course, especially true for the large complex multinational banks. And it is especially true in the middle of a systemic crisis. And therefore, of course, more needs to be done. One thing that has to be done is, of course, that we still have to improve the resolution regimes. At the moment, especially the European regime, has many loopholes. And these loopholes can be used even in normal times. And I think the idea that we will always be strict and never again bailout banks is completely illusionary. So we need something like a systemic risk exception. But this has to be defined very strictly. That means that in the normal cases, we're very strict. We don't bail out. And if there's really, truly a systemic crisis, then there are probably one, I mean, one has to find a solution. And I think it's unrealistic to think that we will be able to stick to every single letter of the rule. Of course, higher capital is beneficial, especially for the systemic banks. And especially if it is non-risk based, I've always been a bit critical about the risk rates. And I believe that the bail-inable debt is a relatively poor substitute for capital, especially for the reason that bail-in often is difficult. I think we have to reconsider the structural reforms in banking. I mean, something has happened in the US and in the UK, but not in the euro area. And I think we have to reconsider that. If we look at a bank like Deutsche Bank, of course, one of the issues is that they have this investment banking division, which functions very badly and which also spills over to the rest of the bank. And finally, I think a problem that is completely not solved is the question, how we resolve the multinational banks. So maybe the best way to have it is a single point of entry approach. But I think this is actually very difficult to agree on. And of course, then in the end, I want to end with a point that is very dear to my heart, which you should not start to create additional national champions now in banking. And as you know, this is happening at the moment in Germany. It worries me a lot. I think it's unbelievable that, given what policymakers have been telling us 10 years ago, that this is happening. But as you've probably seen in the FT, I think that a merger between Deutsche Bank and Komadsbank is a very bad idea. And I'm happy to discuss this more. And thank you very much for your attention. Thank you. Thank you very much. So I guess I'll summarize very, very quickly. So we have a problem where we need to try and find a way for banking and security regulators to play nice. But I'm not sure how we do that. We do need a coherent framework within which we can understand what we're doing holistically at a macroeconomic level. But I don't think we've made a lot of progress quite yet. This one really bugs me. Effective stress testing seems to require that I have daily nightmares and that my nightmares be increasingly bad, or at least uncorrelated, so I'm concerned. And then I guess the question of whether we have recovery and resolution frameworks for global banks is an open one. Aside from the very end, which I like quite a lot, because we have to stop thinking about banks the way we think about our national football teams. It's OK for them to be champions, I think, but not for our banks.