 So let us start. OK, it is my duty to introduce the winners of the ESRB research prize in memory of Ike van der Berg. Ike was, in a way, a colleague of us. I mean, she was a member of the ESRB Advisory Scientific Committee from 2011 to 2014. She was also a member of the European Parliament until 2009. And unfortunately, she died towards the end of her mandate as a member of the Advisory Scientific Committee. And she was a great colleague to have on the Advisory Scientific Committee. And this prize is dedicated to her memory. And since Ike very much held very high the notion that finance should serve society and not the opposite, this prize was created by the Advisory Scientific Committee in this spirit. And so it's dedicated to excellent researchers who do work on systemic risk on macro prudential policy and also to consumer protection in finance. And in fact, last year, which was the first edition of this prize, the prize was won by Celerier and Ballet with an excellent paper on retail products complexity and, therefore, consumer protection issues. And this year, the winners of the prize are Sergei Czernenko. So these are execual winners. Sergei Czernenko from Ohio State with a paper joined with Addis Sundaram from Harvard University. And Matthias Effing with a paper. I mean, both papers will be presented very soon. It's a solo paper, instead, in the case of Matthias. Now, I also want to say that there were two runners up whose papers, as well as those of the winners, are already part of the SRB working paper series. So without any further ado, I give the floor to the winners for a brief presentation of the key points of their paper. Who should start? Well, let's go go in alphabetical order, so Sergei first, and Matthias next. If it's OK, otherwise we can. Oh, it's fine. And this is, OK, all right. Oh, why do we do this here? It's here. Thank you, Marko, for the great introduction. And thank you to the selection committee for picking my paper with Adi to be the joint winner. Very honored to receive the prize. So the paper is interested in measuring and understanding liquidity transformation and asset management. We know that liquidity transformation or the creation of liabilities that are more liquid than the underlying assets is a key function of many financial intermediaries, banks being quintessential examples. Since the crisis, there's been an increasing realization and a spirited public policy debate on the extent to which liquidity transformation by asset managers can cause similar financial stability issues. And here, the concerns basically center on the interaction of what we might call performance flow relationship and fire sales. Was the basic concern being about negative performance by a manager leading investors to withdraw capital, forcing the manager to sell some of the assets at fire sale prices, and basically creating this loop. And these kind of concerns are especially prevalent and plausible, given that a lot of the growth since the crisis has been in strategies that invest in less liquid assets. So here's one example from the US. It's the growth of loan mutual funds. These are open and mutual funds that effectively promise investors next day liquidity, but invest in syndicated loans. Which basically an interesting fact about this market is that these loans basically settle T plus 14. So it takes two weeks for them to settle while the fund promises next day liquidity to its investors. An example closer to home might be the UK property funds that had to suspend redemptions following the vote about Brexit. Now notwithstanding anecdotal cases like this, a key empirical challenge in this debate is really the difficulty of measuring liquidity transformation in asset management. For banks, things are relatively straightforward. Maturity mismatch is a fairly good measure of how much liquidity transformation a bank is performing. For asset managers, things are more complicated because in principle, the assets are tradable securities. Also, the price impact can be at least partially passed on to the investors. Nevertheless, asset managers such as open and mutual funds do perform some liquidity transformation. And effectively, the way that it works is that the pool transaction costs across investors. And so if you're an investor in an open and mutual fund, you can in principle withdraw unlimited quantities at the end of day net asset value without internalizing the transaction cost and the price impact that you're basically effectively imposing on all of the remaining investors in the fund. And that's what basically effectively what it does is that it flattens the price quantity schedule that you face as an investor and makes the assets more liquid for you to trade in the mutual fund than potentially trading in their underlying securities, which might be corporate bonds or properties, like an example of UK property funds. Now, so how do we gonna try to measure how much liquidity transformation funds are really performing? So the key idea in the paper is to take a revealed preference approach and to argue that the way that mutual funds manage their own liquidity helps shed light on how much liquidity transformation they're performing. And specifically, how aggressively funds use cash to accommodate investor redemptions and subscriptions, so net fund flows, and the level of cash holdings are gonna be measures of liquidity transformation. So the first one should be relatively intuitive. If a fund was investing in perfectly liquid assets, there would be no need for cash and it would be simply transacting in the underlying securities immediately whenever it receives net fund flows. The second one might be a little bit more counterintuitive from the perspective that if you force a fund or an asset manager to hold more cash, then mechanically it's gonna be performing less liquidity transformation. The extreme would be a fund that holds 100% cash, while it's not performing any liquidity transformation at all. But the key to recognize here is that funds are gonna be, asset managers are gonna be setting their cash holdings optimally and they are gonna be doing this to mitigate some of their expected liquidation costs. And so the funds that invest in less liquid assets and provide more liquidity to the investors, so have more volatile fund flows, are gonna be the ones that are gonna hold more cash in equilibrium, although their cash holdings are not gonna be fully enough to upset the higher expected liquidation costs. So that's the basic idea. We take it to the data using data on U.S. open and mutual funds and basically find results that are consistent with sizable amount of liquidity transformation taking place in mutual funds. So first of all, from the dynamic perspective, even at a one month horizon, which is fairly long in asset management, a big fraction of net fund flows is accommodated through changes in cash as opposed to trading and the underlying securities. And it's more so the case when the fund invests in less liquid securities. Also the level of cash holdings is strongly related to proxies for liquidity transformation, the ones that I mentioned before, asset liquidity and flow volatility. And there is an important interaction effect between the two that basically for funds that have their liquid assets, the volatility of investor fund flows doesn't matter for their cash holdings. So these funds are kind of close to the frictionless ideal. The typical fund however is relatively far away from that and so for the typical fund in the data, flow volatility is gonna have a substantial effect on its cash holdings. Now this is kind of like validation of the basic idea of what can we do with this? One thing that we can think about is the extent to which funds internalize their price impact. So in the paper we have a simple model but also both are the simple model in the paper and most theories would suggest that funds will not hold enough cash to fully internalize the price impact that they're imposing on other market participants, right? And that is basically the counterpart of the more standard leverage fire sale externality. Now what we're gonna do empirically is look for settings where there is likely to be more internalization of the price impact, right? One setting you can think of is that the monopolist would fully internalize own price impact. If I own all of the security, I fully internalize any price impact in that security. And so what we find is that funds that hold a larger fraction of the outstanding amount of the securities that they invest in, either equities or bonds, hold substantially more cash than other funds. The second setting, second idea is that funds may partially internalize the price impact that they impose on other funds within the same mutual fund family. So we measure overlapping holdings between the fund and other funds inside the family and funds that have greater overlap with other funds in the family tend to hold more cash, okay? Consistent with them worrying about the price impact that they might be imposing on other funds, on other related funds, okay? Now one implication of this analysis is to say that since most funds do not fully internalize their price impact, like the special funds that we're looking here, the kind of like the monopolist idea and the overlap idea, since most funds are not close to that, they're gonna hold too little cash relative to what an agent coordinating across all of the funds would have them hold, okay? The second sort of like extension of this kind of idea is to look at the effects of cash holdings on stock fragility, okay? So the idea here is to measure at the stock level the cash holdings of the funds that hold the particular stock. And if the stock is held by funds with abnormally low cash holdings, it's potentially gonna be more fragile because if investors withdraw capital from the funds holding the stock, the funds might have to sell it, right? To liquidate it and gauge and force fire sales as a result of potentially introducing excess volatility in the stock, okay? So in this analysis, what we do is we first of all at the fund level calculate sort of this adjusted cash-to-assets ratio which is tells us how much cash you have relative to like sort of our model relative to other funds within the same objective sort of size, age, et cetera. And then at the stock level, we calculate the average cash-to-assets ratio of the funds holding the stock, okay? So we do this at time t and this is what you see on the x-axis. On the y-axis is the daily stock return of volatility over the subsequent quarter. What you see is that stocks that are held by funds with lower cash holdings are more volatile and this is especially the case where the stock has a relatively high mutual fund share so relatively high fraction of the stock is held by mutual funds, okay? And another extension that we're working on is using this revealed preference approach to think about corporate bond liquidity. There's been a lot of concerns about deterioration in corporate bond liquidity since the financial crisis. Now if you look at kind of like standard measures that academics have of liquidity in the corporate bond market, they are basically similar to pre-crisis levels. So you see these measures of liquidity spiking during crisis but then they come down and they're similar to what there were before crisis if not even better. The caveat to the standard measures of course is that they're based on transactions. You can only measure them for bonds that do trade. It's difficult to say what's the sort of latent liquidity for bonds that do not trade. And the fraction of bonds for which you cannot measure standard measures of liquidity has actually been trending up, okay? So this is where the revealed preference approach can help, we can look across bonds and the relationship between cash holdings and flow volatility. This relationship should be stronger, the more illiquid the bond is, right? And we can construct these measures and see how they change over time and for different bonds, right? And so kind of like a couple of nice things about this approach is that we're not limited to looking at the transactions. And we don't have to take a stand on sort of like what's the relevant measure of liquidity, it's sort of like what funds really care about, okay? So basically to conclude, we take a revealed preference approach to measuring liquidity transformation and mutual funds. We find evidence of a sizable amount of liquidity transformation and kind of two implications that I wanna touch on. One is that liquidity transformation and asset management is dependent on liquidity provision by the traditional banking and shadow banking systems. And that's in a sense that in order to provide liquidity to their own investors, asset managers such as open and mutual funds that we're looking at hold large amounts of cash and equivalence and a lot of these assets are basically short-term liabilities of the banking sector. So things that affect liquidity provision by the banking sector will potentially have effects on liquidity provision by the asset management sector. And other implication that I already mentioned before is that despite their size, cash holdings of mutual funds might not be large enough to completely mitigate their price impact externalities. Thank you. Thank you very much. Yes, so yes, we'll present, okay. So both papers are about macro prudential stuff more than last year's paper, which was more about consumer protection. Okay, so let me also thank the scientific committee for this honor to be here today. I'm very grateful to share the price with Sergey. So this paper is going to be about investments into the securitization market or in what I use as an umbrella term, asset-backed securities. And I'm going to look at investments by banks in Germany. And more specifically, I'm going to look at how considerations for regulatory arbitrage are going to influence these investments. So the regulatory framework for asset-backed securities came under fire during and after the financial crisis. And it has been reviewed by academics and the media, but importantly also by regulators. So for example, the Basel Committee on Banking Supervision looked at its own regulatory framework under Basel II and tried to identify weaknesses or shortcomings in capital requirements for asset-backed securities. And one important weakness that they identified was the mechanic reliance of capital requirements on external credit ratings, as those issued by Moody Stunner and Poznan Fitch. And there are four different lookup tables. Well, it depends on the approach, but basically you have a mapping of the credit rating of the asset-backed security into risk weights, which will then be mapped into a capital requirement. And the fear was that these credit ratings, and hence the capital requirements, would not be sufficiently sensitive to risk. To make this point more clear, let me show you these box plots, where on the horizontal axis, you have different rating buckets that roughly correspond to regulatory buckets. So securities inside one rating bucket have similar and often even identical capital requirements. But still, if you look at the distribution of yield spreads in these different boxes, you see quite a lot of dispersion in the market price for risk. More importantly, or not more importantly, but also to note is that the relation between the credit rating and hence the capital requirements and the yield spreads is not even monotonic. So for example, a bank could buy a high-yielding A-plus rated security rather than a security with a similar yield spread in the A-minus or even triple B-bucket. And without incurring higher capital requirements. Overall, if you take the correlation between capital requirements and yield spreads, the correlation is only 0.5. And if you compute an approximation for the return on the capital required for an ABS investment, you also see a lot of dispersion. If you use the ratio of the yield spread and the capital requirement for one euro invested as an approximation for this return on the equity required by the regulator, you see a lot of dispersion and especially at the higher end of the rating scale. So especially A-plus rated securities can have very high returns on equity. And the question now is whether banks will exploit that and which banks? Maybe those that are constrained by the regulator more than others. And then what does it tell us about the effectiveness of the regular framework that was in place for an asset class that was at the core of the financial crisis? So the way I tried to get at this question was to look at data maintained by the German Central Bank, the Bundesbank, that records exactly how much a given bank in Germany has invested into one specific asset-backed security. So security-level data. And it allowed me to estimate the probability that a given bank would buy a given asset-backed security as a function of the yield spread and the capital requirement. And this figure summarizes the first result of the paper. So on the horizontal axis, you see the cross-section of banks ordered by how constrained they are by regulation, by their capital adequacy ratio. Banks with low capital adequacy ratios over here have relatively tight regulatory constraints. For them, regulatory capital is scarce. And banks at the right-hand side, they have less regulatory constraints. From the regulatory perspective, they appear better capitalized. And then here on the vertical axis of the figure, you see the effect of an increase in the yield spread on the probability that the bank will buy a given asset-backed security, conditional on its regulatory treatment. So take, for example, a constrained bank with a low capital adequacy ratio at the boundary of 8% and take the green dashed line, which is the A-plus bucket. And you see that if you increase the yield spread of an A-plus rated security by one percentage point, the probability that the bank would buy the security goes up by four percentage points. And this is similar for the orange line that's for A-rated, for the brown line, which is for double A-rated, and for the blue line for triple A-rated securities. That corresponds roughly to these rating buckets that I showed you in the previous figure. So it's consistent. Second, I would like to take away from this graph that the effects disappear as you move to the right. So banks that have lax regulatory constraints don't appear to reach for yield. They do not seem to exploit the low yield or risk sensitivity of capital requirements of ratings. Now the question is, what does it mean for the effectiveness of the regulatory framework? How big is the problem? And I would like to discuss this figure in this context. The horizontal axis is the same. Again, you have the cross-section of banks. On the left-hand side, you have the constrained banks with low capital adequacy ratios. And on the right-hand side, those banks that appear better capitalized in the eyes of the regulator. And this brown line here is an estimate of the yield spread of the average asset-backed security bought by a given bank. So for example, a bank with low capital adequacy ratio here on average buys asset-backed securities that have a yield spread of 110 basis points. And what you see is that constrained banks tend to reach for yield. They buy securities with higher yields or higher market price of risk than unconstrained banks. And now next I'm going to plot the risk weight of the average position taken by the banks. And you see that constrained banks, even though they buy securities with higher yield spreads, they buy securities that on average have a lower regulatory risk weight. Whereas unconstrained banks, they incur higher capital requirements. It might not be bad for them because regulatory capital might be abundant in these institutions. They might attach a different shadow value to capital. But the takeaway is that the securities bought by these institutions appear riskier, at least if you interpret the yield spread as some proxy for risk that is compensated by the market. But nevertheless, they have lower risk weights, which is against the idea that riskier investment should have higher capital requirements. In that sense, the regulatory framework seems to do less well than for other asset classes, let's say. Coming back to this approximation for the promised return on the capital that the regulator requires for a given investment. So not return on equity for the entire bank, but what's the return on the capital that you need to use to fund one specific ABS position investment? If you compute this approximation, you see that a relatively constrained bank on the left can have a promised return on equity of 46%, which is very high and which can be realized due to the extremely low capital requirements and the systematic choice of the highest yielding securities in the buckets. This relatively unconstrained bank has a four times lower return on equity on average for the average aspect security purchased. And the conjecture is that if you can increase the return on the equity required by the regulator for an investment by a factor of four or let it be a factor of three or two, that still maps into a significant increase in risk of the position. So then in the next step, I looked at the exposed performance. What are the implications of this behavior? And I see that the constrained banks on the left-hand side here, they tend to allocate more of their portfolio to the lemons in the market in the sense that they buy more securities whose collateral will perform worse exposed. Nine months after investment, I see that the ABS bought by these banks here have a higher delinquency rate. So more of the collateral that backs the ABS is delinquent, then the collateral here, the ABS bought by the relatively unconstrained banks. Okay, so I don't know how am I doing on time? I have the impression that I was really fast. Subjective bias. So to conclude, this is a paper on the buy side of the securitization market. To be sure, there are other papers on the demand for SAP X securities, but either they look at aggregated data or they look at other investors. Sergei has a paper. For example, he looks at funds and investment companies. But for banks, we know very little what determines their investments into this asset class that became very famous during the financial crisis. And we see that banks actually exploit the low risk sensitivity of rating contingent risk weights for asset-backed securities. I call this regulatory arbitrage, but you might call it differently. And what this may be worrying is, in particular, those banks that appear fragile at least from the perspective of the regulator. So those with low capital adequacy ratios. That they are evading the very capital requirements that were designed to limit their risk taking. I believe that the extent of this, what I call regulatory arbitrage is economically large. And the question is where we go from here. So I just found out yesterday that the BASA Committee redrafted revisions to the securitization framework. I think it became public in July, 2016. And they reiterated that they are going to move further to internal credit ratings, rather than using external credit ratings to determine capital requirements. The question is, will they be more risk sensitive? Alternatively, people have brought forward other ideas like calibrating risk weights to market measures of risk, yield spreads, for example. Rochet did that very early in 1992. The question is, of course, how liquid are prices and how easily can the speed shift? Okay. Thank you. Okay, just a few words about both papers. First of all, basic recommendation, read them because they are worth it. I read them several times in, you know, I was part of the committee within the ESRB picking them among a vast set of people who submitted work and they're really good papers. And they're good papers because they look at new data and they provide new pieces of evidence on issues which are highly under-resourced. The paper, by itself, yeah, belongs to an area of research which is very important for the ESRB and for all prudential, macro prudential regulators because he looks at issues which are relevant for the prudential regulation of non-banks, which is an issue that we have already discussed yesterday, they came up yesterday, and which is increasingly important. So the question is whether, for instance, mutual funds or other financial intermediaries which are non-banks are holding sufficient buffers in general, in particular in this case, liquidity buffers or not. And this is an issue, for instance, on which the ESRB has already done a study based on questionnaires to asset managers to ask, first of all, what is their perception of their own liquidity, of their own liquidity buffers in response to possible market movements and so on. And then there is a second questionnaire which I don't remember if it has already been done or not on second round effects and stuff like that. So it is stress testing in a way of these institutions is increasingly important in prudential regulation. The paper makes a claim which is very important from the point of view of a macro prudential regulator. It says that they have evidence that funds do not fully internalize the effect that providing investors with daily liquidity has on the prices of the underlying securities. I'm not completely convinced that actually they show this but to the extent that they can actually substantiate this point, this is quite important because, of course, it opens potentially the role for intervention by a macro prudential regulator in the form, for instance, of either recommending or forcing mutual fund managers to hold more liquidity than they currently do or other interventions. But I would like to make sure that really, really in the paper there is evidence for this. There is another piece of evidence that Sergei actually mentioned just at the end on his last slide of his presentation which I think is very important from the point of view of a macro prudential regulator. And it is that the form in which this liquidity is held by these mutual funds is essentially deposits with banks or with shadow banks or money market funds. And this belongs to another area which is under the spotlight of the ESRB which is essentially interrelationship between a non-bank and a bank intermediaries in the provision of liquidity. So imagine that, for instance, there is a shock to bank solvency or to shadow banks solvency, then this is telling us that these guys, these mutual fund managers, we have problems in drawing upon their liquidity which they set aside in order to face possible shocks to their net inflows or outflows. And so essentially there is a potential interrelationship between liquidity of banks and liquidity of these intermediaries. So it is all very interesting stuff from the point of view of macro prudential regulation. The paper by Mattias is also very important because it provides the first piece of evidence based on micro data as he mentioned regarding banks on regulatory arbitrage of risk weights. And he explains very well the fact that the way risk weights were set, at least, I mean, looking at this data for German banks, pushes or induces or lets, at least, banks which are more constrained by regulation to actually engage in purchases of riskier securities because these provide, promise them, a return on required equity which is about four times as large as for unconstrained banks. Yet exposed, they had the lowest exposed performance within each risk weight bucket. So essentially, risk weights appear to be utterly counterproductive when they are set because they both increase or at least allow an increase in risk taking by the wrong banks, so the constrained banks, the riskier banks, but also exposed these banks do worse with them. And then towards the end, Mattias was going already in that direction, these banks, the policy question, what to do? I mean, these risk weights are clearly not well set, so should we scrap them, should we base them on more like finer, you know, on basis of finer measures than credit rating agencies ratings, but then which ones should be the actual yields or should be the CDS premium when they're available? This also had a problem, he mentioned, should we go for limits to large exposures? Of course, these are very hot issues, not only for micro prudential regulation, existing one, but also for future possible one. Imagine that we, as we were mentioning before with in the previous session, we have positive risk weights or known as Zelly zero risk weights for sovereign bonds, you know, this all becomes highly relevant also in that case. So of course, this begs a lot of questions, but at least we know that there are things that don't work over and above the prosyclicality, which is generally the issue, which is raised above risk weights based on credit rating agencies' ratings. Okay, let me finish here. I wish you best of luck with the publication of this work. Thank you.