 I'd like to talk about bond markets in the spring of 2020 and how the Federal Reserve acted to stabilize those markets. So I'm going to do this in three parts. First, I'm going to just recap the by now pretty well known dislocations in Treasury and in investment grade corporate bond markets back in March with yields increasing sharply much more than credit defaults drops in each case. Then more interestingly, I'm going to try to sort out who we're selling and why I'm going to document very large outflows from bond mutual funds, especially investment grade corporate bond funds. I'm going to argue that those outflows contributed to both Treasury and corporate selling since even corporate bond funds and other risky funds hold treasuries for liquidity purposes. And those are the first thing that the funds will sell when faced with outflows. Once they then run out of Treasury, they're going to start selling some of their coarsets. I'm going to argue that these outflows from bond mutual funds can be rationalized by a disappearing safety effect, by which I mean investors willingness to pay a bit extra for assets with very low default risk. And then I'm going to view also the role of foreign selling both official selling as well as such fund selling. Okay, so faced with sort of armed with those facts about who we're selling, we can start talking about what the Fed did in order to stabilize markets and in that sense stimulate the economy by avoiding market meltdown. And I'm going to distinguish between treasuries and corporate bonds. So on the Treasury QE side during COVID, the purchase announcement did not stop the yield increase, but very large actual purchases did. So from that, I'm going to conclude that when markets are dislocated due to people selling treasuries for liquidity purchases, then central banking is not about talking but acting and providing that liquidity. So if there's a shortage of people buying in the market, it doesn't help that much to say, look, I'm willing to buy down the road, you have to buy now and have to buy in large quantities. In that sense, I'm going to argue that QE during COVID work very different than QE during financial crisis where we saw very large announcement effects for Treasury QE. Then turning to the corporate bond purchases, which I knew from the Fed's perspective at least during COVID, I'm going to show that things work very different in that case. The announcements lowered yields very quickly, altering today evidence, whereas the actual purchases were delayed and modest. From that, I'm going to conclude that you can stop a run on bond funds buying very little with a sufficiently strong announced willingness to buy. This will sound a lot like the Draghi, whatever it takes speech back in 2012. And I'm going to also review how the Fed's corporate bond purchases work different from the ECB's Corporate Secure Purchase Program in the sense that the ECB program was not designed to stop a run, but was conducted in more normal times and required substantially larger quantities in order to move yields and have real effect. So just touching back to what Ralph was saying about estimating demands, it is a little bit depressing from all of this that there's different ways that QE works during different time. It works differently across different purchases and that makes it, I think, important to supplement demand systems with these even more event study style grabs that I'm going to show you. All right, so let's dig in. So let's start with in blue here the 10-year Treasury year and invert the S&P 500 index. Notice how in January and February and early March you see a pretty good co-movement between the stock market and the 10-year yield. What is abnormal is it starts happening around March 9. The stock market keeps falling, but the Treasury yield goes up by 64 basis points from March 9 to March 18. And then things magically stabilize and you get back to a pretty tight link between the stock market and yields. So that's different from what we saw in 2008. There was no sudden spike in Treasury yields back then and we didn't see that breakdown of the stock bond correlation by then. Across maturities here the blue line is the same as before the 10-year. The longer yields spike more, look at the 30 in the 20-year, but there is still even some spiking down into the 5-year, maybe a bit in the 2-year. In terms of whether this is real or nominal, it's not the case that these Treasury spikes were driven by increased inflation expectations. So one could have been concerned that perhaps due to supply problems during COVID there would be some inflation moves and that's not the case. The green line here is the 10-year inflation swap and the blue line again is the same as before. So if you think about what's really moving here, it's the real rate. You can see that I put in the 10-year tips, the 10-year real rate debt moves, then even more than the nominal. So it's not about inflation. It's also not about credit risk on Treasury. Sorry, I couldn't get Bloomberg to concede this particular series. That's a slight movement in the 5-year Treasury credit default swap for the U.S., but it's pretty small compared to the movement in the yield. All right, so that's why I use the word dislocations because there's something weird going on in pricing relative to a default risk. That's also what we see in the investment-grade corporate bond market. So in blue here is the yield spread between investment-grade U.S. corporate bonds and the 5-year Treasury and then let the credit default swap. And you can see that there's a big dislocation in that the yields run up much more than the CDS. That difference is what's called the CDS bond basis. And if you compare the investment grade to the high yield, of course the scale here is different, but it looks like the high yield CDS tracked the high yield corporate bond spread much more closely. Therefore, if you compare the CDS bond basis, I'm going to flip the sign so that you can think about I'm going to flood the yield spread minus CDS. You see actually that during COVID the yield spread minus CDS has about the same amount of spiking as the high yield, which is very strange because it means that you see the same amount of market dislocation in the market with low default risk relative to a market with high default risk. Let me just flip back to remind you the CDS in the investment-grade market is peaks at 150 basis points. For the high yield, you see the CDS about 8 percentage points. So one would expect some sort of proportionality between dislocations and credit risk. And that's not what we're seeing. And this is abnormal relative to the fall out of the way. There's a nice paper by Biden calling the friend where they documented during the financial crisis. You see the CDS bond basis being about twice as high for the high yield than for the investment-grade market. So there's something new going on in treasuries and in investment-grade corporate during COVID relative to how things turned out in the fall out of the way. So what happened? Who was selling and why were they selling? So I'm going to focus on who was selling. There's other very nice papers who focus on who were not buying, which were the dealers. So one could have expected that once large-scale selling set in that the dealers would go in and make markets, it did that to some extent but not enough. So Del Duffy has proposed and we need to basically unclog the dealer's balance sheets by doing central clearing of treasuries to basically make the treasuries move faster through the system. Hey, Navel and Song have provided an asset pricing model to link the dealer balance sheet constraints to asset prices. But again, I'm not focusing on who were not buying but who were selling to get to sort of the root of what was going on. So let's start with some data from the U.S. financial accounts. And I'm going to show you both holdings. Holdings changes and how much was actually traded. That's because the holdings change in the flow funds will capture not only trading but valuation changes. So it'll be more informative for understanding buying and selling to look at the traded amounts in between the last column. All right, so at the end of 2019, there was 19 trillion of treasuries outstanding. And you can see that about 6.6 trillion O's were held by foreigners. 1.3 by mutual funds, 2 by the household sector and so forth. All right, so look at the last columns. The main sellers in 2020 Q1 were the best at world selling 287 billion, mutual funds selling 236 and the household sector which includes domestic hedge funds selling 170. For the rest of the world in the mutual funds, there's data that show that those amounts sold are concentrated in their notes and bonds as opposed to the bills. For the first time in this last financial account issues, the financial accounts break out U.S. domestic hedge funds from the U.S. household sector. And so for the first time we know like the domestic hedge funds own about 200 billion from the minus 31 then holding change you that suggests that there was some selling from domestic hedge funds. I'm going to argue in a second that hedge funds were more important within the best of the world category. Ralph mentioned how came on islands are important in a lot of these portfolios. All right, so those are the three main sellers. Who are the buyers? Well, the fact buying more than a trillion in treasuries also focused on the longer end and then money market funds which experience large outflows as so which experience large inflows as people were pulling money out of the prime funds and putting them into money. Not surprisingly, those purchases concentrated in bills. Just to get a sense of how abnormal this was, it's helpful to do some time-chase graphs and to the left, the rest of the world, both official and private, you can see the large amount of selling almost 300 billion in Q1 that we talked about. Notice also that that's not something that happened in 2008. Same with the mutual funds, you see vast amounts of treasuries selling in 2020 Q1 and that was not something that happened in 2008. The households net selling net purchases are more volatile and the numbers there look a little bit less abnormal. I should say, though, I'm not focusing on this sector here because in the flow of funds like in the US financial accounts, the households take place the residual and so this may not reflect sort of actual households but more what's not accounted for elsewhere. So that's a that's the worst category to study. All right, so let's start thinking about why were the mutual funds selling treasuries. From the financial accounts, you can document the outflows for mutual funds. So this is how much money people are putting in or pulling out from mutual funds. For now, it's all mutual funds, equity, fixed income and so forth. Notice to the way right that there were mutual fund outflows more than 300 billion in 2020 Q1, much larger than anything we saw during the financial crisis. There's not anything too exciting going on on the ETF side. Moving on to the to a different dataset from the Investment Company Institute, we can split the mutual fund outflows into bond funds and stock funds and go to a monthly frequency. The top graph documents that the bond outflows in March were 265 billion. That's an eight sigma event in in bond outflow terms with not that much. I mean, there's some outflows from from equity, but nothing of the same order of management. And again, notice 2008 there was some bond outflows, but you know, not anything here. Hope you saw this time. Going to weekly data remember how in my very first slide, the Treasury yields spiked at on March 18, the bond outflows peak exactly around the same time with the largest outflows in the week ending on March 18. And the following week, both time both in both cases more than 100 billion in outflows. If you split it up into what kind of funds people were were were flowing out of the timeline here, the investment grade funds. And the orange one is the government, which includes the Treasuries and maybe us. You can see that people are pulling money out of all the risky stuff, but disproportionately out of investment grade funds. So here's my interpretation of what likely was going on. There was funds facing outflows, and they were selling Treasuries and whatever they had that was liquid in order to meet those outflows that contributed to the spike in Treasury yields. However, they didn't have enough Treasuries because this was an eight sigma event. And so then they had to start selling illiquid assets into an already liquid market. And therefore we see these large spikes and large dislocations in the investment grade market. There's a nice paper by Michelle and saying that argue that they actually didn't have enough Treasuries that the corporate bond outflows were about 12% of us in the management, whereas they only had about 4% on average in liquid assets. Yeah, so another nice paper I just want to mention by Trilado Goldstein and Hart-Sachs were studying these mutual fund flows and they argued that they documented that the share of the corporate bond market owned by funds has gone up from 20 to 40% in a decade. They emphasized that this is a really fragile structure in the sense that the funds promise you can get your money out daily or even more frequent and daily while they hold illiquid assets and some Treasuries, but as we saw not enough. Interestingly, they document reasons for redemptions. So they show that there's a world of fundamentals, there's more selling of funds with more COVID affected sector holdings. There's also some fund specific issues of run dynamics and vulnerability where there's more selling of funds with more illiquid assets and there's also more selling of funds with assets that are similar to other funds. Okay, so it makes the fundamentals and mutual fund specification. All right, so what's my interpretation of the economics of what's going on with this? So I'm going to argue that a plausible explanation is that a bunch of people have been reaching for yield into investment corporate bonds and they were not really focused on bond modeling. They thought these were pretty safe assets and for sufficiently low default, these people were willing to reach for yield into those assets paying a little bit more than probably they should have given the default list. Let me just talk you through, you know, where that argument comes from. So as Ralph mentions, I mentioned I have done some work on as a quantities. And so this is the main picture from my work with Chris and Morty arguing that there's something special about really safe assets. On the left axis, on the y axis, we have the AAA treasury yield spread and on the x axis, the quantity, so the debt to GDP, the quantity of the treasury supply. This graph is supposed to illustrate a downward sloping demand curve for the extreme safety of treasuries. And we show also that that safety effect actually is to some extent inherited by the AAAs being the safest corporate bond category. And if you think about what happened during COVID, I think it was that the default probability, you know, in investors minds justifiably went up. Okay, so let's think about essentially this safety effect disappearing that now the corporate bonds, these investment funds, they're no longer safe enough that people willing to pay a premium for them. And then you get a really sharp price decline. In fact, in percentage terms, the losses were as large on the investment grade funds as on the high yield funds. And the way to think about this in a picture is think about the relation between the price of corporate bond and the default probability. In a normal asset pricing model, you would get some downward sloping line like the red line. The safety effect is about investors willing to pay a bit more as illustrated by this blue line for stuff that's sufficiently low default risk. The green line illustrates that that effect is even bigger if there's a small supply of safe assets out there. Okay, so if you look at the blue or the green line, once we then move these corporate investment grade securities out towards the point where they're no longer that special, notice how they're slow, but the blue and the green line is really steep. And you can see these sharp price falls, which is exactly what we saw in some of those first class. This is not something that one could not follow in real time. I wrote a policy note on March 22 arguing the fetched by corporate bonds. And these were effects that one could could follow the fund outflows and sort of the safety argument is something that one could could easily follow in real time. All right, so turning to who else was selling. Let's talk about the rest of the world selling. Remember, there was about 287 billion in Q1. So why were the foreigners selling? So from the New York Fed's custody holdings, the foreign official sectors of their central banks were selling 109 billion of US treasuries in March. Here's a chart from Bloomberg, where you can see that that is actually abnormal. That's also not something that happened back in 2008. As Ralph mentioned, we don't have data broken down by country on the official holdings. The Bloomberg article states that the people sell the country selling were those reliant on oil exports, presumably because they had a shortage of funding for, you know, government watches and so forth, a lack of lower inflows from oil sales, as well as smaller Asian economies, suggesting trade disruptions and countries likely intervening to defend exchange rates, for example. The foreign private sector was also selling here some data from the treasury, particularly the foreign private sales of treasuries were 250 billion in March. It's informative actually to look at the country breakdown, which is only available for official plus private, but nonetheless, because it shows here to the right that the main country selling was the Cayman Islands, which is a hedge fund hub. That fact I think has prompted literature on the role of hedge funds in the market, the treasure market dislocation and there's sort of competing views about how important the hedge fund selling was. So the BIS came out early, the Sushko paper is the earliest one here. They came out and said, look, yes, the hedge funds were selling, they were selling to unwind treasury basis trades, which I'll give you in a second. Then there's a subsequent paper by the very nice paper by Barton Pankin the ORFR that says, yes, they were unwind these basis trades, but it doesn't seem that that costs price distortions to a large extent. So let me just recap that argument. To the left, you can see for Barton Pankin the short futures positions for hedge funds. And it's a little hard to see, but in the right of the picture, you see that decline in 2020. I tracked down the data to the right, I'm just showing you what happened in 2020. These short futures positions by hedge funds are down 131 billion in March. Now, this treasury futures basis, the treasury basis trade goes like this, it's relatively straightforward actually. So you enter a short future position to deliver treasury and get cash at a future date. Now, in order to have a treasury to deliver at that time, you buy a treasury, the cheapest one you can find within the set that one can deliver. And then you fund that position using repo. This is profitable if you get the treasury at a good price upfront relative to the futures. And if you manage to roll over the repo financing at a good rate and meet margin requirements along the way. So in mid-March, the margin requirements increased as volatility went up and that led to trade on mining. Now, was that a key factor in treasury mispricing? So that's where people disagree. So to the left is the BIS argument and to the right is the OFR argument. So the BIS guys say, look, the treasuries look cheap relative to futures and repo. So what they do is they calculate out what is the interest rate implied by positions one and two. This is just by the present value formula. How much did you pay for the treasury upfront? What can you sell the futures for? What can you get in the futures? Okay, so that's an interest rate. It's called the implied repo rate in this market. They said that interest rate was actually pretty attractive, suggesting that the treasuries were underpriced in the regular market compared to the cost of borrowing, which is the actual repo rate. So the treasuries were cheap. Now the OFR guys say, look, that's maybe not the only relevant comparison. They say that the cheapest to deliver treasury, which is the one that the hedge funds would be buying in this futures trade or to support the futures trade, was actually expensive relative to other treasuries. So the graph compares the price of this cheapest to deliver treasuries to other treasuries that have similar maturity but were not deliverable into the futures contract. And say, look, this doesn't suggest that there was massive dumping of this cheapest to deliver treasuries due to unwinded basis trades because the cheapest to deliver treasure was actually expensive as opposed to cheap. Okay, so there's the argument. Perhaps then that means that the hedge funds were not dumping treasuries in severe distress, but instead unwinding in a more reasonable fashion. And then the futures market is still providing some liquidity to the cheapest to deliver treasury. Be that as it may, remember how the foreign private sector sold about 250 billion, that the possible fraction of that that could be due to unwinding at the basis trades is at most half, because we saw, as I showed you, the decline in these futures positions was about 130 billion. Okay, so there's something more going on. So one possibility is the dollar shortage. The, just to give you an example, so where that could come from, suppose you're a Swedish pension fund, you want to invest in the US, but you don't want to try and do this. What you would do is that you would contact the bank, the bank would borrow dollars in the short term dollar markets, for example, by issuing commercial paper to a prime fund in the US. The fund in the bank would enter an FX swap. And normally, the swap would get rolled over every three months, say, but if the bank funding is not rolled over because people are pulling money out of prime funds in the US, then neither is the bank going to roll over its FX swap. So what will happen now is that the fund, which has borrowed dollar upfront and owes dollar later, needs to now deliver those dollars, which will involve either selling dollar assets or buying dollar for coma in the spot market, which then might lead to intervention by the central bank, which again leads to selling dollar assets. From the Fed's financial stability report, they were mispricing in the FX swap market as measured by the FX swap basis indicating dollar shortages. That actually is not new relative to 2008. So there was a dollar shortage in 2008. Two and foreigners didn't sell treasuries back then, suggesting perhaps this is not a big deal. But I'm going to show you suggested evidence in a second that the Fed's swap lines actually did help lower treasury yields, implying that this might have been a factor that the dollar shortage may have been linked to the treasury dislocations as well. All right. So then let me bring you with the Fed did and how that worked. It's not obvious. Exposed sitting now, everyone is saying, look, the Fed saved the world. It's not obvious that the Fed did save the world because if you graph out the COVID cases in red here and the stock market in blue, you can see that the stock market turns as the growth rate of the COVID cases starts falling. So I think we have to be a little bit modest on the central banking side and just sort of admit there's other stuff going on. There's also the CARES Act, the US fiscal stimulus passed on 324. So we need to dig in a little bit more to figure out what was going on with the world in central banks. So the Fed essentially did four things. They lowered the target. They reactivated the swap facilities with foreign capital banks. They had a ton of facilities to stabilize money markets and a bunch of programs to stabilize bottom markets including lots of different QE type problems. I'm going to talk about the three key dates because I had checked all the other dates and there's not as much going on in bottom markets. So the first key date is March 15. The Fed lowers the target rate, the primary credit rate, and the dollar swap mine rate. Then they announce 500 billion of treasury purchases and 200 billion NBS projects. All right. The second key date is March 23. This is the Fed's whatever it takes moment. They announced unlimited treasury and NBS purchases and they introduced for the first time ever corporate bond QE. 300 billion and lending back by 30 billion from the treasury and credit protection. The third key date is April 9. They go from 300 billion to 850 billion in corporate QE with 85 billion credit protection. And they add high yield. So the safest within the high yield category, the following agents that used to be investment rate but has subsequently been downloaded. All right. So what happened? If we look at the treasury market, we have again, this is the same graph. It's the same as the very first slide from today. The blue line is the 10 year treasury. The three lines mark the three key dates I just took you through. Okay. So notice that the blue line does go down around the first mine, but then it starts going back up. So the announcement that they were willing to buy 500 billion of treasuries, which keeping in mind the amount of the magnitude that I talked about how much mutual funds invested the world was selling was like a huge amount. It didn't stop the market from, the market dislocations from getting even worse. And then somehow magically the market started getting better. The peak on the blue line is on March 18. The market's like getting better in the 19th, especially the 20th, where it doesn't, you know, there's no new Fed announcements. And if you look at, you know, intraday data for the second line on March 23, there's a little bit of that. The decline in the 23rd, that's due to the Fed. He's aming today graphs to show that this is returned. The prices do go up a little bit, but the main action, you know, is what went on on the 19th and the 20th. All right. So one interpretation is that something else happened, maybe the COVID cases, but better, maybe fiscal news came out. What's this not due to the Fed at all? Well, I'm going to argue that that actually was due to the Fed, but that we need to look at quantities. Okay. So this is, I think I've shown you probably too many slides. This, I think is the most interesting one. So the blue line is usually a 10-year treasure. The red line is daily treasure purchases by the Fed. Notice how the Fed kicks into really high gear on the 19th. And then they keep buying 75 billion a day or 70 billion a day, roughly every day from the 19th to the end of the month. That's, I think, what just basically hammers down the yields. And so the other fact I want to mention here in terms of hammering down the yields was that if you go in and check this, the first swap line settlements during the COVID crisis was precisely March 19. Okay. So this is when the banks got access to dollars and thus could disperse those dollars to everyone who otherwise might have sold treasury to these dollars. In terms of the amount, the first, the settlements on March 19 were 162 billion of which more than 100 was the EZB. There's also a large amount for the bank of Japan. But remember the Fed was buying 70 billion per day. So this is sort of a couple of days of Fed purchases. The missing factor that we would like to know is for each dollar that say the EZB gets in through the, through the swap lines, how much potential treasury sales by European financial institutions does that prevent? That's a number of probably less than one. And so then since also the total swap lines on this day was 162 much less than the Fed purchases, I think this plays a role but a secondary role to the Fed purchases. So what did we learn then? So what I conclude from all of this is that treasury selling was driven by liquidity needs. This is not about a loss of confidence in treasuries and high CDS rates, sort of, you know, European Southern debt crisis style. This is about liquidity. Because it's about liquidity, central banking needs to be about action and not talking. Therefore, the March 15 announcement was not enough to stabilize the market. It took actual large purchases to bring the news down. So really what we have learned here is that what we could call market functioning QE, flow effects are crucial. You cannot evaluate the policies by just looking at an announcement effect. You have to look at the actual purchases. That's very different from how things work in the 2000s. If you have followed this as long as I have, you might remember that when the Fed announced treasury QE on March 18 in 2009, the 10 year drop by 50 basis points within a few minutes, that's very different from what we're seeing. And I think that's because back then treasury QE was not about providing market liquidity. The treasury market was actually doing quite well. And therefore, things work differently. Back then it was more about the signalling channel of what does investors infer about future short rates. There was also scarcity effects with treasury supply reductions leading treasuries to become more special. I've written about that previously. But the main point is since back then the treasury market was not dislocated, things just worked very differently. Okay, a few words on the corporate markets to wrap up. The Fed's corporate purchases, let me be precise, the Fed's corporate purchase announcements had large effects on corporate market markets and I think stopped those outflows from the mutual funds. These two graphs are for investment grade at the top, but high yield corporate bonds at the bottom. It's returns since the announcement, the investment grade on the first corporate bond QE announcement date on March 23, investment grade corporate bonds go up 8%, that's an 8% return within an hour with smaller, but nonetheless, big effect for the high yield market. So I think that's consistent with the causal effect of the policy, both because the timing lines up and because the investment grade effect is bigger, I remember that the first announcement was only about investment grade. Then the second announcement, you see that the high yield bonds do better, that's because high yield were added, high yield ETFs as well as the following units were added, again consistent with the causal effect. There's another nice paper by Gil Cresen, co-authors arguing these effects are causal, but you can take a look at it if you're interested. So does that imply then that this was all about which exact corporate bonds were purchased in terms of yield effects? The fact that you see higher impact on investment grade when investment grade produces an ounce and similar for high yield does suggest specific effects, but there's actually a very interesting general reduction in risk premium on March 23. So this is a VIX exchange-traded fund. That's just a massive drop in the VIX right around the announcement, so that suggests some general effects. Let me conclude here with the equivalent of the Treasury quantity picture for corporate bonds. The green line, which is the actual corporate bond purchases, is just tiny. So this works very differently for the corporate and the treasuries. They stabilize the market with very low purchases. Essentially, that was because they stopped the run on the funds, boring some graphs from flattables, and it's actually you see basically that the fund flows river after the emergence. All right, let me just say one more thing about real effects. I want to just finish with the word linking this to the ECB's corporate sector purchase program. So after the corporate market stabilized, there's a huge bond issuance boom in the US. Does that mean that the Fed is going to generate large real effects? So that's not clear. There's some debate about the impact of impacting corporate bond yields from the ECB's program. I think what I learned from reading up on those papers is that they've had real effects, but that they're not exactly what you might expect. In particular, if you facilitate corporate bond borrowing from large firms, that could actually be good for small firms, because then the large firm use less of the bank's lending capacity. So what we really want to know is if the Fed hadn't saved the corporate bond market, would the small firms then have done even worse than they currently have? We see actually that during COVID, it's the large firms that increased their bank lending, not the small ones, but that doesn't tell us what the counterfactual would have been. So this my last slide here is would the SMEs have done worse, even worse without the Fed programs in terms of getting bank loans? And I'm suggesting here sort of a couple of things that we want to see tested. The question we want to know is if the corporate bond market had not been saved, then the large firms would likely have used their credit lines even more? Would that have heard the small firm borrowers? One way to test this would be whether the small companies that were customers of banks with larger credit line exposure from large firms did those small firms borrow less during COVID? You can see how one could set up the test. Okay, so apologies for running a little bit late. Let me just conclude by saying that it succeeded. They stabilized the market, but of course, now moral hazard and financial stability is an even bigger concern than before. Thank you very much. Thank you, Annette. A very nice, very instructive overview of the facts and also an interpretation of why those things might have happened. We've got two questions, one from Tillman Bletzinger. You explain the dislocations by demand side dynamics. Have you also looked into our actual or expected supply of new bonds affect your analysis? So there's a large amount of issuance at the same time. I have not come across anyone mentioning issuance on the key dates. I don't think there was particular changes right around March 19 and March 20, but that's a great suggestion to try to control for Treasury supply coming online at the same time. Second question, Wolfgang Lemke. In your longer write-up of this presentation, you compare the strong announcement effect of recent Fed measures to the working of the ECB's OMT. What do you think is common across the transmission of those measures? One may argue that OMT did address a very specific European problem. Yeah, I think that they're similar in the sense that, so I mean in the multiple equilibrium sense that if the ECB in the OMT case or the Fed in the proper context comes in and makes kind of a buy whatever it takes announcement, then people stop worrying and they stop selling. Sort of in stark contrast to the Treasury liquidity where, you know, if you're a hedgehog facing a margin call you just have to sell and it doesn't matter if the Fed say they'll do whatever it takes they need to actually do it right now. So I think that's the parallel between OMT and the COVID-1 market is that the central bank can say something and that'll make people stop selling and that'll be enough to stabilize the market. Thank you. I have another question just came in. Klaus Masuch. The Fed rightly stepped in, stabilized the markets and also rescued large investors, hedge funds. How can the Fed avoid higher future risk taking? As investors see this as another clear evidence for a strong future of Fed put for which they do not pay. But I think you mentioned this as a moral hazard. That's, I completely agree. You know, I've written papers about the Fed put, I think it's a real concern and now it's even worse. You know, some people say that there's no moral hazard from COVID because the financial sector didn't cause COVID. Obviously they didn't, but they did start out pre-COVID having huge amounts of leverage, both especially the non-financial corporate is actually where the issue has been this time. And so in that sense there is more on hazard. They, you know, allowed all kinds of reaching for yield. They issued a whole bunch of bonds that were barely investment-grade and then the COVID shock hits. That's clear moral hazard. So I think actually that it's time to start charging for the Fed put. And one possibility would be to think about the deductibility of interest. You know, if you've ever taught MBAs and invariably when you teach capital structure you start talking about how there's a tax preference for debt. And then some MBAs will always say, well, why is that? And you try to go through some sort of public policy argument and it's actually pretty hard to give a good argument for why there should be a tax preference for debt over equity. So I would, you know, one possibility would be that we start thinking about not giving that preference to debt. That doesn't necessarily mean we have to tax corporations more. We could just sort of equalize the incentives for debt versus equity financing to, you know, not at least we're not really charging for the Fed put, but at least we are not giving as much tax preference for doing this stuff that leads to crisis done alone. Many thanks, Annette. We are slightly over time, my fault, but we are going to take a break now and reconvene at 17 Frankfurt time. Thank you. See you soon. Bye.