 Welcome to everybody. I'm really pleased to have all of you connected to this event that we are very proud of, you know, of having the opportunity to provide the IK Vandenberg Prize. Let me just give you, let's say, a summary of on one side why we have this prize and who was, in a sense, IK Vandenberg. Well, she was a Dutch politician and a member of the European Parliament. She was also part of the European Parliament's Committee on Economic and Monetary Affairs. And from 2011 till 2014, she was a member of the European Systemic Board's Advisory Scientific Committee. And really, in recognition of their contribution, the European Systemic List Board has established this annual research prize. And it is a very special research prize because Mrs. Vandenberg was dedicated to the notion that finance should serve society. And this prize is administered in that spirit. This is from this from 2014 that the ESRB is awarding this prize. So this is pretty much this. This is the seventh edition. And it is devoted to young researcher. So applicant must be under 35 years old, both the authors or, you know, if you have more authors, all of them has to be below this age. The topic should be related clearly to the ESRB mission of preventing and mitigating systemic risk and clearly also to try to improve financial stability. This year we received 24 very nice submission, you know, really research, young researcher are making huge progress in their ability to produce research through the different years that we saw the different application. And as you may already have observed, the winner are Kaster Müller from the National University of Singapore and Emil Berner from MIT. And the paper that is the winner one, the title is Credit Allocation and Macroeconomic Flotation. The research question that this paper is addressing is on what is the role of finance and credit in particular on macroeconomic flotation, but also on financial crisis. And the key aspect of this paper is that it is distinguishing between good credit and bad credit by looking to different credit provided to the different sectors. Clearly this research is in the spirit of the IK main purpose, so making finance society. But I do not want to anticipate too much. I'm now spending a little bit of time on the housekeeping information. So, as you know, you know, I'm suggesting to all of you to mute your microphone and limit the background interference. Do not use your welcome when unless you are you are talking, please use the chat function and to all panelists to ask questions. And we will have a section where question will be addressed and during the Q&A. But, you know, going back to the main, let's say, program of today, the plan is that we will have the presentation of the paper by, you know, the authors. And this will be for about 30, 35 minutes. Then, and I would like to thank. We will have remarks by Philip Hartman from the European Systemic from the European Central Bank, and seven of us come that is a member of the advisory scientific committees. They will provide on one side comments to the paper, but also they will also give to us a broad view about the topic of the paper. And then we want to dedicate 20, 25 minutes to a general discussion with question and answer. So I will give also the opportunity to the author to reply to the point rise by Philip and seven and clearly also Philip and seven and will continue to intervene. But on this part, I will really appreciate if a significant part of you will ask question and try, you know, to develop knowledge about this very interesting topic. We are planning to end at about a quarter to five. So I don't want to take other time to, you know, to this section, and inviting the author of the paper to start their presentation. I want to remind them high. I want to remind them that they have that you have 35 minutes. And, you know, I will then, by the way, congratulations for the price. We really appreciate your work and we are learning a lot from your paper. So congratulations first and the floor is yours. Thank you very much. Thank you, Laurie Anna for that very, very kind introduction. It's an absolute pleasure and an honor to be here. Before I start our presentation today, I just want to say it's really a great honor to receive this recognition and this project has been in the works for many years. And so as a result of that, over the years, you're really accumulate a lot of debts we really owe many people both that have worked with us, and that have taken the time to talk to us about these issues. But most importantly, of course, we want to thank the ESRB advisory scientific committee for selecting our paper. We're just really happy that we get to be part of this conversation. And so let me start my presentation right here. And then I would appreciate some feedback on whether that works for you. Again, thanks so much for having us. So the motivation for this paper here is that rapid expansions in private credit are often but not always followed by deep economic downturns and financial crisis. This is a well established finding, going back not only to the 2007 eight crisis but also seems to be a more general pattern. But despite this well established finding many questions about credit cycles about how credit markets interact with the wider economy, we think remain really quite poorly understood. And so, for example, why is it that some credit expansions tend to end badly. So with these economic downturns, but others are actually linked to positive growth experiences or even growth spurts. And if that's the case, then how can we tell apart these good booms from bad booms to use some language from our recent paper by Gordon and Ordonis. And perhaps on a more fundamental level, does it matter who gets all that credit during those credit booms, which we think is kind of an intuitive question to ask. And so what we try to do in this paper is to show that the sectoral allocation of credit matters for understanding the link between credit booms, macroeconomic fluctuations and the incidence of financial crisis. Now, why should we focus on the sexual allocation of credit. Our motivation here comes from models of credit cycles that feature heterogeneity in terms of sectors. So most of the models in this literature distinguish between three major sectors in the economy. So think of one about firms in the tradable sector. So for example, large manufacturers, then you might have firms in the non tradable sector. Think here of construction firms or small restaurants, and then you have households. And what these models really emphasize is that these sectors may differ in how sensitive they are with respect to credit supply or financing conditions more broadly. And these sectors might differ in how sensitive they are to changes in household demand. And what these theories would then predict is that credit to the non tradable sector and households in particular can lead to economic downturns by fueling unsustainable demand booms by contributing to a build up of systemic risk or build up a financial fragility and by lowering productivity growth through a misallocation of resources across sectors. Now, if you think about perhaps the most prominent theories of credit cycles that that are out there, they on the other hand do not really tend to emphasize this kind of heterogeneity on the borrower side. For example, you think about theories that stress the net worth of financial intermediaries, or you think about behavioral factors, they are this heterogeneity is not important. So, whether the allocation of credit matters empirically is we think very much an open question. And so together that we need to measure who in the economy actually receives credit and it turns out that that's a very difficult question to answer. In fact, if you look at existing efforts in terms of credit data, it's not really possible to measure that. So just as one example here, there are very widely used data sets by the Bank for International Settlements or the IMF and widely cited work by George Shuler and Taylor. But at best in those databases, you can see how much credit goes to firms and how much credit goes to households. So to make progress here in our theoretical understanding really of how credit interacts with the microeconomy, the backbone of this project is a new cross-country panel database that we think is quite the departure from these data sets. So what's new here is that we can measure the allocation or distribution of credit across an average of 16,000 in the economy. We cover almost 120 countries and we can go back for many countries to 1940 or 1950. And it's exactly these data that allow us to study empirically the link between different types of credit expansions, business cycles and banking crises. And so just briefly preview the results here. We find what we think are quite striking differences in macroeconomic outcomes following different types of credit expansions. So the kind of headline finding we have is that growth in credit to firms in the non-tradable sector and also to households is systematically associated with a slowdown in economic growth in the medium run. But if you look at credit expansions in the tradable sector, that predicts stable or even higher economic growth. And consistent with models of credit cycles with this sectoral heterogeneity, it's lending to non-tradables and households that predicts a boom-bust pattern in demand, a higher likelihood of a systemic banking crisis happening with defaults being concentrated again in the non-tradable sector. And it's also these types of credit expansion that are associated with lower productivity growth, which could suggest a misallocation of resources across the sector. And so the takeaway that I really want to emphasize here is that whether credit booms tend to end badly or not seems to depend to a significant extent on what that credit is actually used for. And the central finding here is really that distinguishing between different types of firm credit is really important for understanding credit. Now, of course, you know, we're far from the first to think about the link between credit markets or finance and the real economy. And this slide here cannot do justice to the important work that many people have done in this area, including many people of you present here today. And so what we most directly speak to are the many, many studies on the link between credit and financial crises on one hand and credit and economic growth on the other. And so our contribution here is to draw on the insights of this large, mainly theoretical literature in international macro to build what we think of as a bit of a bridge between these two empirical literatures. And so more specifically, we show that differentiating between different types of firm credit along the lines that are emphasized by this international macro literature is useful for understanding when credit booms predict bad outcomes and when it predicts not so bad outcomes or even positive experiences. And we think that this is quite a departure from thinking about the world through the lens of how leverage the private sector is or how leverage banks are. And what we want to emphasize is that who in the economy gets credit so this allocation or distribution seems to be just as important or you know maybe even more important than just thinking about how much credit is there in the economy. So as I mentioned, the paper introduces a new database on sectoral credit that's the result of a multi year effort. And so to construct these data, we drew on more than 600 individual sources, many of them were newly digitized. Many of them were previously unpublished and we got them by simply contacting more or less all the world's central banks financial regulators and statistical offices. And a key contribution here is that we systematically harmonized the classification of sectors both across countries and across time. And we were only able to do that because we got a lot of help and we got help from more than 150 people working at the national central banks, the statistical offices, etc. And so the result here is a data set that covers domestic credit, mainly bank credit, but also bonds and non-bank lending where the data is available. It spans as I mentioned almost 120 countries and in many cases going back 70 or 80 years. And we'll make these data available as soon as possible sometime this year, including an extensive documentation and we hope that this will be useful for many other people as well and we plan to continue working on. Now just to give you some intuition, I'm strongly here comparison between total credit relative to GDP, averaging this across countries, where this is kind of the dark blue line down here. And then you know we're plotting that against the same data from the Bank for International Settlements and bank credit, which has been widely used. And what you can see is that these two lines essentially just overlap and both of them suggest that in the long run, the ratio of private credit to GDP has increased. What you see at the top is the BIS data on total crediting kind of encompassing all types of debt and you can see that you know we're tracking the trend here as well, but this has kind of a higher level than our data. So how I want you to think about our database is that essentially we can provide a breakdown of what actually accounts for credit to GDP that you can get from existing data sets. Now what's nice with these data is that they allow us to provide some stylized facts that we think are kind of extending existing work or our novel. And so for example, we can confirm with a much larger sample than previous work that it's really household debt that in that explains most of the increase in this credit to GDP ratio over the past 50 years or so. So you can see that in the graph here on the left hand side at the dark blue shading, this is household debt and there hasn't been as much movement in firm credit relative to GDP. But what's really novel here is that we can also look at the breakdown of lending to firms across different industries, which is what I'm showing you here on the right hand side. And so what this suggests is that the share of agriculture, but also manufacturing in mining in particular has really decreased quite substantially since the 1980s or something like that. And you can also see the role of lending to firms in construction and real estate. So that made up perhaps three or four percent if you look at, I say the early 1960s, if you look today, lending to construction and real estate. So these are not residential mortgages. This is firm trade makes up something like 25% of all outstanding firm credit. Now, next, I want to show you a case study that I think is very much representative of what we see happening in our data around major credit bulls, both in advanced and emerging economies. Now, as many of you know, Japan experienced a major credit boom in the 1980s that was followed by a banking crisis in the early 1990s, which I'm shading for you here in great. And what this graph shows is how the ratio of credit to GDP in different sectors has evolved over time index to be 100 here in 1985, but you can do this in many different ways. And what the patterns here suggest is that this Japanese credit boom was concentrated not only among households, but also among firms in real estate development and more maybe more importantly also other non-tradable services like for example, accommodation and food. So it's not just about kind of a housing story. There are also these other non-tradable services. Now, in contrast, if you look at manufacturing, you see that despite this massive ongoing credit boom, if anything, there was actually a decline in manufacturing credit in Japan during that episode. Now, another entirely different example here comes from South Korea, which experienced a kind of prolonged boom in GDP growth starting in the mid 1960s. And that boom followed a major banking reform that happened in 1965. And what we see in our data is a really sizable increase in manufacturing credit following that reform that is not mirrored by other sectors such as lending to households, construction, other non-tradable services. And one reason why that may be the case is because there was a huge policy initiative to basically force banks to lend more to export intensive activities. And to get at these types of patterns more systematically, we'll start by constructing three measures of credit growth that are motivated by these models of credit cycles with sectoral heterogeneity. And so that means we look at lending to the tradable sector by which we mean agriculture, manufacturing and mining. We look at credit to the non-tradable sector by which we mean construction real estate, retail and wholesale trade, accommodation and food, transport and communication. Although you can define that slightly differently and I won't make much of a difference. And we look at lending to households. And so a key question here, of course, is how do these firms in the tradable and non-tradable sectors differ? And of course, they differ in many underlying characteristics, but the first difference that's emphasized by many of these open economy models is that industries might differ in how sensitive they are to household demand. And for that, some data we're using here from the World Input Output Database suggests that domestic households consume something like 36% of the output of non-tradable sectors compared to only 15% of tradable sectors. So that means here that the non-tradable sector is much closer to final household demand than the tradable sector. Perhaps unsurprisingly, the tradable sector also trades more and has much higher export to value added ratio. The second difference that's emphasized by these models is that these financing constraints may vary across sectors. And here we find that small firms that may be more financially constrained, they're much more common in the non-tradable sector and real estate collateral is much more common in the non-tradable sector. This is true even outside of construction real estate. Now our third important difference is productivity. Here we're just using data from a paper by Manon Castillo and that suggests that both in levels and growth rates, the tradable sector is much more productive than the non-tradable sector. Now, because it's difficult to pin down empirically which of these characteristics exactly are more important, in most of the paper we follow the reduced form approach of the theoretical literature and focus on this high level distinction between tradable and non-tradable sector. Okay, so equipped with this kind of basic intuition, I'll now show you impulse responses from local projections that are commonly used in this literature. So what that will mean is we'll look at regressions of changes in log real GDP in a country I between a year T and a future date T plus H and we'll regress that on country fixed effects and then on different measures of lag credit growth. So we look at credit to the non-tradable sector, the tradable sector and households relative to GDP for now. And we're going to look out 10 years into the future and make sure to saturate these models with lags of the dependent variables of the credit growth measures to kind of soak up mean reversion in those variables. Now I'll show you two sets of standard arrows here, both of which take into account their serial correlation here. So, you know, we want to make sure that we take into account that these credit booms are not random in terms of timing and they're also not random across countries. And while the impulse responses I'll show you they don't necessarily capture causal effects, but we think they answer an interesting question which is conditional on using a particular type of credit expansion, what on average happens to GDP and other types of micro outcomes. Now the patterns you see here are the main result of our paper. What you can see is the response of real GDP to innovations in credit to the non-tradable sector on the left hand side and the tradable sector on the right hand side. And for now, these do not take into account household credit. And what these impulse responses suggest is that lending to the non-tradable sector firms in the non-tradable sector systematically predicts a slowdown in GDP after four or five years, but lending to the tradable sectors associated with stable or even higher GDP. And in the paper we show that these patterns are not only highly robust, but importantly, they also hold if we control for the composition of the economy. So if you control for the share of the non-tradable and tradable sectors, this picture will look almost identical and it will also look almost identical if you look at credit relative to value added. And so how I want you to think about this is that it's not just about a real location of economic activity towards non-tradables that's associated with lower growth. It's this real location towards non-tradables with an increase in leverage of that. Now from previous work, we already know that household debt is associated with lower growth. And so this is something we can replicate here as well if you look at the graph here on the right hand side. But importantly, even when we control for household credit, this leaves these patterns for non-tradable and tradable firm credit unchanged. And we also find a similar pattern when we replace the dependent variable here with unemployment. And here we can see perhaps even more strikingly that it's really lending to firms in the non-tradable sector that's associated with spike in unemployment in the years going forward. And for these other types of credit, these patterns seem to be somewhat more muted. And what that really suggests is that we have to think about firm credit as well when we want to think about macroeconomic fluctuation. Now to this point, we've taken this kind of reduced form approach following many open economy models and only differentiate between tradable and non-tradable sectors. Of course, we can also just split firm credit using potentially important variables that are kind of the primitives in these series of credit cycles. So what I'm showing you here are the results for splitting firm credit based on the median value of two of these characteristics. So in panel A here, we're splitting credit based on the proximity to household demand. And in panel B, we're splitting it based on the share of small firms. It's kind of one of those proxies for financing constraints. And what you can see here is that these results line up quite closely with our main distinction of tradable and non-tradable sectors where firms that are close to final household demand or industries that have many small firms, they predict kind of a boom-bust pattern in GDP growth while if you're far away from final household demand as an industry, if you have very few small firms and maybe you're less financially constrained, these types of firm credit just have a positive and kind of constant correlation with the GDP over time. Now, an alternative approach here is to ask what is the path of GDP look like around the peaks of different types of credit booms. So to look at that, we first identify slightly over 100 credit booms in our sample based on a detrended measure of total credit to GDP. And then what we do is we divide these booms into two buckets. The first bucket is what we call non-tradable biased credit booms. So these are 75 cases here where during that credit boom, there's an increase in the share of non-tradables and households. The remainder goes into the tradable biased credit boom bucket. So these are cases where the share of the tradable sector in total credit increases. And so here you can see, if you look at the orange line, that it's really the cases that look like Japan where you see this increase in the share of the non-tradable sector in total credit that are associated with a slowdown in economic growth going forward. Well, if you look at cases that look more like Korea where you see an increase in the share of the tradable sector, those do not look like kind of economic crashes. If anything, you kind of see just growth just continuing or only slightly slowing down following the peak of the credit boom. Now, what we've seen so far is that credit to the non-tradable sector and households is systematically associated with lower economic growth and higher unemployment in the medium. And so models of credit cycles with this sectoral heterogeneity, they mainly emphasize three key predictions that could give rise to such a pattern. So number one here is that lending to non-tradables and households should be linked to a real exchange rate appreciation and should be linked to a real occasion of economic activity towards non-tradables. And you get that, for example, in work by Schmidt-Groen Uribe, the 2016 JPE paper. Then second, lending to non-tradables and households should predict financial crises. Once you're in a crisis, the losses, they should be concentrated in non-tradables. And you get that, for example, in the work by Schneider and Tornell. Now, third, lending to non-tradables and households should also predict lower productivity growth where in work by Ricardo Reiss or Beninio and Farnaro, this is because the tradable sector is just more productive. So if you give more money or more credit to the less productive non-tradable sector, that will kind of slow down the macro economy. And so in the remainder of the talk, I will show you some empirical evidence that is consistent with all these predictions. Now, first, in the data, we find that changes in credit to non-tradables and households are associated with an increase in the ratio of non-tradable to tradable employment. So that's what you see here in column one. That is intuitive. But we also find that lending to non-tradables in particular is associated with a real exchange rate appreciation, which you can see here in this column two. And so what this might suggest is that these types of credit expansions really boost demand in the economy rather than the productive capacity of the economy. Now, next, we turn to look at differences in financial fragility across sectors. And so our starting point here is some data on banking crises, which we take from some joint work by my co-author Baron Vernon-Chieng in 2020 and a widely used work by Levin and Valencia. And so here I'm just replicating the existing literature, which has found that before the onset of the systemic banking crisis, we usually see an increase in the ratio of credit to GDP. So here we kind of get positive changes in credit to GDP. But what our data allows us to do here for more countries and existing work is to first look at patterns of firm versus household debt around crises. And here really the key takeaway we have is that household debt tends to expand somewhat earlier than firm-traded around these crises. But more importantly, we can also drill down and look at what is it that accounts for the growth in firm credit in the run-up to financial crises. And here we find that almost the entirety of that firm credit growth before financial crises is driven by the non-tradable sector, where there are very few changes in tradable sector credit. And what this key here is that this non-tradable sector firm credit expansion is not only driven by construction and real estate, which you might imagine, but we see a very similar pattern for other non-tradable services such as trade, food and accommodation, or to a lesser extent transport and communication. If you look at tradable sector credit, if it's agriculture manufacturing, especially these are just flat around the incidence of crises. And we can also confirm these findings using standard predictive regressions where the dependent variable now is a dummy for the onset of a crisis and independent variables here at different measures of credit growth. And what we find here again is that it's really lending to firms in the non-tradable sector. That's a statistically significant predictor of future banking crises, even more so than households, while tradable sector growth is just essentially uncorrelated with the incidence of banking crisis. And just to give you a sense here of the magnitudes, the probability of a banking crisis is something like 3% unconditionally in our sample. If you have a one standard deviation higher growth in credit to non-tradables, that jumps to something like 9% or 10%, which we think is quite sizable. Now, a key question here is why exactly is it that these non-tradable sector credit expansion seem to matter so much more for the likelihood of a financial crisis. And here, theory again gives us clear guidance where in particular models like Schneider and Tornell 2004, the source of the crisis, the immediate source of the crisis are large steel defaults in the non-tradable sector. And so in the data, we indeed find that the non-tradable sector is really critical for understanding defaults after banking crisis. So here in these figures, I'm focusing on a case study, which is the Spanish banking crisis of 2008. But in the paper, we also have a graph now with the total of 10 banking crises for which we were able to gather these data. And so let's start on the left-hand side. What I'm plotting you here is the ratio of non-performing loans to total outstanding loans by sector. And I think of this as kind of a sectoral default rate. And so what you can see here is that following the 2008 banking crisis and kind of going into the eurozone crisis, the default rate of firms in the non-tradable sector was around twice as high as that in the tradeable sector. Quite a pronounced difference. Now, as you can imagine, credit before this Spanish crisis was really concentrated in the non-tradable sector and housing sectors in particular. Now, taking together with a higher default rates there, what this meant was that the vast majority of non-performing loans after the Spanish crisis was concentrated in the non-tradable sector. So the share of non-tradables in total non-performing loans in Spain, 56% according to our estimates here, really much larger than the non-performing loans accounted for by households. And they make the NPLs in the tradeable sector and others irrelevant. And so what this suggests to us is that if you want to understand the non-performing loans that are kind of the likely immediate source of banking crisis, you really have to look at firm credit. And you really have to look at firm credit to the non-tradable sector in particular. Now, we also find some evidence that the sectoral allocation of credit is associated with differences in productivity growth. And so in particular, lending to the tradeable sector predicts stable or even higher labor productivity growth, total factor productivity growth going forward. If you look at lending to non-tradables and also households, that tends to be associated with lower productivity growth. Again, both in labor and TFP terms. While we want to be careful in interpreting these patterns, they are these consistent with the intuition from these models by Ricardo Reis and Benino Fomaro that credit flowing away from the tradeable sector towards sectors with lower average productivity could reflect a misallocation of resources. Now to conclude here, we argue in this paper that the sectoral allocation of credit is important for understanding these linkages between credit markets and the wider macroeconomy. And we believe that the patterns that we document here provide a bit of a new perspective on why credit is sometimes associated with economic growth, this kind of finance growth view. And sometimes and perhaps more often with economic slowdowns, this kind of credit booms gone bust for you. And so we also think that our results have a number of potentially interesting implications for theory and also for policy makers, which is that number one, we think they show that this heterogeneity in firm credit is really important for understanding credit cycles. And differentiating between household and firm credit and housing and non housing credit, a lot of focus has been put on that both in terms of theory and regulation but we think this is very important but it's not, perhaps not the entire story. And of course, last but not least we also want to be super careful here but if you take these findings at face value. And they do suggest a potentially stronger role for for sectoral regulations, although there are many many caveats that you could come up with why this might be a bad idea, not least political constraints with which I think I potentially quite important. So, let me just take the opportunity again to say thank you for the privilege of getting to present all work here today and thanks so much everybody for engaging with it. So thank you very much custom for the very nice presentation and also you know I will give time later on also to a meal to to intervene. I think that thank you also to be on time actually earlier. So we are moving now to, let's say, the discussion or they say part of the panelist. So the first one is Philip Hartman from the European Central Bank, and he has of course a lot of comments on on your paper. Yes, thank you I need the sharing option enabled. No, thank you. So now I can. Yeah, in the meanwhile of course I don't think that I need to present you Philip you have a long, let's say series of paper of research and clearly this topic is very close to your agenda. Yes, thank you so much you should see now the my presentation, can you see it well. Let me do it. You can make it a little bit bigger. I do it in full screen now. Yes, that's perfect. Okay, so yes, thank you Lorena that I'm, I'm very happy to to just to to make a contribution to this nice ceremony. And I, you know, obviously could do a long loud ratio about what's so great about this paper and so on but I will gloss over this a bit quickly and come to the essence quickly. But let me just say that. So why does this paper to preserve the price it provides a public good to the research community. Great compliment to Jordan Schuler, Taylor, make it available online as soon as you can when you have the latest data. It finds evidence for an interesting and theoretically well founded hypothesis about the non-tradable sector and it does us a favor to alas harmonize to reconcile micro and macro again after a somewhat frustrating paper by Moritz Schuler short while ago. It's inspiring for policy I will say in my remarks later that actually I think it's a nice input into policy making very much in line with the current policy process. And actually by doing so, I realized while trying to do some charts on these type of issues that you have in the paper that we absolutely need more harmonized sectoral credit data for the EU and the era, the even in the EU and the era, you have to go to the national sources in order to know how much is flowing in the construction sector in one country and how much in the accommodation sector in another country and so on. And I think for macro potential policy, this is critical that we have data right away rather than having to use how unharmonized and the paper is relevant for Europe. So, and so it resonates very nicely with the story of the European financial crisis and the role that non-tradable sector played in there. Just want to warn that there's an additional component that's not the focus of this paper in that crisis, which is the short interbank lending that created the big southern spot which partly finance the flow the capital flows to those sectors. But let me go to the essence so basically what I will do I will make one or two remarks on the paper very quickly but then I will go applying it to current situation to Europe or the last 10 years say to Europe and to kind of policy relevance. Before I do so here in this chart. I put down what Carsten just presented as the main result of the paper. And this is the business cycle part not the financial crisis part. What I want to alert you to is in the business cycle part that if you throw in the household credit into the estimations. The non-tradable credit survives, but one sees it's kind of weaker and less statistically significant, which is as he just explained not actually the case in some of the crisis regressions and he provides a battery they provide a battery of evidence that how relevant the non-tradable sector is and in some crisis in cases even more than the household sector. But what I want to point out here is just obviously and the authors don't claim that but just for the general audience, the household relevance of household credit for systemic crisis is not dead at all. Anything in the business cycle context is still slightly stronger and more robust, but that may not be in every crisis and that it's not so significant in the business cycle may just be a usual lesson from history is that even though we find a statistical regularity. It may sometimes be that actually you have to still look at the cases in one case it may play a bigger role the non-tradable credit in others less big role and that may be reflected here. Let me go to Europe today. So here I borrowed from what Moritz Schulerich presented a few days ago in the ECBU forum on central banking, because he actually replicated part of this for more recent data because your data stopped in 2014. And he wanted to know whether the corporate credit the non-tradable corporate credit is actually a problem in Europe today. And so what you see here is the red line is the US, the blue line is France. These are the countries that's all normalized to one at 2015. And you have none left on the left and untradable on the right the tradable sector, and you see that actually in the countries that he covers here. The problem is not so pronounced in the last few years in Europe, except for France to some extent and the US which is not not our So it doesn't seem to be a generalized issue in the euro area at present that we are in danger of a non-tradable corporate credit boom that may turn into some type of crisis. Now let me talk a few words about the French case because this seems to be the exception here. There has been strong credit growth in France for a while. It's not limited to the non-tradable sector. You see on the right it's also to the tradable sector, the corporate sector and also the household sector. I have a backup slide where I could show you that also the household sector is quite similar. Now this development would also be more attenuated if you would calculate these figures on a net debt basis, rather than a gross debt basis, because in our days, many companies seem to be cash rich and nevertheless take a lot of debt. So that may actually be attenuated this dash blue line if you take this into account. Nevertheless, since this credit issue in France is very well known has been addressed by the authorities. First of all, through a very interesting macro prudential measure, which is the tightening of large exposure limits by banks to highly indebted French companies. This is a special macro prune measure under this famous article 458 of the capital requirements regulation, which is still active today. It was adopted in July 2018 and it is still valid. It was also adopted a more broad based counter cyclical capital buffer and one year later in July 2019, but then released it of course when the COVID crisis hit. So, so I would argue that the general impression that even there may be other countries where non-tradable sector corporate credit at present is a major concern but here I would say it's not special compared to other credit and it they have been some measures to address it. Okay, then the next thought is about, you know, is, as you yourself address is this a traditional construction and real estate story or is this a is a really peculiar story for also non non real estate related sector so I put it here. What we did here we took the, the authors data together with the Schuleric data to extrapolate the authors data to more recent data and broke down the non tradable sector for two countries France and Spain into the sub components into construction real estate wholesale and transportation communication information in green. So what you see is that the French case is actually where the real estate part or the construction part actually dominates. However, if we turn to the right where you have the Spanish case where actually most of us would think well they should really have construction in in the forefront and real estate. You see that actually in this case, it's not dominating so actually wholesale accommodation food and transportation communication information is actually similarly important in this cycle that you see here. And, and actually I could even I didn't show it but they could show you a third case where actually you see that actually Austria is one I think Germany is another. Where actually the transportation for example would dominate the other two sectors. So yes I agree with the authors this is not just a resurrection a re rediscovery of a, even for the present European situation for the for the big role of housing booms and busts for business cycles and financial crisis. All the sectors the sub sectors that how the authors are able to decompose or compose them for a long historical database actually are kind of relevant depending on the country and depending on the time. So that leads me to my last part which is the question the relevance for policy. So first of all, as the authors point out this has been embraced by the Basel community to have a work on sectoral corporate credit cycle, counter cyclical capital buffer sectoral counter cyclical capital buffer there was a working paper, and there was even a guiding principles issued a few months later in 2019 about how to operationalize such operationalize such a corporate sectoral sectoral counter cyclical capital buffer. In Europe. Actually, there's also already things being implemented. The capital requirements directive and the new capital requirements regulation that became that actually were adopted in 19 and came became applicable as of January this year, except they haven't implemented the CRD five yet actually contains a systemic rig by sectoral systemic ricks buffer option. And this is systemic risk buffer can the sectoral systemic risk bar can be used for actually in a similar way as if it was a sectoral counter cyclical capital buffer by the countries, the European Banking Authority has already issued a guideline about how to deal with the sub sectors which subjectors can be covered and how can you choose them. Obviously they have to be systemically important and so on and so on but other than those criteria actually you can go to the sub sectors in the European market potential framework now and apply systemic risk buffers, even in a particular way, for the purpose of dealing with or addressing the banking risks related to the to non tradable corporate credit booms. No country has used it so far obviously because we are playing still in the crisis so we have other problems to deal with right now. That's my last slide. So obviously, as you should from a policy side so we are in reasonably good shape. The research is accompanied by frameworks policy frameworks at the G 1020 level but also at the European level at France I mentioned has done something but obviously there are a lot of covers as also I think Carsten finalized, obviously by adding these more granular instruments we making the framework more complex. Think about if you have several sectors that move in certain directions and you add to the overall counter cyclical buffer this the sectoral ones it can be kind become really quite complex. The process is still unreformed it's quite complex to actually adopt these measures and as countries have proven to be a bit slow in actually using these counter cyclical buffers which left us with a situation where the macro potential space to release them in the downturn situation was actually not so large actually rather tiny, but maybe having narrow measures actually makes it easier to adopt them rather than the quote unquote nuclear option of a cross across the board. Now, then another caveat is obviously this is only for banking. This doesn't cover non bank financial intermediation we have said that very often in the ECB we would like to see the European framework to evolve more towards the non bank intermediation where the credit growth is growing more and obviously the corporate debt securities widely held outside bank balance sheets are also not covered here. In adopting these measures you have to look at the trade offs and do cost benefit analysis and mind about leakages and let me stop here and and you back to Lorena and Shabnam. Thank you very much Philip also for being on time I think that you rise several points and I will give them the opportunity to the author to answer to your question but now is the turn of 7M even for 7M she's clearly member of the advisory scientific committee she's doing a lot of research on small and middle enterprise also with with an eye on what happened during the COVID in terms of don't perform loans and so on. So again, we are very curious to know what you are thinking about this paper and what is your view about the issue that they are trying to address. Thank you very much, Lorena it's a pleasure to be here let me just make sure to see my slides in the full review is everything good. Everything is fine, perfect. Okay, thank you so much it's a pleasure to be here and talk about this paper. So my talk is going to be first of course, first and foremost congratulations to the authors it is very impressive work on credit and macro fluctuations. I'm going to try to highlight how impressive the contribution of this work and also the policy elements, and then combine it with some new insights from Europe and United States. Now that doesn't come from sector level beta but firm level beta. So, and this will give an idea to you, I hope that I mean what the author started has been great and now we should push this agenda further to a more granular level so we should think about where does credit go credit allocation not just by sector but also by firm and and authors work makes this extremely clear. Okay, so let me start by putting this in the context of the literature so there's an extensive literature first of all linking credit cycles to business cycles. This literature, in fact, has been around because this was the story of emerging markets for a long time. But of course it came to even on, you know, under a bigger line after the 2008 2009 advanced country crisis right even crisis European crisis. And now the work has been done to carry this to a historical setting in financial work by Jordan Shuler Taylor, and also me answer the burner and other papers. So they these papers show that, you know, in a historic context, which of course, we all realize that you know advanced economy is going to be what is out there, especially when you go to 1800s household credit is extremely important. Then, as I said, we know firm credits extremely important emerging markets of course this, you know, I don't cite all the all the papers here but this literature goes back to 1980s, starting with Latin America that crisis first focusing on the sovereigns and state on firm and then extended to Asian crisis, Argentina, Turkey, I mean, there's this understanding that, you know, the firm credit is extremely important. A recent incarnation of this work is shown in my Jackson Hole paper in 2019. And the recent forthcoming a restart paper using very detailed granular credit register data, which exactly tells you which firm gets what, which again seems to enforce this conclusion that you really cannot ignore corporate. Now, the debate is a little bit on, okay, can we also say firm credit is important for advanced economies. This is going to be harder because of unfortunate regulation. Most of the advanced economies firms are not going to require to report regulatory authorities, the debt actually. And that's exactly why emerging market literature is more grounded because they started collecting that type of credit register data. Well, before right in the 1990s after the crisis, and that's exactly the problem right so and that's exactly where this conflicting results and the debate in the advanced economy literature is, we have some new papers looking at us. So what I cite on this paper is exclusively on us using actually micro regulator data matching the census data, and those papers show actually form credit has been very, very important and investment. Again, not an historical set, but at least since 1990s, and especially since 2000s, early 2000s, it has been extremely important in driving boom bus cycles. Even it's not a financial crisis like 2018, even just a regular like recall the 2001 recession in the US regular business cycle recession is very much linked to firm credits just that it is really, really to get at this data in advanced countries, especially in the US I'm going to talk more about that. Now, at this juncture, the work by Miller and Werner comes I think it is great right because no one is like, look, I'm going to use a lot of countries and I'm going to take this back historically. Of course, I cannot do this at the firm level because of the historic dimension and a lot of countries dimension but I'm going to do that the sector. I mean, this is I think amazing because this really bridges these two literature that they like also that non tradable sector is really what matters. And of course, you know that most of the firms are in the non tradable sector, the declining role in manufacturing since 1960s. So, and they are saying, you know, these both these type of credit is going to be very important for this boom bus. It's a very important message, very consultory message bringing many different strands of the literature together. Okay. And then they also have a very intuitive backing to these facts they show they say like, look, this is going to come from three channel versus the consumption boom story which is obviously going to be very relevant for household credit financing demand and consumption. And they say this is also going to be relevant to a certain degree non tradable sectors. Again, this is the lesson we've learned very well in emerging market crisis literature since 1990 and that's exactly how they refer to open kind of model a lot of model actually center around this narrative. They, they don't stop there they are going to say that look there are other stories financial fiction story and celebration story. And they are and that's again very intuitive because half of the non tradable firms in non tradable sectors, of course, they might have title financial constraints and they might be less productive and having the resources allocated one is going to be a problem. Again, here, I think it is very impressive that their work is connecting different pieces in the literature. They can speak to the papers that says there are negative effects of corporate debt overhang on outcomes. And they are basically telling us like, look, maybe this is coming more from firms in the construction sector, but now maybe in the firms from firms from manufacturing sector, which is of course extremely variable for for European. Okay. Now, as I said, it is very impressive and policy relevant work. I have nothing but praise for this amazing work by mother and burner. And let me also tell you that they also not only like, you know, show amazing facts and and connect the two strands of the literature and many other pieces together. They also provide me a great service with a profession. They are constructing a brand new historical database from several countries bought at not an emerging not just advanced at the sector. And they are their key. A driving force here is this very important question that we should all focus on and as policymakers and I think if you really want to understand the link between finance and macro. This is the question to focus on where does credit go. Okay. And this is exactly their driving force guiding life is whatever you want to call it, but they are exactly on the right path. And let me tell you, constructing these type of new data set is, is, I mean, it is a huge, huge deal. Since I have been involved pretty much in the last 15 years of my career, many of these type of projects. This is, this is, this is huge. So it is not only that they spend a lot of time on this. They are going to make this available. So it's a great service to profit. So what I'm going to do in the remaining part of my presentation, I'm going to present some results from further database literature and then bridge the policy implication that that literature is telling us and what the work by Miller and Berner telling. And here, let me also tell you this. This, it is the same guiding principle. The first level that we also exactly starts from the same question that these alter stuff where does credit go and the state of the art in this literature right now in terms of data construction and measurement is the credit register data. Okay, which is why us is such a big problem. And Euro is also a problem. It is getting there. ECB on a credit is a great start, but it's only at the household level. We of course know different countries in Europe like Spain, Portugal, until they have great credit register data to going back. Unfortunately, we need this harmonize at the union level. I mean, in a way, Miller and Berner doing that, right? They had a harmonized historical data at the sector level for many countries. So the goal here, and this is a very high bar, but if we really want to do proper quality and if we want to avoid policy mistakes, this is where we should be going. Okay. Doing this credit register data set. So which is going to tell us exactly where does the critical for each household for each firm, more harmonize for more countries. And that is going to be like amazing. And this is basically what emerging markets have, right? They didn't also combine it, but they have it. I don't, I can't imagine any emerging market right now not having these data sets. Why because they learned their lessons from many financial right. And that's exactly what ECB down with on a credit. But I think they should push further on the corporate side, which is actually what you asked. So, I mean, this is the bittersweet for me because I really want Europe do this first. When they started on a credit, I was, oh my God, great. This is one thing that Europe is going to beat us. Unfortunately, it didn't happen because when they started on a credit us has also started. Frank and started collecting credit register from firm. And now US still has has the better data here than, than, than Europe, but I ensure, you know, Europe will go the same way. But I will, I will clarify this further. Okay, let me show you this figure. So this is the corporate data investment to GBP for Europe and US. And this is the last of 2018 year. And, you know, looking at this figure, it is very hard to say corporate debt is not an issue for outcomes. It is an issue. It's just that we have to collect it from corporate right and that's really what it comes down to. So here, everything is normalized at the start of the interaction of euro, you see that periphery countries as an amazing corporate that boom, you're already overall also United States is less. So of course, now, you know, you extend it till 2020 United States is going to be on the optic, especially if new regulatory data and not the full funds and I'm going to tell you why that's exactly the case. And here on the right, there is the investment counterpart and you see that again, everything's normalized. So, you know, it's kind of like there's an investment boom and like, you know, basically the big picture message here is everybody falls off the cliff and you see that periphery has been recovered by 2018. Okay, by 2018 investment, corporate investment still didn't recover from in periphery. There was some recovery and going down in United States, some very sluggish recovery in your area. But basically this does tell you that like, you know, this corporate debt overhand is something that is really holding back investment in periphery open countries. Clearly in my paper with them. And then, okay. So, and here's the imposter responses for that that looks at like, you know, high leverage firms in the periphery countries on the top, and the low leverage firms in the center country on the bottom. Of course, there has been a lot going on during these European crisis episode in Europe, but basically, you know, accounting for those things and that's what the figure is trying to do like, you know, all foreign bank linkages and all the other problems. You have that firms with high leverage. So, the zero here is 2008 drop investment the most. And then they really didn't come five years out of the crisis. And as low leverage firms in the center European countries, this didn't happen. And then we have added exercises that linking back with the investment that shows that this corporate debt overhand has been a big driver of not recovering from the investment in Europe. So, why, let me just say one thing here about the sugar paper that Philip refer to why that paper cannot pick this up, because aggregate data and we are is not going to be able to pick that. Okay, and let me just leave it at that. And I can answer questions later in Q&A but he showed this very clearly in our new US paper, using regulatory data from Federal Reserve is just that you cannot pick that up with the aggregate macro level rate, which again, shows us importance of going sector level in Northern Europe. Okay, there is an important dimension, the boom dimensional. We know what is happening in the bus. In the bus, you know, households and non-tradable sector really hurts you in the bus. And, you know, I'm basically saying the same thing at a firm level. Look, firm credit during the boom really hurts you in the bus. Okay, so I just put what Miller Werner said at a firm level but it's the same. Okay, so why these firms, you know, allocate accumulated credit so much during the right because if you want to think macro potential policy and regulate these firms and these sectors households and non-tradable sectors, you have to understand what they are going on in the book. Okay, this is the boom here and this is from my Q&A paper on location with Gopinat, Kara Parfonis and Gula Verstensian. Again, it focuses on Europe. And this paper argued as look, declining interest rate, which is completely normal part of the EU integration process incentivized from to finance investment with Gopinat. Okay, so this is a channel actually in essence similar to demand and also brings the financial friction story together like that. It combines two channels, actually three channels, demand, financial friction and misaligned that is highlighted in Lodren Werner. So this actually starts in the Mastery Treaty. So this is the interest rate on short term debt for corporates in Europe in blue and what we have in the model. Why is it huge? You can finance now short term debt at a much lower cost and of course you are just going to borrow more. The problem is the heterogeneity in financial friction, right? If you have terminal heterogeneity accessed in finance, in Spain, in Italy, in Germany, and this is not a country or sector level heterogeneity. This is a within sector terminal heterogeneity. Then you are going to have a lot of problems in your life. Why? When you have such a real interest rate decline, everybody wants to borrow and everybody wants to invest, but some firms, these are going to be the larger firms, higher network firms. They can borrow and invest. They are going to face low returns to capital. Other firms, small firms and low network firms, they unfortunately cannot borrow and cannot invest. That means they are going to face high returns to capital. That type of dispersion in returns to capital is going to draw that product down. And as we show in our paper and also in many, many other papers, this was from a Southern Europe experience declining productivity in the first 10 years of Europe, as opposed to Northern Europe. And that's really about these differential financial frictions. Specifically, if you want to pin that down heterogeneity, that is a size-dependent bottleneck. So, in the extreme, you have a scenario where large firms borrow, small firms cannot, and the composition of that type of size distribution in your economy is going to affect it. Is there any evidence for that? Yes. This is from Southern European countries, and if you plug actually the same thing from Northern European countries, you are not going to see this. But in the Southern European countries, interestingly, there is a size-dependent financial friction, meaning large firms borrow more and more. Okay, so this size is on the x-axis, borrowing leverage on the y-axis, and this is the convex relation. So there are millions of firms here that says, largely you are, you are going to borrow more. Okay. Now, let me take a stock and link it back to the Miller-Burners. So Miller-Burners' story told us that mislocation is very important, and then they were talking about households and not construction sector. So here, I'm telling you that there is mislocation even within the trade-offs. I just showed you a figure for the manufacturing sector for all the Southern European countries, but it's at the front. So this says that, I mean, you know, what Miller-Burners is telling us is extremely important, you know, obviously if it is even there, within the manufacturing sector, obviously it's going to reverse for construction sector between sectors and all. And the other story, financial friction story, the same thing, right? So financial friction is going to be tighter for households and construction, but I'm showing you that it is even there within manufacturing. That means, you know, we really have something very serious on our hand because that means it's going to be applicable for a buyer. What is the policy implication? The policy implication is obviously you are going to regulate by firm on top of the household, on top of the sector regulation that Philip told about, which is of course a great set-up on Basel, but if you think about it, actually, it is very hard to pull off. And I think here is one place where advanced economies really should look at emerging markets. Emerging markets did try to, they did try to regulate construction and households and not manufacturing because of course they have their exporters, their beloved exporters in manufacturing, and they love to keep those exporters happy, right? They did try to. It failed miserably. Why? Because the construction sector goes in a joint agreement with your exporter sector, goes to the bank and gets them, okay? So this completely leads. If you try to regulate the construction guy and not the manufacturing guy, it is actually going to fail. We know that it failed in emerging markets. This is actually, is really good. What we learn from the Vermeer paper and what we learn from the literature, look, I actually can do better. As I am regulating households, when I'm giving them debt, I look at their LTV and all that, I can do this by firm, right? Now, the big question is for this implication, how do I do that? I mean firms are obviously going to, all firms are going to try to find people around that. Now, the big advantage is most firms are going to be bank dependent in Europe, so you can regulate them through banks and that's actually what emerging markets did very successfully. All right. What about yes? Let me say one thing about yes and then I'm going to conclude. So is this something only applied to Europe or only applied to emerging markets? No. Unfortunately, US data limitation is extremely serious. What do we use in US for all funds? Okay, for all funds, it is going to be based on tax records on income and assets. Unfortunately liabilities and debt is going to be a residual debt. So you are going to miss all the small firms who are going to borrow from banks. Corporate bond markets, of course, very deep in US, not much budget in Europe, in US they are, but that's why we go and rely on the bond market data and miss all the SMEs, which is actually accounting 70% of US employment and 60% of US output. So if we miss all that and then a crisis like COVID happened and then they say, why is it disconnected between Wall Street and the Main Street? Why I have this huge unincreasing unemployment, then the Wall Street is like breaking profit after profit. So I don't measure that cost. Of course, I cannot understand this. So here's where the US new regulatory data comes. It is an amazing data set. It's basically as credit registry known as FR by 14 federal user by 14 data. It is started being collected as part of the Frank Act after the crisis, very similar experience to emerging markets. That's what emerging markets is. And that's what ECB did to down a credit. It's just that ECB didn't extend it to firms and they should have actually. This is the picture you cannot get some form of fun, right? So this says when I look at the by 14 firms, by the way, this is regulatory data. So banks reporting to federal reserve exactly the loan they give to the firms and then in the paper, I show that these firms are going to account very large part of US employment investment output. Then I look at their balance sheet, the large ones, very large ones, only 30% of their balance sheet is bank loan. The rest is in the ball. Okay. But so this is the ones over 75% out of asset distribution, this bottom line. But if you look at the firms in by 14, less than 75% out of the distribution, the entire balance sheet that is banked. Okay, they are not in the ball market. And this figure is also going to be the case for Europe. In fact, in one of the latest things we did for ESRB, we show that very, very little part of the corporate sector in Europe borrowed from the ball markets, largely they borrow from bank. Yeah, and there's nothing different between US, once you know where to. Okay, what does it mean? Well, in US, we are going to get the exact same complex relationship of the size dependent borrowing constraint, once we look at these private firms. So this is exactly the same relationship I showed you for certain European firms, leverage and size of the important you see the blue convex in 2006. In 2009, at the heyday of the Lehman collapse, this goes down because not everybody becomes constricted, right? But in no time there's no such thing as public firms contest that list of firm data that is widely used in the US. They are not to settle. This is right and still figure 90 in 2006, nor in 2009 actually. Why? Because these firms can go to wants, these firms can go to equity, these firms can do whatever they want. There's no such thing as financial constraint. Then we talk about Amazon's and that's that's okay. And then finally, the linking the boom bus cycles, what we did and this is basically matching the census data to this. So this is the employment group at the sector level and revenue growth at the sector level in US census. You see that, you know, it is going to be positively correlated with short term leverage during boom, right? So the firms who increase the leverage, increase the credit is going to register higher growth at the sector level and pulling that revenue during the crisis. They are going to crush. That's the boom bus cycle and it states it is person. Okay, that's basically what these regressions show. Okay, what is the implications of policy here? Again, so what I did is like basically confirming what we're saying at one final granular level using firm level data. Of course it says the importance of factor for measure regulation for household sector leverage corporate sector leverage. Now, to do the corporate sector part right. We have to collect regulatory credit data for every agent. In European context, this means extended credit to European firms in US context. Thank God we now have my 14 and that's exactly how we understood during calling large firms doing on their credit line small firms couldn't. And that's why the PPP program came in the actual design a very good program in US PPP based on that observation, right? And PPP is completely designed for SMEs firms that's now 500 employee and it is a pandemic long term to a grant if you keep your That's what we need. If you don't have that thing that we cannot design a PPP program in a week in the middle of the bank. Okay. We're telling us that we need to watch the household leverage and non-tradable sector leverage. More granular look tells us that you have to watch the leverage non-productive large firms for sure. That's also a time bomb there. And you have to watch what is going on with financial constraints small firms because of course those firms are going to borrow more and more during low interest rate environments, which is what we haven't. So in terms of aiming for policy promoting growth, limiting boom bus cycle, we should limit leverage on household and non-productive firms and make sure high-productive firms have access to finance, especially during pre-soft mortgages. Thank you very much. Thank you very much. You know, you give to us a lot of information and also clearly pose a lot of questions to Karsten and Emil. So I'm suggesting that I'm giving now on the floor to Karsten and Emil first to maybe react to all the questions that first Philip and then seven and, you know, pointed on your paper. I saw ready that there is one question on the chat, but in the meanwhile that you're answering, you know, on their points, I'm suggesting that other are going to point question right question on on the chat, please. So I don't know who is starting Emil Karsten. Thank you, Laurie. And yeah, I can start and thank you so much, Philip and Shevnam for the really insightful discussions. I think I just want to pick up on three of the points that were raised that I think are useful to talk more about. And I'll just take them in order. So Philip mentioned sort of this connection when we look in the data and we look at household credit booms and non-tradable credit booms. At least when we look at growth, you see that these household credit booms, they seem more important for subsequent growth slowdowns kind of once we once we control for them in the data, at least in the business cycle part. One point I think that's important to keep in mind is these often really do go hand in hand and I think that you sort of have to impart lump them together for two reasons. So one is kind of this demand channel that we talked about where if, for example, you have a credit boom where households obtain more access to credit, that's going to indirectly benefit disproportionately the non-tradable sector. And they're going to rise in their activity and they're probably also going to borrow more just to finance, for example, that working capital, right? So actually often when we think about sort of just our simple separation of good and bad booms based on sectoral data. In the end, we want to think about households and non-tradables really going sort of hand in hand. And there's kind of some correlation across those. But that brings me to the second point, which is actually I really like the focus on the different case studies that you see both in the 2000s and also in the 2010s looking at Spain, for example, versus France. Because in the paper, and this is something that we're doing more work on actually investigating the episodes one by one and seeing, well, what was the narrative account of what happened during different credit booms and crises? And then what do our data tell us about which sectors were sort of playing the prominent role? And there we see, as Carson also showed, for example, for Japan and Korea, we see some interesting commonalities, but also some differences. So sometimes it really is also credit booms, especially in advanced economies that tends to be more prominent, that households are really leading along with construction and real estate. And then sometimes households are sort of taking a backseat to rising leverage more and other non-tradable sectors, construction, but also sectors like food accommodation and so on. And so there are sort of different varieties of credit booms, and this is something that we're actually currently working on. They're not always the same, but they have sort of these common features. And what's interesting is not just their commonalities, but also some of the heterogeneity that you see. The third point I just want to make is following up on Shabnam's discussion, which I thought really made sort of an excellent point of thinking about we've had this macro literature that's looked at aggregate credit cycles. First, we had aggregate credit from the IMF's International Financial Statistics. Then we went to households versus corporates. We saw for advanced economies it seemed like the household sector that was a better predictor. And then you had this micro level literature that especially focused on emerging markets, where we know in emerging markets, household credit is a much smaller segment of the credit market, especially when you go back in time, although households have been borrowing more and emerging markets. And there was a little bit of a disconnect. And I think that in some ways we're sort of moving toward each other. We're moving from the sectoral perspective where we have, you know, in our data, for example, typically we have one digit sector. So broad sectors like manufacturing, construction, you know, real estate, mining and so on. And we can think about the characteristics that those sectors have as being, for example, differences in financing constraints, how differentially sensitive they're going to be to episodes of credit supply expansion, for example, or other factors. And then in the micro literature, you can really see how, even within sectors, these sources of heterogeneity that we're pointing out that matter across sectors also matter within sectors. And so, you know, I think hopefully in 5, 10, 15, 20 years, one day we'll get there, we'll have sort of a fully integrated perspective and it's going to require this accumulation of data at the micro level, where we can think about the heterogeneity both across and within sectors. Because I definitely, I completely agree that within sectors as well, within manufacturing, you have some firms that are actually less tradable, producing more for the local economy, you have some firms that are more leveraged and we know that these measures that come from micro data like the high yield share, for example, the share of risky debt that's being issued, or the share of debt that's going to highly leverage firms that also contains lots of important information for policymakers about what's happening in the credit cycle, and the riskiness of that. I think that in some ways, where we're sort of taking one step, and I'm actually very excited about sort of the future agenda and the way that Shevnan is pushing us on that. So let me just stop here and then we can maybe take some other questions that I fear popping in. Yes, so let's start with some of the question and eventually, you know, even Karsten can catch up later on, also on some of the other points. So the first question is from Francesco Manzaferro, he has actually two questions. The first one is, you know, is it the case that any real estate boom is leading pretty much to a boom and then to a crash? Or is there any evidence that you have some residential, let's say, real estate boom that was a good thing. So it contributes to the global wealth creation, but it was not generating, you know, perverse effect like a crisis or a bust or something like this. What is the evidence? I think that's that's a really good question. Let me go first and then Karsten can just jump in. I think when we talk about real estate booms, especially in the current discussion that we're having now in the US and other advanced economies like Sweden, for example, where house prices are really high. It's important to distinguish between what exactly we mean by a real estate boom. So there's some episodes when real estate prices rise very quickly, and that's what's what we're seeing right now is real estate prices have increased quite quickly. And for example, what we've seen in the past 10 years or so is real estate prices have increased, but credit has actually not increased so much. And so, for example, in our data and also work by Greenwood Hansen, Schleifer and Sorensen, you see that it's really the confluence of credit growth and asset price growth, real estate price growth that seems to signal most downside risk for the economy. So when you have house price growth, for example, that doesn't come with as much credit growth, that just doesn't create as much vulnerability for the macroeconomy for the financial system. It's really the interaction of those two factors that helps predict the severity of subsequent crisis. And that's sort of the reason for that I think is what our work is suggesting and what other work has pointed to is that leverage is really the important factor. In terms of the boom that end badly, they're really the ones where not only property prices are rising but where leverage in the housing sector broadly defined is rising both for households and corporates and not just the size of that sector. There's certainly some episodes where house prices rise and even where there's construction and obviously we know that there's big housing deficits in many countries, but the booms that are problematic from a systemic risk perspective are the ones where there's a strong increase in leverage that feedback, feedback with asset prices and creates those vulnerabilities for excessive increases in prices followed by a bus. Good. So the second question is instead of, you know, what are the implication of your work when you are looking for the perspective of, you know, a very large constituencies like the US or Europe, the Euro area in general, where there is a lot of heterogeneity across region in terms of specialization in terms of tradeable and not tradeables. And, you know, for Europe also the fact that we have also the possibility of the exchange rate to adjust due to the different, let's say, characteristics of the different countries or areas and so on. So what are the implications of your study? What is the measure that you can adopt considering that, you know, there is all this heterogeneity both in the US and also in Europe. But this measure, you know, in principle should be the same at the European level. I think that's a really good point. And Karsten, you can jump in, but I think one of the things that we learned from the 2008 financial crisis and we had known before was, of course, the role that different regions play in the economy. In our integrated economies in the Euro area or in the US, we saw that there were large flows of credit that were going to specific regions in the economy. We saw this in 2008. Actually, in the US, we had a similar cycle in the kind of mid-late 1980s where there were big lending booms, housing booms, for example, in the Northeast that turned into worse local recession. So not only do we need, you know, macro data, sectoral data, and firm level data, as Shibnan was saying, we also need to have a perspective on how the regions are diverging. And these regional divergences, of course, as you mentioned, when you don't have the ability to adjust through exchange rates, for example, in order to readjust relative prices, that's going to create distortions that we know take a long time to unwind. And so that's going to create an additional reason for policy to potentially intervene both by mitigating those really big regional booms that we have where capital flows from, you know, for example, the core to the periphery in Europe or from, you know, Central US to the Sand States in the 2000s boom, so both preventing those misallocations from happening. And then when the bust happens, if you don't have, you know, region-specific monetary policy, region-specific exchange rates, then you have to think about other policies to help clean up after that boom has happened. And so that's been obviously a lot of discussion post-2008 financial crisis around fiscal policy, around other instruments that you can use, financial policy of cleaning up these debt booms and the overhang that they leave in terms of the reduction shortfalls and demand and mismatch and demand in different places that you're going to have to address. Yeah, then there is this question by Antonio and, you know, his point is also going on this direction of heterogeneity across countries in terms of the balance between tradable and non-tradable, tradable sectors, you know, and I'm talking for example about Italy, you know, clearly tourism is a typical non-tradable sector and it's not that Italy clearly, you know, it is a fact that this is an important sector. Does this mean that forever this country has to be fragile due to the fact that there is this type of composition? Or, you know, do you think that you should then try to create a balance between tradable and non-tradable even if you have this structural, let's say, predominance of the non-tradable or it is better just to maintain the structure and try, you know, to control more the non-tradable, let's say, sector in terms of credit boom? Yeah, let me take that one. I have two good thoughts on that. So number one is I think you want to be careful in differentiating levels and growth rates, right? So our results are kind of about how quickly do you see a credit expansion, right? You can come from kind of a high level or low level. I mean, we played a little bit around with this. I think it does not actually matter that much. It's really about how rapid of a credit expansion do you see and presumably bad stuff that accompanies that like excessive risk taking, etc. So question or point number two is should you incentivize kind of credit to the tradable sector and so on? So I think there's a really important question here. When you start going really granular with these regulations, you really have to think about them and historically they've really been part of broader discussions about industrial policy. So what you see in many emerging economies is that, you know, these kind of really granular credit regulations like how much credit goes to high-tech manufacturing versus, you know, construction sector. I mean, obviously, there has to be some kind of accountability for why that's being done because you're really shaping, you know, the structure of the economy by extending credit to one sector or another. So, you know, in that sense, I think it's kind of almost like a bit of a divorce point from the financial stability considerations, but I think we need much more empirical evidence to gauge what's going on with this kind of tension. Okay, good. So then there is Javier Suarez, but also Alessio de Vicenzo that are both congratulating and I think that pretty much all the participants are congratulating for the prize. So Javier is having a question pretty much related to causality, you know, what is really the most relevant thing? So what matters for the quality of a credit boom is the composition of the aggregate demand in the background. So it's more the demand that then is generating, you know, the fact that the real estate sector construction has to grow and then, you know, you are, in some sense, trying to boil down what is coming out. But in some sense is the demand that is putting the fire on the, you know, behind the pot to create all this, let's say, construction boom in, or credit boom for the construction sector in that country. So what is going on? What we should care the most, you know, because at the end you're just trying to reduce the supply but is largely driven with the demand. Yeah, I think that's a great question and that goes to the heart of what is it that actually drives episodes of credit booms, credit cycles? Is it mostly driven by credit demand because, you know, consumers want to consume more, they want to consume more housing? Or does it come more from credit supply in the economy from lenders, banks being willing to make more loans, perhaps being more optimistic about borrowers' ability to repay? So here I think two pieces of, two types of evidence are quite useful. So one is just the case study evidence. You see that in many of these episodes of very rapid credit expansion, there's a really big role of credit supply. So one international, for example, factors, large capital inflows, international lenders being willing to lend more to a given country, perhaps because of reforms in that country. You often also see financial liberalizations, liberalize the supply of credit that increases credit significantly. Sometimes it has to do with financial innovations like in the U.S. about securitization and being able to, you know, pool risk in a greater sense to channel more funds, for example, to the housing credit sector. So I think credit supply just in the case study narrative perspective plays a big part of the fluctuations in credit. And then the second piece of evidence on credit supply is other work, for example, by Krishna, Murthy and Muir shows that if you look at what's happening to credit spreads during these episodes of rapid credit expansion, they tend to be quite compressed. So it suggests that it's not just that households want to go out and borrow more. And as they're asking for more credit, you know, credit is getting more expensive. Actually, you see that lenders are willing to lend credit at kind of more compressed cheaper terms. And so I think that drives, you know, a big portion of the rise in credit. Of course, then there's going to be feedbacks with demand. There's going to be feedbacks with asset prices that are going to increase demand, for example, by relaxing collateral constraints. But I think credit supply is a sort of crucial part of the picture. Then once you sort of take this credit supply view as given, then what's happening, I think there's two things. One is some sectors are just more sensitive to credit supply than other sectors because they're initially more constrained. So as a stereotypical example, very, very large manufacturing firms, they're not really that constrained. They can always go out and access credit in the public market, for example, or from banks. But when banks start wanting to lend more in the economy, it's going to benefit, you know, it's going to affect, differentially, these more constrained sectors. So these smaller, non-tradable firms and households, and that's where we're seeing this reallocation. So if it's less than when supply reverts because there often is a mean reverting component of credit supply where financial conditions are tightened, then it's exactly those sectors that are also going to be the cause, the symptom and the cause of the subsequent slowdown and the subsequent losses for banks as Carsten showed. So I think we need more work on understanding the causality of the factors that drive credit fooms. These are obviously very kind of complicated, dynamical interactions between different shocks in the economy. And I think we're making some progress and I think that there's more to be done here. May I perhaps just add one minor thing, which is just to underscore something that Shebnam said in her excellent discussion, which is, you know, I think we've really, and also a fill of I should say. We really need better data on measuring the financing conditions of firms in different industries. So it's nice to have these credit data and we see, okay, there's an expansion here and expansion there. And if we want to say something about kind of the ultimate source of this credit growth, we also need to see prices at least, right, or ideally some estimates of risk premium, right, because then we would be able to say, okay, you know, it's really say in housing related sectors that we've seen expansion, credit supply, lots of lending at very low risk premium, and then, you know, maybe the credit expansions and other non-tradables, perhaps that is more kind of the demand story, right, that now those, you know, restaurants, they all want to open because, you know, people feel like they're richer because their houses are have more value or they, the construction sectors expanding, and perhaps that's really driven by credit demand. But we've tried to look at it and with current data that's available, it's simply not possible to disentangle these kinds of stories. Can I add something here? This is super, super important, what Carson said. So there'll be absolute price data. In fact, if you look at my new paper using the Y14 data from US, US credit is just amazing. You see the price of every single credit for every single firm and the collateral posted down, right, and it doesn't tell, you know, the supply, so it tells actually demand story, right, but it's the firm credit demand and, you know, so you can exactly separate them, and there is a risk premium story for sure, but you can separate that from the price pressure driven by firm credit demand during a low interest rate, by risky firm and by not risky firm. To be able to do that, you definitely need, you know, how much they pay and, you know, what they post, and this is definitely very important. In the US, it is as far as I know, it is not in the European credit register data. It is also in the emerging market credit registry data, by the way. So we should, if we start pushing this rather unaccredited collection in ECB, we should definitely push it this way. I mean, not just collect the amount, but collect the all dimensions of the contract. I mean, that's why it's called credit risk. You need to collect the data on the contract. Yeah. Yeah, indeed. I don't understand why they are not collecting the interest rate paid. This is really something that is not clear to me. And the same is in Bundesbank data. So, you know, it is all over something for some reason that it will be nice to figure out. So, I'm moving now to another question that I think is also very interesting. And it is, you know, we are now facing this, let's say, this incentive, this challenge to try to address issue related to climate change. Clearly, we are creating transition risk. And this transition risk means that a lot of money are going to, let's say, to green firms, and clearly we are going to penalize potentially brown firms. It depends on how this money will be allocated and what type of incentives, you know, public and the private sector will provide. But in any case, this is the way that we are going to observe. So clearly, looking to the result of your paper, is this going to create a boom and pass the game? That's a hard question. Carson, do you want to take that? I have some thoughts I can maybe start with. So, I think that there's, I think our paper doesn't speak directly to that question, which is a very important, I think, you know, one of the big questions for finance going forward is how we're going to finance this green transition. I do think that there are perhaps some historical lessons one can learn from credit policies that were used in different countries around the world with mixed success in the sort of immediate decades after World War II. So, you know, credit policies varied a lot across countries, but in some countries that were explicit, you know, credit ceilings and that's in part how monetary policy was conducted, was conducted by setting specific limits on how much banks could lend to different sectors. And that created both, you know, shape and allocation of credit that in some cases was, you know, at least seem to be associated with successful outcomes. In some cases, Korea is perhaps the most famous example. And now we're seeing some micro data evidence from Korea that suggests that actually some of these credit and industrial policies were successful and credit was a big part of that. But we also know that it led to lots of distortions. There was lots of distortions in a number of ways. One, sometimes, you know, the right firms weren't being lent to. In other ways, they created migration of lending from the traditional banking sector to non banks that weren't as directly affected by these rules. And so I think that, you know, as we think about whether we want to do, you know, whether we want to use credit to shape the allocation of resources in the economy, there is this useful historical perspective. And I think that there's both sort of benefits to doing that are positive lessons, but also negative lessons. And it's important to think about exactly what role credit plays and credit policy could play in dealing with the kind of market failure. The externalities that we think are associated with them with whether credit is is is one of the best instruments to do this. So I don't think we have the answers, but I think I think that there's some perspectives and I think there's there's much more work that needs to be done on this. Very good. So I think that, you know, there is still time, eventually, for one question, because we started a few minutes after, but in any case, otherwise, if there are no other comments, I have one. Diana, I could say something on the climate transition, just a little bit. Just add a perspective to what was said, if you wish, if we still have some time. So is it is very clear that the carbon transition cannot be seen. I mean, so the major obstacle to the carbon transition is a lack of availability of the technologies that are greener. And the financing side can only do so much to to solve this, even if you subsidize, you know, all types of things. You will only achieve so much in the innovation that you would need that would allow actually the the carbon emissions to run down by the relocation that comes from there to to greener industries. So the financing is only a part of the picture and that resembles actually what Carsten said about the question with respect, what do you do with a country that has a large non-tradable sector. There is this industrial policy aspect, somehow, here innovation policy that is actually critical. And that will have a major bearing on the question on the on on the greening. Another concerning aspect, a worrying aspect is that if you look at the current situation in the corona crisis and how the productivity race so far plays out, it seems to be that is the already large digitally very apt companies, according to work by the OECD, some people in the OECD, that actually seem to win even more. So the gap between the high the large highly innovative firms and the smaller local firms in the productivity race, the distance to the efficient to the efficient innovation in this becomes actually larger right now, because they are less able to actually catch up on these new technologies. So there is actually needed on this side of the innovation policies or support of the private sector in doing the green transformation, a major role to work on this side that I think I understood Carsten to be a bit reluctant to reshape the economic structure of a country, and to this type of planning social planning approach here that needs to happen on the innovation productivity side and I do think this applies to this question was just just said without that. And a lot of things that you do on the finding side will will actually, you know, not not be effective, why not as effective it would need to be. Thank you. Thank you very much, Philip. Yes, for pointing this. I think that on top of this is also related to the question I was asking, you know, we have all this digital transformation and that is also changing the characteristics of tradeable versus non tradeable sectors, you know, because if you're thinking about Amazon and, you know, and how we are now selling goods and so on, something that was pretty much non tradeable now maybe is tradeable so we need also to consider this other aspect that maybe will change the way we are thinking and if we are moving more eventually versus tradeable sector this is can be also a good, a good thing. I know if you also think about these aspects. Yeah, so I totally agree with that. So in fact, I mean, you know, this, this definition kind of how do you think about, you know, tradeability versus maybe you know these these other sector characteristics that are just highly correlated with I think that's really where more micro data right so actually go into the firm level where that is really critical because now there are many advantages of what we're doing here in terms of just, you know, getting to a reasonable number of business cycle fluctuations getting to a reasonable number of financial crises. But of course, if you look at the sectoral data you have to make compromises right in terms of how you measure things. And you know, you're exactly raising the right question there which is, you know, okay so if tradeability or these other characteristics change over time right, how should we think about that so that's why we focus on these kind of course definitions, but the hope would be that you know with more micro data and you know other efforts we hope of course that we can provide a bit of a little bit of literature going forward. So I think that now really we are far away from the in terms of time so thank you very much. I want to really, I'm asking everybody to close to both of you for the great work and thank you very much to, you know, Philip and and I want also to thank Shirley and Antonio and all the staff at the European Systemically Board they were really supporting us a lot on the selection of the of the let's say of the winner and also you know in setting up this this day everything went very well and I learned a lot and I saw that also the other, you know, both Javier and even Stephen I'm sure that all of us learned a lot from your paper and all the let's say the member of the Advice and Scientific Committee and the rest of the participants so thank you very much to everybody and you know we are ready for the next prize for the next year. Bye bye.