 No conference these days is complete without looking at, you know, some type of policy response or policy implications of COVID. And so here, Shabnan, together with Tsepiali Vygurin-Shas, Vironika, Ben Chakkova, I hope I pronounced her name correctly. And Nick Sander have looked at fiscal policy in the age of COVID. Does it really get in all of the cracks? So, Shabnan, like, was the case for again that you would have 30 minutes for your presentation. And the discussant of this paper will be Fabiano Schiwardi from the Boerischer University. All right, great. So thank you so much for putting our paper in the program. It is a privilege to be part of this amazing two-day conference on fiscal governors in the European Military Union. And our paper is also very well placed, so it is going to follow very nicely from the paper we just heard on Big G. We will be also looking at different fiscal policies, you know, direct support to firms versus transfers, but we are going to do it for COVID. And this is actually the paper we wrote for the 2021 Jackson Hole paper. Again, joined with Pierro de Gonshaw, Vironika, Ben Chakkova and Nick Sander. And I also have to say because two of my co-authors are working at central banks, these are our views and do not represent the views of the institutions. So it is going to be an applied policy paper, but we are also going to have some very interesting modeling exercise in the paper. And we are going to do two different exercise for that. So to make that clear, I would like to first start with this slide in terms of what kind of a shock COVID is. Of course, we all know now that it is a very atypical macro shock. It is a very uneven shock. As I said, the paper is written for the Jackson Hole conference this year and Jackson Hole conference was about unevenness of COVID and also unevenness of the policy response. And that's really what we were trying to do in the paper to really get at these issues. Now, to start just like beginning yet right now we are two years into this, but just starting 2020. Of course, everything starts with a great lockdown as going by Gita Gopinat at IMF. So there are going to be these several government-boundary lockdowns with amazing cross-sector and cross-country heterogeneity that we are going to take into account. So that's your starting supply shock. Then this sectoral heterogeneity hits you, which really makes it very different than how you think a supply shock in as of 1970s style oil price cost was shocked. So clearly this is not that because you are not going to be subject to this throughout the COVID episode all two years. Every country, every sector, there is going to be this extreme sectoral heterogeneity and of course country heterogeneity. So the original negative supply shock in certain sectors is going to create these negative demand shocks in other sectors. You know, you cross borders demand for travel goes down and this idea is, you know, a model now by in several papers by Verneke Groveri. You want learning, you pick up and others. So this idea of chain aging supply shock. So supply shock creates its own demand shock. Then of course, COVID is a demand shock at the same time against sectoral, right? It doesn't have to be negative for every sector. It's kind of fear factor. You are fearful of going out because of the transmission of the virus. Well, that's not true for every sector, right? It's actually going to be positive for something negative for some sectors. This is the typical restaurant versus grocery deliveries example. Then on top of that, you have this propagation and amplification through the trade and production. And this is going to work, obviously domestically, but globally, right? In fact, one of our arguments is going to be that thinking this in a domestic closed economy model is not the right way to go. Because trade and production networks are global. And this is obviously what we have been living through now over a year. The shortage, not just in labor, but everywhere from microchips affecting car production, restaurant affecting the luxury of food and all that. So this has a global nature to it, not just domestic. And domestic and global is going to be entrapped and amplifying things. Then we come to the later phase where demand normalizes. This is, of course, the big achievement of the policy, especially the fiscal policy. And unfortunately, that is going to create further supply shocks because higher demand in certain sectors are going to create supply concerns in other sectors. And that's the later phase we are living through in terms of both the inflation debate and then this great, what I actually call great supply chain disruption, which started early on, but intensified because of this unbelievable stimulus to demand, which again going to be not even very uneven, not only across sectors, but across sectors. Okay, so this is the setup we are working with. The international dimension is going to be super important because, again, we are trying to understand fiscal policy during this era and obviously inequality and fiscal space is going to be a number one issue here. As I said, COVID is going to be global and production networks are global, which is going to carry things around. But of course, the fiscal transfers, the large ones are going to be made in advance. So that's going to have important implications in terms of cross-border spillovers of fiscal policy, which is going to be very different than the standard models tell us, right? So that's something we are going to be focusing on. And on top of that, of course, fiscal policy is not just the transfer policy that has been extensive, directly critical support to firms and also household but mostly to firms. And that's what we also want to look at in the paper. So, again, the paper is going to try to understand these issues of broad effects of fiscal policy on the current but in a globalized way. So that's going to be very important for us to open economy dimension because a lot of the standard results are going to be different once we start thinking in terms of open economy. And I think this is really the priority right now for COVID. We know that we are not going to get out of this until everybody gets out of this. So this is a global pandemic. But this is not just like the pandemic side. This is not just about, you know, virus mutating in South Africa and coming to U.S. Omicron part. It is going to direct the impact on economies, right? You just cannot recover your economy by yourself without the entire global world recovers. And so that's kind of the approach now I have been taking in several of the COVID papers I did. And here we are going to look through that lens and try to understand the fiscal policy under COVID in a globalized way. Okay. So we are going to do two quantitative models to answer four questions. And of course, in a limited time, I'm not going to go through details of the models. Everything is in the paper. And I assume my discussion is going to comment some of those things. But basically these two models are going to give us eight results. Okay. So there are going to be eight main results in the paper. And that's what I'm going to show you in this 30-minute presentation. Now, why do we do two quantitative models instead of one? Because we would like to answer these questions you see on the left-hand side of the screen. First is how we started this research agenda. We have another paper saying four sort of co-authors looking at the SME failures under COVID. So first, we would like to understand that, right? How does it provide enough because it supports struggling SMEs? So here we are not talking about spending and transfers. Transfer spending as we just heard in the previous paper. So it's very much like this big G concept, right? So there have been, if you want, over 500 programs in 30 countries that we will be looking at just to support SMEs. Just to support SMEs. This is not about supporting airlines. This is not sending checks to household. This is literally to support small, medium enterprise. So we want to understand how successful that is. And the answer to that question is these policies were super poorly targeted. Understandably, these programs put together in a week basically, both in Europe and in the United States and other countries. So they were poorly targeted, but they were largely successful. We define success in terms of reducing the SME failures. They literally reduced SME failures because we are going to show you a counterfactual SME failure from our model in the absence of this immense liquidity support to firms. So in that sense, policy is going to be successful. And also there is this other silver lining because there's going to be a base that inefficiency as they are poorly targeted. The silver lining will be the agent of create zombies. So in spite of the entire media coverage of this panic move last year in terms of all those sorts of zombies coming and we are going to face a wall of zombies. Of course, this is really linked to the 2008 2010 crisis, you know, because that was largely the case then, although the literature still didn't have a consensus on that. My cold, my discussant did a lot of work on that for this crisis for COVID and the policy support for the SMEs, we did not find the result that this policy support created zombie first. I'm going to show you detailed numbers on that. Okay, our second question is this fiscal stimulus health support active. So now this is more in line of the previous paper presented we are going to be looking at transfers here. And we actually going to find a very, very low multiplier, but that doesn't mean fiscal policy is unsuccessful. We will explain why we found that point or six multiplier and there are going to be some similar reasons to the previous paper, but more importantly, the supply constraint and I think it is going to play a very important role for us on top of the nominal rigidity. So we still think fiscal policy is successful here because it is reallocating demand towards sector itself. So low demand sectors are going to be getting higher demand, and that is helping overall employer. So in the traditional sense of fiscal multiplier, we are going to find a low number, but in terms of if you wish an employer multiplier actually, the policy has been large. Then we go to our global dimension and we have how big our fiscal pool is globally because there is this sense that especially early on, a lot of policy commentary was on the fact that there is tremendous spending in US and Europe in advanced economies, and that is good for everyone. That is of course a standard, you know, change in terms of trade US consumers now is going to spend a lot. That's going to be great for them, you know, Mexicans and all the other countries. We actually find that that's not the case because this is a very different shock, and there are all these issues with the supply chain so fiscal policies flows are going to be small. They are all positive on the employer. So you are definitely going to be helping the employment in other countries, but unfortunately, they are actually going to be negative. Beggar dye neighbor on the other countries, other countries be emerging markets here. So we are not going to find a result that advanced school and fiscal tide is it's not going to lift all the boats. And then finally, of course, this is going to put the importance of this question even higher implication of a two speed recovery. I mean the, this is completely about unequal global vaccinations in the world, and we are going to be in this for some time right now, unfortunately, given the much higher vaccination advanced economies versus markets, and that is combined with this unbelievable fiscal stimulus. We are going to show you that there is going to be an effect on unfortunately, global interest rates, and that is going to be a double value on emerging markets because this premier emerging markets is going to increase and that this is going to mean that there's going to be strong So overall, we are going to find that fiscal policy can get in all of the domestic cracks, but it is all about the fiscal space, something advanced economies has but emerging markets countries do not. Okay. Let me start with the first part of the paper. Again, without going to the details of the model. There's a very simple model there that tries to understand the optimal decision making on the firm side. And we are going to model this individual firms behavior, they will be operating in different sectors and the economy is linked to each other to a very rich and full of structure and extensive margin meaning that firms fail in certain sectors, the demand is going to be So the price adjustment is going to help that so wages are going to be rigid, but prices are going to be adjusted. And here we are going to calibrate this by doing a very detailed calibration of the COVID shock. Now, for which shock is going to be several things is going to be a sectoral supply shop sectoral demand shock and I'm just dementia, and we are going to pin this down using real time data at the country time sector level. So the sectoral supply shock is going to go back to the lockdown stringency from Oxford and the owner data in terms of if you can do work remotely or not. Demand shocks again, we are going to Google mobile data and also the face to face interaction at your job in that sector and the active demand we are going to get from realize GDP and forecast GDP growth from the host IMF. So the idea here is basically we have an optimized equation and then decision of firms failing when firms liquidity and operating profits are falling short of to cover the expenses. And we map that we estimate that equation using firm level financial data at the entry of COVID. So we are going to use firm level financial statement from 2018 for 18 advanced economies and nine emerging markets. And basically we will be shocking these firms balance sheets with these COVID shocks. You calibrated the real time data and then first show you a counterfactual firm failure rate without the policy. Of course, this is not the real world because there was tremendous policy support, as I told you, and this part of the paper we are going to only focus on support the firms directly. And as I told you, they're just over 500. So we are going to focus on tree, because that is, these are the not only the common policies across these 18 advanced economies online emerging markets. We also have data on them for ECB, ESRB and OECD. And the policies we are going to be looking at tax waivers cash grants and pandemic. Okay, so they are going to be calibrating fiscal support to this and then, then calculate, okay, that counterfactual SME failure rate in the absence of policy. How much that goes down with the support, tax waivers cash grants and pandemic. Before showing you that result, let me show you the shocks, because this is going to be very, very important. Of course, everything is about how you measure these COVID shocks rich heterogeneity heterogeneity sector dimension. So here, the color coding in this slide is going to tell the sectors that are essential, they are going to be orange, and the dark blue sectors are going to be non essential. Now the left figure is going to show the sector specific supply shock using the data I mentioned and you see from the y axis, it is going to be very intense. So supply shocks are going to be more intense here than the demand shocks. That's what the data tells us. And the demand shots are going to be on the right where it is not always negative as you see it is, it is, it is relative shock right it's going to be negative in certain sectors large negative like entertainments and recreation but it's also going to be positive in other sectors such as electricity transportation and storage and construction. Okay, so there's going to be this positive negative demand shock sector and the sector of supply shock. Now, in terms of intensity, I told you supply shock is always going to be, oops, sorry, more intense like you can see from the scale going up to almost this indicator of 60 and there's going to be amazing country heterogeneity. So this depends on your lockdowns and demand shock is going to be also going to depend a lot how, you know, consumers behave and all that but basically the intensity is going to be less you can see why. Emerging markets, everything is just worse, right? I know this is a supply shock is going to be worse, emerging markets, they did much more lockdowns, we do all that, but there's also going to be higher intensity for the dimension. Now, this heterogeneity is going to be very important. So we are going to be working with sector country heterogeneity to calibrate the COVID shock and firm heterogeneity of course in terms of your initial balance sheet, initial financial position as a firm in all these countries. We have firm balance sheet data for all these countries. How, what type of a financial position as a firm you were in when you face these sector COVID shocks depending on which sectors you are operating and of course some firms might be operating in different sectors. Our first result is as I told you before, so fiscal support is going to reduce SME failures more than that. So where do we see that? So in the first column here you see the failure rate is nine percentage point increase. So we are going to show Delta's always. So Delta mean how much the failure rate increase in COVID relative to non-COVID. So there's a nine percentage point increase in the failure rate in all countries in COVID and you see that much higher number in emerging markets, almost 13 percentage point increase and advanced economies around 5.65 percentage point increase. But the fiscal support, you see what happens with numbers with the fiscal support. So first of all, overall increase is healthy, right? So this is our first result. Fiscal support reduces SME failures more than how your nine percentage point increase in the SME failure in the counter factor scenario of no policy help came. That becomes 4.3 with the policy support. And remember the policy support we are looking at. We are just a little cheater, pandemic loan tax waiver and cash grant. Now there is a full offset in advanced economies. Advanced economies is so much that actually you are doing better on the COVID relative to normal time business cycle exit. And in emerging markets you still reduce 12 to nine, but obviously it is never, it is nothing like a full offset in the advanced economies and that's solely due to the size of the fiscal program in advanced economies. Now, as I told you, these are going to be poorly targeted. The support in terms of percent of GDP is going to be around 4%. And most, as I said, is spent by advanced economies, 6% of GDP, where emerging markets are spending only 1.9% of GDP. Again, these are these three programs we are looking at. And they are very poorly targeted. We show that 88% of the funds dispersed, these are the dispersed funds, right? So you can get some back. In that sense, we are now calling them the fiscal cost, but dispersed funds, 88% of them went to viral funds. So it went to firms who didn't need it actually. And so that's the poor targeting. And the third result, the second result is poor targeting. Third result of no-zombification comes from the fact that our estimated 2020 on failure rate is only 2.6% above normal, meaning, oh, is it the case that we saved all these firms in 2020 and there are all zombies that they are going to fail in 2021? The answer is no. The firms you are saving are actually mostly mild after the policies, okay? So the answer to the question, the fiscal support provided, we could have struggling firms, yes. It is poorly targeting, but it's definitely reduced SME failures fully offsetting them in advanced economies without creating zones. Now, there is, again, please do not forget the extensive country heterogeneity. I don't have time to go into detail in the presentation, but basically you can see a simple picture of this year. On the left, the baseline failure rates country by country in dark blue and what the policy does in orange in advanced countries. You see several countries turn negative, a big spender like Germany here. You see that it is more than full offset. And of course, Germany is an interesting case because there is also a regulation and ban on filing a bankrupt. So it is actually extremely difficult to understand what is going on also in real time as we are not going to get real time data without before two years. And then countries like Germany, Benning, the filing of bankruptcies, you know, it is going to be actually quite a heroic exercise to do this. But basically, you know, there's going to be a counter heterogeneity, but still for advanced countries, you see a huge decrease in emerging markets. Policy is reducing that there are going to be countries that are also going to negative like Poland, but also there are going to be countries like Romania and Bulgaria so little spent, you know, policy is not being that effective. So again, it is about fiscal space. Now, from this model, we are going to go to a global model. Why are we doing this? So this firm failure model that gives you the firm failure without the policy and then shows you what policy did. Then we want to understand the added effects of fiscal policy in a globalized world. Remember, in an open economy setting, here we really want to look at fiscal trends. So the firm failure model cannot do that because there's going to be, first of all, everything that is going to be exorbitant. So we cannot look at transfers to households. The IO linkages are not linked in an open economy sense. Everybody's IO linkages to themselves, domestic, even rework with 27 countries in the firm model. So we cannot talk about international spoilers. And it's also a static model. There is not going to be any savings channels. Now, the global model, we are going to solve and we'll spend it in a good object. There is going to be set temporal issues with constant household. So we have this full-fledged heterogeneity in terms of the fiscal transfers and multipliers, literature, there are going to be households constrained. There are going to be households not constrained and the NPCs are going to be important. So that's going to give us a very nice framework to understand the effectiveness of fiscal transfers. And one of the most important, as I have been saying, this is a global model. So IO linkages is going to be very important in terms of these cross-border fiscal spoilers. And we are going to use data on that from OECD on global production and trade network for 64 countries and 36 centers. So these are going to be improved. The second model, the paper on the first model, the cost is, of course, we cannot use firm data. Let me first show you this picture to make the point. Again, the model here, I don't have time to go to details, but it's a global dynamic heterogeneous model with normal rigidity and we are going to calibrate it using data on global trade and production network. Now, what do we mean when we say global trade and production network? Let me clarify that because the first thing people always think about is the trade network. It is not simply a trade network because there is the sector of the measure. So the left here, and by the way, these figures are from my paper titled the economic case for global vaccinations that we have written in January 2021, and make the case that the pandemic is not going to be over until it is over everywhere and no economic recovers until everyone recovers. Because in this paper, basically it is an epidemiological macro model, a global model, and there in that paper we simply made the economic case for global vaccinations, not the ethical case because not vaccinating Brazilians, Mexicans, turkeys, and Africans of the world is going to cost a lot to European Union and United States. And in fact, yesterday at the opening panel, Helen Ray quoted these numbers from this paper showing the cost of not vaccinating the rest of the world is 20% for France and how this is going to into a fiscal governance in France. So this was in Helen Ray's presentation. So basically this cost is going to work through this trade and production linkages, but it's not just simply this trade, but it is this non-linear, this domestic IO linkages that intersect with that trade and that insight is very famously from the recent work of David Bakai and late Emmanuel Fahri who we lost last year. So this vaccination paper fully takes their framework and basically carries it to an empirical setting to make the economic case for global vaccination. So here in our paper, we purely via Bernika and Nick, we use exact same data and these figures actually from my vaccination paper, but I'm showing you here to make this point. So on the left, you see the trade network and you see this color coded because, and by the way, this is a step down version where I am showing the largest trade like over 15% because if I show me everything, it's like a spider web and you're not going to see anything. Right. But the point I want to make here is that there are large countries, small countries, and the color code is they are very open countries like Ireland here, very dark blue, and they are not that open countries like China and the US, they are larger, but they are less open in terms of trade at like, okay. Now, what you should envision is you incorporate this domestic IO linkage, which we mostly think more like a structural relationship, right. We want to produce a car, he needs four tires and all, but some of these sectors are tradable, some of these sectors are not. This is going to be fully incorporated into this. That's what a global trade on production network and that's exactly why we are leading to the supply chain issue, great supply chain destruction. Then we have these sectors with different demand and supply shops demand normalizing at this point. This is going to be very important here for our paper when we understand the effect of these proposals. If you look at the figure on the right, again, there are going to be these very tradable sectors like computers, electrics, you know, manufacturing sectors, but there are all these sectors, construction, wholesale and retail, you know, they are not agricultural and fishing, they are not that tricky, but they are all linked to the other sectors. And these links are the things that are going to be, and non-linearities are going to give this amplification. Again, this insight is coming from Bakke and Pari's work. This is going to be very important and that's exactly what it means when we talk about the lumber stuff and the other stuff in the, in the literature, and then you just go to the Starbucks and you start seeing signs of like, okay, we are not being able to get our inputs and labor. So the service is going to be so, so it is not just, you know, what we trade, but it is all these linkages together with the domestic, so we are going to use this data and how we are going to make use of it to understand the fiscal policy in a global context. Basically, we say, okay, now let's focus on transfers. Again, this is going to be very related to the previous paper. We are going to look at transfers. And again, of course, fiscal space is going to be very important. If you look at this video, these are the countries we work with. And here, obviously, advanced economies, United States at the top, on average, the transfer spending is going to be almost 16% of GDP in advanced economies. You look at the emerging markets, high is Brazil, but it is, of course, very, very little compared to what advanced economy spends. I mean, you can see that, like, you know, Mexico almost did nothing. It's 4.9%. Okay, so in a world like that, of course, is very important. I mean, a country like Mexico wants to, you know, understand, well, they didn't spend much, but thank God they are trading partner US spend so much, so is it going to be good for Mexico? This is an understanding model, yes, it's going to be good for Mexico. And we are going to find it's going to be good for Mexico, but it's going to be bad for other countries, exactly because of these different linkages and the rising global rates and the amplification and then the differences. Okay. Now, let me go through that results. First, our first result, result number four says that there are going to be sizable demand constraints. Okay, so we are going to separate sectors in terms of supply and demand constraints. Of course, 70% of sectors are going to be supply constraints. 30% are going to be demand constraint. But that means like 30% of global GDP, global GDP is like our 27 countries is going to be in demand constraints. Okay, there is going to be lower. And that goes back to the figures I showed you, just the sectoral shock, the COVID shock in terms of intense supply shock is going to be more intense. Okay. Now, but what does it mean? I mean, you start with this, you have 31% of the global GDP in demand constraint sectors. What does it mean in terms of what can fiscal transfers do? Of course, the low multiplier is going to be very low, right? So we are going to have a 0.67% stimulus for 11.3% fiscal impulse. 11.3 is the average of those numbers I just showed you. 15 and 4 in advanced economies and emerging markets. The multiplier is 0.06, right? This is going to be very low multiplier. But we think just looking at numbers misleading, because in fact in the paper we show we started a standard textbook multiplier and we bring it down to 0.06. Why? Because of supply. But for mix and two main policies, the transfers are not designed to stimulate output as we have discussed in the previous paper. However, the policy is successful in terms of employment support. Why? Exactly because of this heterogeneity of animals across sectors. If you have these sectors where there is slack in demand, fiscal policy can reallocate the demand. And in fact, that's what happens. Fiscal policy is going to reallocate the demand to those demand constraint sectors. That is going to help the employment. And this Canadian employment, right? Because it was there at the first place because of these Canadian supply shocks. So that is going to decrease a lot. So that's going to be from 2.67% to 1.4%. So even though the change in real GDP is 0.6, the change in the Canadian unemployment is minus 1.3. So that is another success for fiscal policy. Okay, despite the multiplier, fiscal policy is reallocating spending towards the sector with slack. So that is again something very important now. What about schoolovers? In spite of this domestic success, the schoolovers here in Tunisia, schoolovers are going to be small and mostly negative. And that's going to come from interest in terms of trade effects. Now, employments are going to be small, but mostly positive because demand, extra demand created in the sectors with slack. Here's a figure showing this and I'm only going to, we have several things in the paper. Of course, doing it advanced economy stimulus, that kind of stimulus. Let me show you the only US, the biggest spender, US stimulus. What is the spillover of that to other countries? Of course, the effect is largest on the US itself. This is output. And you see Canada and Mexico, the most important trading partners, they also have caused the effects. I mean, small, but positive. At least in the right direction. Everybody else exactly in the wrong direction. Okay. This is going to be a very important result out of our paper, which obviously you are not going to get with the standard shock and standard amount. So the shock is also very important here, not just the global nature of the market. Employment spoilers. Again, you know, we are doing good here. They are going to be small, but still positive. So, because the share of the demand cost to sector is declining around 3% points. Unemployment is declining around almost a one person point more, more in US, and, you know, but also in emerging right so it is so in terms of unemployment. This US fiscal policy helping reducing unemployment in other countries. So again, the fiscal policy flows are not that big, small, but in terms of employment, they're all positive, good news in terms of unfortunately, they are not going to be positive. So there isn't this sense of advance going spending, lifting all the votes up, so which again makes importance of global vaccinations. Okay, my final thing is going to be on to speed recovery. Now, this is again, this is this is I think the only important issue right now is the stupid inequality in global vaccinations, which I think none of us understands why we are still in this world. This is a big turn to investing in this global vaccinating the world is huge. You know, January paper we calculated 166 times return to advance country investment in vaccinating the rest of the world. This is still not done. IMF actually put out a proposal in May, you know, calculating the total amount, which is going to be 50 billion. So it is it is a very small cost to vaccinate rest of the world, but it is not done and behind the two speed recovery. So combined with fiscal stimulus, the huge fiscal stimulus in advance economy is going to have very damaging effect on more. First of all, obviously, the decrease in the private saving is going to increase global interest rates to advanced economy channel that's going to increase the global imbalances. So in terms of the trade balance, of course, this is a standard model. We are going to get a charity trade balance and advanced economies improving the market. But what is important here is, yes, I'll put this increase in advance, but it is declining. Okay, so there is going to be a almost one percentage decline in output in real GDP in emerging markets relative to this is real. Okay. And the increase in the global interest rate is going to be 2.6%. Now, they're going to double them into emerging markets. Why? Because this high global rates is going to translate the high risk for emerging markets. And on top of that, if you have multiple response with a title, because of obviously this inflation issue, driven by the supply chain disruption, that's going to be even worse for emerging markets because of this high risk. And we show it here in this figure. This is, by the way, updating this result from my 2019 Jackson Hole paper there. I there showed first time this estimated effect of a contractionary US monetary policy on emerging market spreads here on the left. And the contractionary monetary policy increases the spread in emerging markets, so it spreads and it decreases the net best comments. Okay, so here we are updating that result. And with data pretty much to the end of 2020 and, you know, with the surprise contractionary monetary policy, there is going to be a higher risk for emerging markets. So even higher external borrowing costs because higher global rates times the clusters, you are going to have huge costs for external financing for emerging markets. So what are the implications of the two-speed recovery in this world of unequal global vaccinations increased from our model, increasing global rates, widening global imbalances, financial habits for emerging markets. Okay, let me conclude. So our paper shows that fiscal policy did get in all of domestic cracks. This is very important. So this is now a title inspired by Jeremy Stein's famous code of monetary policy getting in all cracks on top. So fiscal policy is getting in all the cracks of domestic. So your domestic fiscal space matters above all. In terms of firms, SMEs, fundamentalists are strong and they don't need additional support. This is, by the way, not futuristic factors down the road statement, we are talking about 2020 and 2021. The multiply from fiscal transfers is very small. That's not to be surprised about because we look at transfers, supply bottlenecks and IO linkages. You start with the taxable multiply, you bring it down to 0.06, then you incorporate all these. Now cross-border, cross-fiscal policy are also small and actually are better than neighbor for most countries. This is going to depend on the standard of complementaries in trade and production network. To solve our model in this paper, we are actually assuming a cup bubble. So we are assuming not much complementarity. So with that, there's going to be much bigger amplification as we know from Baku Fahni's work and also as I showed in my global vaccination paper. Now, there's also this problem, unfortunately, in terms of this fiscal policy schoolover is that unequal global vaccination, of course, works against the policies, right? Again, going back to the course that landed yesterday, overall, in that vaccination paper, we chaired 49% of the costs are going to be borne by advanced countries, even if they are fully vaccinated via these supply chain issues, and especially if complementaries in the global production network is very strong. And I think one lesson you learn from this COVID shot is that at least in the short term, within a year, time frame, these complementaries are actually very strong. So a two-speed recovery is going to create headwinds for emerging markets due to rising global rates, which is also going to be combined with rising risk premia if US, you know, not if then actually US monetary policy and other advanced economies, monetary policy and online. Thank you very much. Thanks a lot, Shannon, for an extremely rich paper and so many results that are very useful in the policy discussions that are ongoing on the implications of COVID and impressive what you managed to cover in the time that we allotted to you. I may pass on directly to Fabiano for the discussion. I know he has also done some work at least, I remember still from the initial part of the pandemic about the liquidity support that firms may need, and he did some very interesting calculations, so I'm sure he can bring an interesting view on your paper, Shannon. Thank you for having me. It's of course, you know, and I have to discuss, and get Olivier and co-authors paper, you know, that contribution to our understanding the pandemic cannot be underrated and the fact that they got to present it. And Jackson Hall shows that the cooperation between the academia and institutions work well. So, so what does the paper do? As Isabel was saying, it's an extremely rich paper so you need to get to find your way through it. It basically studies the effects of fiscal policy during the pandemic using a very rich micro macro model, the first one and then macro model for the open part. And it's basically a more paper where you run counter functions. You set up a model which is rich enough to allow you to, you know, to see what happened, but also to see what would have happened, for example, if the policy support would have not been there. And, and then given that they have all this richness in terms of from level and different channels through which the pandemic hit the economy, they can also say who benefited from the interventions, what type of bottlenecks were more important to explain the lack of effects, for example, from these interventions. So they start with this close economy model and just say the ingredients just to give you an idea of how rich the model is, but it's from heterogeneity that are distinct demand and supply shops, and they differ in the country and at the sector level, and then there are these IO linkages. Okay. So, you know, modeling effort, a lot of work to collect the data. So it's really, right. And so, you know, when I, when I was reading the paper and I got to this point I felt like when I go and visit my mother, after a while I've seen I haven't seen her, and she prepares this fantastic dinner, a lot of great food. And at some point I say, oh, I'm so full and everything was so good and she tells me, what do you mean, I mean, we still have the second course, the dessert and the, and the, and the grapple because you know, then, once you get to page 30 that is actually this multi country model on which I will not really comment. My international finances at the graduate of steel and level, I know something about firms I don't know about. And, you know, maybe would belong to another. What are the results. So the policy work that we were seeing the failure, failure rate declined from 9% without policy to 44.3% with with policy and in advanced economies, it actually went below what would have been in a normal. And then, you know, given that the model is rich, they can actually say that the IO linkages were very important, you know, I actually got in a paper with a because I think we got a similar results when we estimate the effects of zombie landing during the great recession meaning that zombie landing somehow doesn't seem to be that bad for the aggregate economy, because it prevents the disruption of supply chains and therefore, during the session, keeping zombies alive is not as bad as we expected before doing all the The support was purely targeted 90% basically went to sort of the wrong firms meaning firms didn't need it. And then, but the support was not too bad in terms of quality of firms meaning that not many zombie firms but actually were actually finance and I think this is actually in line with the emerging evidence that's coming out, for example, of using an accurate data. And it also means that the provisions that the government is put like for example, not having bad loans to receive a guaranteed credit actually work. And then, and then the model predicts a modest surge of failures in 2021. And I think this is also in line with what we're seeing. Actually, I think in Italy, we see that even in 2021 failures are below but what they were in 2019. And maybe there's even too much of a good thing. There is a persistent effect of this policy. And then, you know, there is this G model and as I said, then, explain but is there are these negative spoilage in the spill overs that you know I don't, I mean I can't really tell where they come from because my recollection is, is the fiscal expansion with very accommodative monetary policy should benefit other countries but I think this is one of the many specificities of these crisis compared to a start. But anyway, I don't have much. I want to have anything to say. I have three comments. So the paper goes a long way in describing the data used to feed in the different counterfactual exercises and also to feed in actual policies, etc, etc. It's a little bit more. It's a little bit quicker in terms of discussing parameter estimation of calibration well calibration I would say, and I think it would be useful to have a section dedicated to that when you have, you know, when you do model calibration. And also, the same, on the same page, it would also be useful to discuss a little bit the model feed before diving into, into the counterfactuals because this is the typical, you know, partner in those papers you write a model, calibrated, you show what happens when you send the predictions with respect to things that you can actually measure you measure in the data to show is, you know, the predictive capacity of the model along dimensions that you have to calibrate, for example. And that would help because you know there's so much in a moment like this that it's really hard to assess, you know, many different assumptions, etc, etc. And at the end of the day what's posing what to so to have a little bit of a model feed would be useful. And indeed some of the assumptions are strong as it cannot be differently. For example, but our firms in essential sectors, and they face no demand shift we need that if you are classified as essential, the demand doesn't change doesn't there's no, there's no dropping demand. You know, when you look at the data, for example, there is this transport and storage sector that has a positive demand. It doesn't show up and I don't know it doesn't confide. It doesn't actually agree with my expectations, we I think in transport section is I was a big loser in this recession as far as I understand. And then there is a very big negative shock education and that is not super clear to me where it comes. The same thing this will say in retail is a is a is also a big loser without the policy bankruptcies would actually skyrocketed skyrocketed and I understand there are many shops that actually suffered but there are also grocery shops there. Probably, you know, they are essential. And while food and accommodation actually even has a negative effect on failures absent policies, which is also surprising. I'm just saying that you know, when you have so much things, something is going to come up. I mean, maybe not in line with what we expected, but to have a little bit more, you know, discussion of calibration and model would have to assess the capacity of the model. The question is about the why what was the support so poorly targeted and I think because already in the chat asked, what could we do to do that and here I'm going to talk a little bit about my work and I felt a little bit bad about it but I think it was important to discuss because question says that maybe it was actually a good idea. So I think there is no question that a lot of resources like to refer to them and set them encoders also acknowledge it. It was no time to be picky about what to do. You know, it was an impressive crisis. The government went for let's do something rather than let's let's understand what we should. The assessment is actually is actually very large 90% of support went to firms that didn't need. Okay. And I think this is actually a very important message and given that this paper is about policy. I think it should be important to qualify more because it might actually give rise to reactions, even by the fact that is not fully understood. So, and the paper says that in advanced economies, the GDP, the, you know, the, the interventions that were actually needed to really save SMEs at the risk of bankruptcy was was really like point 13% of. On this number, I have two points. One is that you know when I computed my, my numbers for liquidity shortages and liquidity shortages by the way is exactly the way they define failure is not they don't look at bankruptcies as a legal event. They say a firm goes back to the exit the market exits the market when it's liquidity becomes negative. And this is exactly what we do. And when I did that I actually got that around. Here it is if you look at this draft, let's only look at the last bars. This is in in in. And billions, billions, billions, billions, billions. By December, around 70 billions, you know, we're needed to save all firms given that they focus on SMEs, but number four SMEs in Italy was around 50 billion. And around, you know, whatever to when something per se. So, I don't think point 16 point 15 seems a little bit. The next point is that again on again on this is that the model only focuses on the extensive margin, which of course, you know, it's a, it's a, it's okay, it's a choice. But at the same time, I think these policies also work that they intensive margin meaning that we know, for example, from a lot of work on the great recession that that liquidity constraint firms, firms that were not at risk of going bankrupt, actually maybe to cut their operations because they, they couldn't finance the working capital. So to avoid bankruptcies, they would just hire workers and reduce production. And I think, you know, the interventions also work that dimension. Okay. So, I think that's, it's a dimension as important as failures in the sense right because breaking workers for matches that actually were are still valuable today is like a micro example of failure because we know that it takes time to make cheese, etc. And so this is, I'm not saying that this is something that should be should be taken into account in the moment, but in the discussion of, you know, the fact that most of the, most of the aid went to firms that didn't need it, I think it will be important. So, certainly didn't mention it, but in the, what's next, they say that they just talk about tapering, right, and they talk about tapering in advanced economies vis-à-vis emerging economies and so I think here, I don't think the job is totally done for policy, right? And indeed, when you look at the data now, the data, I just looked at the financial accounts of Italian firms for 2020, there are many firms that became fragile from a, from a financial point of Okay, and we know, and, you know, of course, they're aware of it, and they mentioned this point in the conclusions, but only there. And so this is not really a criticism of the paper, it's more a contribution to a workshop on what should policy do now to get out of the, of the pandemic, right? So it's another paper, maybe, and it's something that I'm working on. So we know that from corporate finals that overhang and shifting can actually prevent entrepreneurs from, you know, fully exploiting growth opportunities, and that could actually be a drag to the speed of And so the targeting, targeting point that Nico was asking, and what should we do? Well, you know, there was a document by the G3, by Mario Draghi and Ragu Rajan, and they were saying already, like, maybe a year ago, you know, now that the really a good phase of the pandemic is gone, we might actually become a little bit peculiar to be more selective in our So let me give you just a very quick summary of what I think we could do. This is what I'm doing with Andrew Lou of Indiana. So the idea would be that now we can distinguish between firms that are economically viable, firms that were good before the pandemic, and we know that quality is persistent, and who are in sectors that have good growth prospects after the pandemic, and we measure them from the stock. Some of these firms actually had deteriorated accounts just because they were in sector heavily hit during the pandemic. So they got support, but a lot of the support was through, through loan guarantees. And so now they're actually, you know, economically viable but financially. And so we could think about interventionals targeted to this firm. We do some preliminary calculations we find that around 27 SMEs in Italy are in those conditions, we do economic prospects, but actually, you know, with a lot of debt in their balance sheets. And we also do some preliminary counterfactuals and it turns out that it looks like doing some equity injections in this firm could be a good idea, also in terms of using that, you know, that becoming bad. And I think here in issue will be, what will banks do once, you know, all these guarantees come to an end? Will they actually, on firms that don't look that great, will they actually decide to exercise the guarantee with the government and let these firms fail or not? And I think this is a point that will become important at some point, and we might as well start thinking about it. So, again, paper, thanks for giving me the chance to discuss it. That's it. Thanks Fabiano. I'm just checking with seven and the other react to meet you to the points of Fabiano or shall I have two questions in the chat as well. So, okay, let me quickly react. So, I mean, great comments and I actually like there's nothing that I disagree. Let me just clarify. 1.2 points actually. One is on this counterfactual cost, right? So, our 0.13 number and Fabio's 2% GDP number for Italy. I fully agree with that. And I think this is very important because that counterfactual is basically, you know, what the model gives you when you go and pick these viable firms with at least, right, which you cannot do in real life. But obviously you can do within the context of model, but still you are using the firm level balance sheet data. So that number is an average number. It is going to be higher in countries like Italy than you have a weaker position, which also explains the importance of this recent work Fabio doing and others. So, how we should be dealing with doing that work, right? I mean, look at the real time data. The site is out. I mean, you know, how weak they were entering the call rate, which was related to you. But now what that change in this current financial position and see what can we do in terms of their financial needs and all that. I fully agree. Fabio is doing great work on that. Ben, I created a paper looking at that French data. Let me be very clear. This type of work needs very detailed administrative data, right, as the Italian credit register or European credit or in the in the in the French word detailed also confidential data for France. Then our paper is remember the point is we wrote this paper, we wrote the first draft starting in, you know, for paper in May 2020. So the idea is what can you do in the middle of the crisis when you don't have the data and then and even there's confidential data that cannot be collected like that. And it will be even if it's there, some of it will be there only available to some regulators, right? So the idea here is like, okay, what can we do? Then we don't have any of that with a simple model. And now actually we show that this works pretty well. So that's that's the idea, right? But just in terms of understanding, you know, who is why we're bringing more confidential data on that and then try to understand maybe the kind of smooth crisis by making an economic viable, but now they have all this leverage. So we should do more. I fully agree with all those points. And in fact, now I'm working with US credit registry. And this is even more important for us because with Italy and the ECB on a credit, we always have access to these data. Unfortunately in US we knew nothing about the small firms financing because US what they don't find. But now, thank God, there is this regulatory data at the Federal Reserve start being collected after actual legal collapse as part of the Frank Act. And we have written a paper using this data and showing that actually small firms in US are no different than the European small firms. They are fully bank dependent. They are going to have all these leverage problems, even more so when I'm moving forward. So this is definitely number one policy issue. I fully agree with you. So, great. Thanks. Let me immediately jump to the two questions that were in the chat. And meanwhile, I also solicit anybody who wants to still ask questions to raise their hands when they're a panelist or otherwise put it in the chat. So the first question I think actually Fabiano gave a party already suggested answer on it is, you know, how could a more targeted policy support for SMEs look like in real time. So this was a question from Nicole Zorel. Yeah. And the second question was really on the data because, and I think it's written by what we know about your area, where of course a large part of the support was in the form of government guarantees basically right and so of course, only a small part of them were called upon. Many didn't materialize in reality but at least our analysis would say that they still had a positive impact on private sector expectations and I mean you also see especially actually in combination with the monetary policy measures we see that they have been quite powerful and supporting the macroeconomy. So the question is did you somehow or how would you somehow include these guarantees in your analysis? Did you do that or how could you do that? And if not, what would be the implications of that kind of support if you consider if you would reflect them as well in the analysis. So these are the two questions that I see here in the chat. Yeah, no, on the target is exactly. So this is what we should do more work, right? I mean, let me be very clear. None of the policies were targeted. I mean like PPP to put together in a week. I mean the German and French programs also like you know, this is less than two weeks. So nothing is started. No policymaker went to the battlefield and let me take a surgery and find out. So this is not that. But again, completely fully agree with Fabiana that we should do this now moving forward. This makes this data issue like number one priority, okay? So we are doing a model-based exercise because we cannot do targeting using data. Why can't we not put? I wrote the policy proposal being as early as April 22, 2020, literally one month in the pandemic title, negative SME tax. I argue that, okay, we should not be dealing with banks. So the IRS and I wrote it for you as immediately should be a negative SME tax now and call back later. That's like, this doesn't happen. I mean, we had all these policies done through banks, through a combo of ways. When they were trying to put the PPP, they were trying to access from financial books. They cannot because in the US, unfortunately, the census data and this regulatory credit registry that's completely separate. And they cannot match them because there are different regulatory jurisdictions. So this actually makes the importance of data sets like on a credit. Data sets like real time, we can see what is going on with from finances during the shock. So we can go talking about this is, this is super important. In fact, now, you know, this work is being done. This is the paper, this is the red paper in San Francisco. I know they are matching the PPP data to the regulatory credit register of the US. They are trying to understand how targeted it is, you know, which one's got it on all this. All of this needs real time data to be done. So our, our contribution is like, we have none of it. What can we do as a policymaker in the middle of the crisis? So how much we can. So that's, that's the goal without saying. And then the second question is the pandemic loans are the ones with guarantees, right? We are, we are using those. We use data from the SRB on that. Of course, again, we are going to fall short of really which firm got what, right? We take up rates and disbursement we use from the SRB data, but then we don't know firm X got this much firm Y got. This is exactly this new work now being done, right? So that's exactly why we are very careful saying this is the disbursement, right? The actual cost to the government that has to be evaluated once this is all out. But in our paper, we did it, of course, an exercise saying, okay, say the firms all, you know, paid it or they, they are forced to pay. It's the five year loan with the government guarantee. They are forced to pay right away. What happens? And even under those scenarios, we didn't find the meltdown. The meltdown comes if something goes wrong on the monetary policy side and we are back to a banking crisis issue where banks refused to roll over the existing debt and things like that. If you go back to the scenario like 2008-2009, then we are going to have problems, okay? So, and to calculate all these things, we do need to know what is going on in real time. But we did say, okay, if these government guarantees everything is paid back, what happens? What is the cost of government? And the cost to government, of course, is going to still be very high because the programs are not designed as I'm going to slow back this excess profit tax. You know, so that's exactly what was my proposed about, that they are not designed like that. I mean, most of them design, you know, you have to keep employment at PPP, then it turns into a grant, or you pay almost like zero interest in over five years. And that's done because governments made banks do it like that, right? And in fact, now, in the US, people are finding that before the government involvement, no bank is doing any lending, which is very interesting because it's not a banking crisis, right? It's understandable that we are in 2008-2009, when banks got the shock themselves, they are just like cutting the credits up and all. Now, I mean, nothing happened to them. The monetary policy is like record, you know, accommodative, and they still don't do it. Why? Because this goes back to this issue that I originally showed in my QG paper, the financial policies are size-dependent. Small firms are financial policies, okay? If governments weren't involved, there's no way banks are going to go and tell those restaurants, oh, I'm so sorry, your earnings crash. Let me give you some money so you can snoot up this liquid shop. That's why you look at the liquid, okay? They are not going to do that at all, even themselves, not on facing a shock. That's what I think government programs achieve very successfully. Yes, not targeted, inefficient money-based, but they achieve if the idea is to prevent a failure and unemployment from skyrocketing, they did achieve it. Thanks a lot. I can actually compliment. We have a survey for small companies in the ECB that we run. I don't know if you know it's a safe survey. Yes. Pointed towards it because it gives a bit of indication as well. Exactly. Yeah, we cite your survey. You say the same thing. Exactly. Yeah, but I mean it's like in our paper, we kind of model this and we show how these shops can amplify in a road like that instead of everybody's facing the same financial frictions versus small firms facing a much, much rigid one. It is very clear in your survey. Yeah. Very good. No further questions came in. I think people are still adjusting all the results and the riches of the paper. And actually we're also over time for the conference overall. So maybe I just suggest that we close it here and I'm sure people may come back to you with questions. Please email me. Any questions you couldn't ask, you can definitely email me. Thank you so much. And then maybe I would still like to spend everybody's time for one minute to thank all the organizers of the conference. So this was Jakob Pochimadomo, Demo's Joannou, Bartos Makovjak, Leo Vontan and Niko Zorel. And I would also very much want to thank Anna Maria Borlescu for making sure that everything runs very smoothly operationally as well. So thanks to them and thanks to everybody who joined into the conference and in particular to the presenters and discusses for the great papers and the great discussions. And I wish everybody a very nice weekend as well. Thanks. Bye.