 Welcome everybody to the next COVID-19 ECB webinar. Today we are very happy to have Veronica Garieri from Chicago Booth with us. And actually we have a bit of an unusual setup because she's flanked by her three co-authors. But I'm sure we'll take you to the top questions at the end. So we have Guido Ador and Sony, Ludwig Straub and Ivan Werning. And Veronica will take us through her recent research with them on how supply shocks can also affect the demand because of this pandemic and give us sort of also a general flavor on the topic. And Veronica doesn't need any introduction. She's a leading macroeconomist and we're all following her research. So without further ado, I pass on to Veronica. Hi. So thanks a lot for inviting me to give this talk. So as I look mentioned, this is a work on the recent pandemic and it's joined with with the Lorenzoni, Ludwig Straub and Ivan Werning. So everybody, unfortunately, is aware that the current COVID-19 pandemic is having a very bad and weak effect on the economy all over the world. I mean, last data that I read yesterday on the newspaper is new five millions unemployment claims have been made last week and six millions a week before in the US. And the data are not much better, actually are not any better in Europe, if anything worse because the pandemic is more advanced. All governments and central bankers around the world and I know this well, the ECB but the Fed and all other countries have been realizing that this situation is particularly dramatic and they have been trying to help and to think what they can do to help the economy. And clearly when central bankers and governments say we can do whatever we are willing to do whatever it takes, a natural question that comes about is, is it a pandemic, a situation, a type of shock for the economy that needs stimulus for spending? And so to frame this question in terms of the textbook macro model, one can, the natural question that people have been debating is, should we think of a pandemic like COVID-19 as a shock that is primarily about the aggregate supply of the economy or is it a shock that affect mostly the aggregate demand of the economy? This is very important to then think about policy. In our paper, the way in which we approach the problem is, well, let's think about the pandemic. When we think about it, the direct effect that we can think of a pandemic is something that looks more like a supply shock because either because people, there are some sectors where you know that there is more transmission and so people, consumers and workers, prefer not to be interacting with other people or because of lockdown measures, some sectors of the economy simply shut down. And so there is no production of some goods and services. And so this is a clear supply shock. However, our question is, can this shock that it starts as a supply shock propagate to the economy in the form of a demand shock? And can the demand component become even larger than the supply? And we are going to call a supply shock that becomes more importantly a demand shock, a change of supply shock. When the demand shock becomes the initial supply one, we are going to say that this is a change of supply. And we are going to use some models to show you that a series of models where we add ingredients to show you that there are two main ingredients that can transform a supply shock in a Keynesian supply shock, so mostly in the demand shock. And these ingredients are, first, we need multiple sectors. And in particular, we need these sectors to be very much complement to each other. And so when we think about different sectors, we think about some sectors have been shut down, like sectors where there is a higher contact among people, like restaurants, hotels, and sectors that may be complement to those are sectors like if you think about hotels, maybe people decide to buy less luggage or less cars to travel because they cannot go into a town. So if you think about restaurants, well, people won't buy fancy clothes to go to the restaurant. So naturally, these are not all goods are complement. And there are goods that are substituted. For example, if you don't go to the restaurant, you may want to do more takeout and so forth. So then we have to evaluate what is the stronger force in the economy. And when complementarities are strong enough, then we're going to have a change. The other ingredient is incomplete market, and we are going to talk about that. We are going to also mention two additional ingredients that can enrich our model and amplify the demand effects of supply shock. First of all, input out to linkages and across sectors. And second of all, we are going to endogenize business exit and show that this can generate cascades that multiply the effect of the original. Then we are going to draw some policy lessons from our theoretical insights. And first of all, the first point is that, well, when a supply shock becomes mostly a demand shock, clearly a stimulus can help. So monetary policy easing and fiscal policy may be desirable. But interestingly, we are going to show that for the special nature of this type of shock, the standard fiscal multiplier is going to be smaller than usual. And then second, we are going to show that in particular, what is going to be very important element of policy, if you want to think about optimal policy, social insurance. So ensure the workers in those sectors or the firms that operate in the sectors that are mostly affected by the shutdowns. And then we are going to also think about incentive that another, I'm going to show you that another important component of policies incentivizing firms to preserve matches, so to do job hoarding rather than break the matches. And this is going to be very important, not only for insurance dimension, but also to think about the longer run when the recovery is coming. So let me summarize our main results through this table. So we are going to consider four variants of building up on the basic one sector, complete from market model. And I'm going to show you when these supply shock that we think as a pandemic can transform into the demand shock. Of course, I'm going to show you that in the standard one sector, complete market model, a supply shock is a supply shock. And then this is a well known result. I'm just going to do it to set the stage for the other results. Then I'm going to introduce incomplete markets. And I'm going to show you that incomplete markets is not going to be enough to make the demand component strong enough to offset the original supply effect. And then I'm going to show you that what you really need, so an ingredient that is going to change things around is multiple set. If we introduce multiple sector, then we can, even with complete market, then we can have engine supply shocks. And when we add to the multiple sector framework in complete markets, the results are even stronger, meaning that the demand effect of supply shock becomes even possible and the range of parameters for each occasion supply shock that can arise is going to be even larger. So let me just with the one sector model in a one supply shock is going to naturally stay a supply shock in a sense that it's going to naturally create an upward pressure or the natural interest rate. Why? Well because in a sense it's like some positive news about the future. If you are hit, if you lose your job today and you know that tomorrow you're going to get your job back, then what you want to do is you want to borrow to do consumption smoothing. So this is going to push force towards demand and push up. This result is going to be strong enough that it's going to survive even when you introduce incomplete markets. Because basically what you can do if you are, if you cannot borrow, well at most you just don't borrow anything but you'll never go on the other side of saving. So in the end it's going to be a wash for the people who cannot borrow but these effects if there is some people in the economy that can borrow, there is going to be always a positive upward effect on them. Once we introduce a multi sector we have to think as I mentioned before about our goods complement or substitute. Again if we think about restaurants as a leading example, so restaurants are shutting down, we can think of a different type of sector. Some produce goods that complement to restaurants and some are substitutes, take out are more substitutes for restaurants and maybe are more of a complement. If you think about hotels then cars could be another example of complement goods. So what we are going to show is that when the complementarity force is strong enough actually then supply shut and become Keynesian. And this is because once, because once just some sector shut down if there are goods that are mostly complemented to that then demand for those goods is going to reduce beyond the original level and so there is going to be an additional drop in activity on the top of the sector so they are directly affected by this. Now this with complete market even when people who are working in the affected sector are insured. But if we introduce in complete markets the effects are going to be even stronger because now workers in the shutdown sector lose their income and so they have to cut spending. So let me show you these intuition with a simple, with a graphical representation to make it more concrete. Just imagine think about two sectors, sector one and sector two. Sector one is the sector where that is more where there are more contacts among people and so is the one that is going to be shut down. Sector two is the sector where people can work from home and everything is so it's not going to be shut down after the shutdown. Before the shutdown all the workers sector one sector two get income and they spend both in sector one and sector two and so everything works well. Once we have the shock we're going to represent the shock as the shutdown of sector one. Sector one is completely locked down so workers in sector one don't get any income. However they are insured because we assume here that there are complete markets so they still have some resources that they can spend and so both workers in sector one and sector two are still going to spend in the sector two. The question is how much are they going to spend? Are they going to spend as much as before? So the demand in sector two is going to be the same as before the shutdown more or less and the answer to this question of course depends on how complementary or substituted the sector two is relative to sector one. In particular if sector two is a mostly is complementary now, sector one it may be that the demand in sector two is going to be lower than before because now there is less need for the vice versa if it's mostly substituted it could even be that the other way around that there is a higher demand that offset in part the drop down in the lot. Now let's imagine in let's think about incomplete markets so what happened when we had in complete markets simply workers in sector one are not going to get any income anymore they are not insured so they cannot spend in extreme. So now the demand in sector two is going to drop even more and it's going to drop even more and in particular if we think about the situation of complements that we consider before it's going to drop more but even if goods are a little less complement than before it still may be that the demand component now dominates because of the workers in sector one that don't have any more possibility of spending. Okay so let me move into the model let me start with Veronica. Could you go back to the previous slide before we go to the model just to clarify a question from Michele Lanza. So could we think of the model could we think of these sectors also as countries so especially you think about if you think about EMU if there are input-output relationships for instance between Italy and Germany or even more importantly if the lockdowns are not synchronized across countries could your model also speak to that? Absolutely so yes I was planning to talk about that later but this is a natural extension of our model or another way of interpreting our model is thinking about different countries especially in Europe where countries are so where trade is so integrated you can think about countries that are locking down before than others or they don't and you're going to have similar results. And of course this is an impact on policy and I can go back to that later when I talk about policy. And then there's a second question from Florian Haider. What is the role of uncertainty? I mean you said we can think of this as positive news about tomorrow but right now people are very much worried about tomorrow. Yes so we here we we would abstract we're going to abstract about uncertainty of course if you add uncertainty into the picture you're going to have an even stronger demand component that is going to pop up but in a sense there are many other ingredients in the real world that may add on the demand side for example issues and things like that I can't touch on that but what what we do here is like more we want to abstract from many other reasons for demand shortages and try to see purely the supply water the the pure effect on the demand coming purely on from the supply shock that's why we are going to abstract from uncertainty paper but it's certainly very important. Thanks for clarifying so let's see the model. Yes so um so let me start from a simple complete markets one sector model and this is again these are well known results but it just will set the stage for next versions of the model. So preferences are standard there is one single consumption good there is a fixed endowment of labor and barn and technologies. So how do we introduce a pandemic shock a pandemic is going to be uh we are going to think about a pandemic as an MIT type of shock so it's going to be an unexpected shock the economy is in steady state and then unexpectedly there is a shock that is going to be temporary in our model and it's going to last one period so at time zero we are going to have a temporary reduction in labor supply so instead of n bar the endowment uh goes down to 1 minus phi n bar so phi n bar is the is the size of the shock and then from time one hours we are going to go back to normal to n bar endowment goes back to n bar and we are thinking about a flexible price allocation. Okay so uh the the question is um one way of thinking about is this a demand or a supply shock in the frame in the term in terms of our model is to think about what is the the pressure on natural interest on the natural interest rate. So think about the flexible price economy and so let the interest rate adjust what's going to happen to the interest rate at time zero if you want to keep full employment and then we can simply use the earlier question of the um of the consumers to back up what is the interest the real interest rate um knowing that people are going to be at time zero the um consumption has to be equal to uh production that is 1 minus phi n bar because we are in a representative agent model so we use market clearing here and consumption in period two is going to be back to n bar and so you can see that immediately this tells us that there is going to be a upward pressure on the interest and then um another way of of thinking and so this is the supply shock in that sense the supply shock another way of thinking about it is well let's just add that to the opposite example if interest rate thinks that 1 over beta the interest at the steady state level of interest rate and so in a sense let's add the downward rigid nominal wages and then let's think about if there is excess demand or demand shortages in that case in the in the first period okay so this exercise is useful especially if you want to think about the zero lower bound the kind of of environment what it cannot actually adjust really the interest rate and uh if you look at the same error equation and then you flood in the interest rate 1 over beta you can back up what is the demand for in period zero and the demand in period zero turns out to be equal to n bar like in the second period if the interest rate is the same of the of the steady state 1 over beta so this means clearly that there is an excess demand because we know that production needs to be 1 over phi n bar because there has been a shock and so I can summarize this example so one sector of complete market there one sector of complete market model a negative supply shock is a negative supply shock in a sense that generated an increase in the natural rate or in other words generate an increased excess demand if they interest rate it what's the intuition well intuition is as I mentioned before the idea is that if you have a negative supply shock today it's good news for the for the future at some point maybe not tomorrow there may be uncertain for when it's going to be but at some point we know that we are going to go back up and so what agents want to do if we are in a standard stylized one sector model they want to borrow not say so that's why there's an upward pressure on the enough now can complete market save us and kind of overcome these upward pressure on the interest let's see so we're introducing complete market in the simplest possible way so we're going to use a simple version of the model because we want to emphasize the mechanism and then we can think about how extent I mean generalizable these results are let me start within complete month so we are going to have agents that have access to a zero net supply one period bond it's going to be a it the budget constraint for agent i and and we are going to assume that the fraction mu of the agents face a border constraint so new agents are going to be constrained one minus mu of the agents are not going to be constrained and the ones that are going to be constrained are going to be handleborrow at all so a it needs to be okay and then the genetic dimension comes also in another dimension that is about the shock how who is going to hit by the shock so each agent as a labor endowment and it that is equal to n bar and when the shock happens so when there is this drop in the labor supply in the economy some agents are going to be affected some agents are not going to be affected so five of the agents are going to be affected and are going to have a labor endowment goes down to zero the others are unaffected so let me first of all let me say that this new help us span the different extreme of the economy so when mu is a zero we have no constraint agents the model boils down to the represent to the complete market version that i've just described when mu is equal to one we have the extreme where everybody's constrained okay um so how we think about the effect of the shock in this economy so we have to think about two sets of agents that is one set small let me one set that is the new five of the agents are going to be the ones that are affected by the shock so that they lose their labor supply but they're also constrained so if they're constrained here because they cannot borrow at all their income go to zero they're going to consume zero okay so they're kind of boring we can kind for the analysis we don't really need that so let's focus instead on the others agents in the economy who are the other agents in the economy well on the one hand there are the other affected agents who are unconstrained so these agents can actually borrow from the other agents who are unaffected and are willing to lend that to them okay and then there are the unaffected agents who are could be face both the constraint or not but the constraint is not going to be relevant for them because they are not at the heat by the shock so all these agents are going to group together and they each of them can be the behavior of each of these agents can be represented by a standard earlier equation and then thanks to omotetic preferences can aggregate these into an aggregate earlier equation that is a seed that is a looks exactly the same of a and then once we have that we are going to think about what's going to be the effect on the interest rate and so we have to plug in what is again the same exercise of before let's think that we are in a flexible economy and then let's let the natural rate unjust and let's see if there is an upward or downward pressure well if we do that and we want to keep the economy full employment c zero needs to be equal to one minus five and bar so this is the endowment at the time of the shock and then in period one what's the demand this is the market steering condition what's the demand for goods well the demand from the unaffected and unconstrained plus new fine and bar this is the demand from the people who are constrained and affected at time zero who are at the zero lower bound and sorry at the zero constraint and so now they're going to spend everything that they get from back from endowment and then this has to be equal to unbar the supply of labor and so from here we can back up what is what is c1 plugged into earlier equation and see that again there is an upward pressure on the interest rate meaning the natural interest rate is going to be higher than one of our beta so to summarize again one sector model even within complete markets a negative supply shock is going to generate a rising in the interest rate or we can do the same exercise as before if we kept the interest rate fixed we're going to have an excess so again why well the reason is basically the same as before as I mentioned at the beginning now there is this negative supply shock again agents that are not affected by the shock there is no problem for them but agents who are affected by the shock they well then they want to borrow and so they're going to who are constrained they cannot borrow but they're not going to save so at the most they're just downpending the effect of the others in fact in the limit where everybody constrained is constrained there is going to be no change in it all right now so we saw that with one sector model we are not going to go very far so with one sector model a supply shock is a supply shock so it doesn't seem the stimulus should be very useful if we think purely about this standard effect less than introducing an ingredient that we think is natural if we think about the COVID that is multiple sector why again we think multiple sectors are natural well because some sector we're you're going to think about lockdowns and you're going to think that some sectors are going to shut down but other sector may still be still be active in the economy and so we want to understand if the effect of the of the original shock is going to limit to the sectors that are shut down or is going to spread over into the other sector so we are going to do that with a simple two period sorry two sector model first and then this can be generalized but oh one thing I didn't mention in terms of generalization the result that I gave you before that with one sector model incomplete market is not going to make it to to generate tension supply shock is very general so can be generalized easily to more general framework of incomplete market okay so we can add this just so the result that one sector model is and the intuition is clear because it's the same as before if you have some constraint I mean at most you're going to kind of reduce the increase in interest rate but you're not going to generate within complete markets and on a force to create our pressure one okay so now two sectors c1 and c2 again this is generalizable to a continuum of sector and I'm going to show you this at the end and how do we model the preferences so we're going to think that there is a constant intertemporal elasticity of substitution both intertemporal and intratemporal across sectors right so one over row is the elasticity of substitution across sectors one over sigma is the intertemporal elasticity of substance technology is linear in both sectors so let's focus on the steady state so consumption in sector one is going to be equal to production in sector one is market clearing and production in sector one is going to be equal to phi and bar so we are going to assume that phi of the total endowment and a shared phi of the total endowment is allocated to sector one and one minus phi of the total endowment is allocated to sector okay this is by assumption and then consumption is equal to production that is equal to phi n bar at one minus phi n bar notice that we picked we chose this phi that's defined as share of endowment of labor in the two sector being the same of the phi that is in the preferences and that is the kind of the weight on the two bundle of goods five to the one one minus five to the wrong and we do that exactly to normalize the the price the relative price of the two sectors to one the p star is going to be equal to one by choice of the normality now how can we think about the pandemic shock in this two sector model we are going to think about that as an asymmetric and so temporary again and so we are going to think that the pandemic is going to completely shut down sector one so again the size of the shock is phi and bar because the size of sector ones okay the sector one is going to shut down so consumption production labor is all equal to zero times zero in sector one now how about the effect on the rest of the economy so we are going to think to think about what is going to be the effect on the interest rate we are going to think about the real interest rate in terms of good too for good too so why is that why we picked to do well because the interest rate for good for good too is a good in our simple version of the in our simple model where there is downward wage rigidity basically it is a good for which there is no inflation so the real interest rate is going to be equal to the natural interest rate so it's interesting to look at that and also a good one is going to be a good that disappears from the economy so the price is not going to be it's going to jump up to infinity so it makes more sense to think of or think in terms of but the result more general and I'm going to go back to that later when we talk about inflation so again here we have multiple sector complete model so we're going to have an a any a a little equation and we are going to for good for good too and we are going to look at this other equation and we're going to plug in what is a consumption at time zero and at time one in sector one in sector two in sector one consumption at time zero is going to be zero by the shock and the consumption sector 2 is going to be C2 star and similarly at time 1 we are going to go back to the flexible price and bar so we are going to go back to the steady state level. And you can see that once this is the case we plugged in this number we obtain the interest rate there can be a approximate downward on the interest rate under some parameters and in particular the result is that when we have a model with multiple sectors and complete markets the negative supply shock becomes occasional supply shock in a sense that generates a downward pressure on a natural rate when 1 over sigma is bigger than 1 over rho so when the elasticity of intertemporal substitution is bigger than the elasticity of substitution across sector and why how can you interpret this this result well and we can think about it in two different ways from two different angles let's first like a fixed sigma and elasticity of intertemporal substitution and think about in terms of complementarity across sectors then the commission tells us if 1 over rho is small enough then the demand effect becomes strong and and this is natural because this means that if the two sectors are complement enough then there is going to be a feedback on demand of sectors 2 coming from the shutdown of sector 1 as I described before restaurants closed then people are not going to buy new clothes so this is going to have an impact on demand for the sectors in the other sector. The other way of thinking about this is well let's fix the role of the complementarity across sectors well when the elasticity of substitution is large enough then we are going to have again demand shock dominate so why is that well one way of thinking about that is that you know that today some goods just disappear so if you can substitute consumption between today and tomorrow really or or we are eager to do that because elasticity of intertemporal substitution is large enough well then you want to substitute that so you want to consume more tomorrow because tomorrow some goods are going to go back into the picture right so these goods that you really would like to consume today that are infinite price today you you can consume right and this is a graphical representation of this condition so you can see that the the pink area is the the shaded pink area is the area of the parameter space where we have one over on the x axis and one over sigma on the y axis where Keynesian supply shocks arise and the other areas where a pandemic is just a simple rider supply and again we can restate the result in terms of the patient access demand if we just now how about incomplete markets so uh if we if we add on top of the multiple sectors in complete markets what happened well let's do again that in a um keep the same structure of incomplete markets model as we did before we have now a fraction mu of the agents who are constrained one minus mu who are not constrained and five of the uh workers are going to be affected and one minus five are not going to be affected so again we can focus on the unaffected and unconstrained workers to think about the effect um on the interest rate and in this case you can see that this is the error equation written in terms of the interest rate at time zero and then you can see that what is c2 zero so c2 zero is the consumption in sector uh uh two at time zero if again that we let the natural rate adjust and to keep the full employment this has to be equal to one minus five and bar which is the total uh employment and production in at time zero and how about consumption in two sectors in uh in time one while the we are going to have one minus five mu times n bar as a total uh income and then five of that is going to go uh two um sector one and one minus five so this is going to be this this can show once we plug this into the error equation that there is going to be a downward pressure of the natural interest rate for a larger for loser conditions than before for a larger set of parameter spaces so when um then then before and you can see that in this graph before in their pink shaded area uh we could we the pandemic shock was occasional supply shock now this area becomes blue becomes blue includes this triangle on the right this is drawn for the things where mu is equal to one so where everybody's constrained if you mute those uh is zero we go back to the pink shaded area so we knew in the middle we are just adding a slice of this fire okay so this means that even in cases where one of a raw um uh bigger than one over sigma you could have so for example here i'm showing you some parameters for which one of a row is bigger than one over sigma and so where in the case of complete markets you have um that a pandemic is a simple supply shock and so you have an output gap that is positive in the non-shocter shock sector then if you introduce enough marketing completeness so if mu the fractional regions were constrained increases enough then this is going to transform in an occasional supply of shock in the sense they all could get in the known in sector two start to be negative and the law and interestingly enough for parameters uh for which the shock is larger so for five larger where you have a bigger boom in the case of complete market you're gonna in sector two you're gonna have a bigger bust once you have enough uh marketing uh i have now uh completed my my table so uh just so some uh and uh complete markets or incomplete markets uh a pandemic is a simple supply shock once we introduce a multiple sector uh then even with complete markets occasions of play shot maybe even so they demand the facts of the supply shock maybe even larger uh when we have uh incomplete markets uh on top of uh multiple so now uh let's think about uh uh Monica before we move to policies there were two questions coming in so one is on the the which and the model is sort of a spillover right it's a kind of uh exogenous technological complementarity the question is whether this could also honestly so specifically Freud and I was saying right now I cannot go to restaurants but that also means I voluntarily avoid other sectors down also makes people not want to go to other sectors could this endogenous reaction been alternative to the more technological complementarity you are considering so yeah sure I mean that that's uh a little bit my next point after I took a policy I thought I talked about that so the complementarity of sectors I'm gonna talk I talked about here I just modeled in a very simplest way as pre as part of the preferences but I agree with Florian you can think about the complementarities in a broader sense and what we are thinking is what Florian mentioned about like endogenous decision about the other I mean the endogenous feedback effect of what happened in one sector to the other but another way of thinking about it is to think also about input output linkages across sectors is another way of thinking about complementarities and I'm gonna go there in a few minutes so I totally agree complementarities you don't want to take that as a simple parameter that you look at in a data standard model but it's something that is broader than that okay and then the second question that came in is by Christoph Camps he's asking whether your paper now reviving the law of markets by Sey so Sey's law which basically was about a fallen supply destroying demand yes absolutely I mean we mentioned that in the paper yeah we mentioned that in the paper I mean we think that we have a different interpretation I mean a different take on that but we mentioned that that there is a link to that okay and then there were two further small questions one is by Luca Dedola whether in the incomplete markets version of the model also the persistence of the shock matters yes here we have a pure temporary shock so it's only a regular shock but if you embed a more shock with persistency for sure that that would be important too okay and then the last question by Philipp Hartmann is whether sort of the distribution if you had more than two sectors right if the distribution of containment policies matters so for instance if all sectors operate only at 20 percent because everybody's working from home with low productivity because it's on a form containment policy would the balance between demand effects relative to supply effects change so you're saying if each sector operates at the lower the some intensity then yeah yeah so if it was more of a uniform thing across sectors sort of a partial lockdown yeah we're gonna I'm gonna show you a more general model at the end if I have time with the continuum of sector and then you can interpret that more broadly as an intensive measure of activity if you want but yes okay perfect thank you how about policy so the first thing like let's just think about a simple policy like government spending or transfer that are sector specific and then it's gonna become our workers workers model that is multiple sector plus in complex model so it's striking that if you if you think about government spending and transfers in this model you can see from our proposition that the fiscal multiplier is one so you lose the multi like the the the the second round of Keynesian cross-operating in a standard model with nominal rigidity is and why is that because of the nature of the shock because if the shock is that you cannot have the feedback effect on the fact that people's the people that government spending is going to increase the income of people the income of people are gonna spend more they cannot spend more in sector one because they're gonna spend more in sector two but that's not really where it's needed it's where it's needed in sector one because it's where the workers are the highest right so fiscal policies are still beneficial because the fiscal multiplier is positive but they're not as strongly beneficial as in a standard model how about optimal policy so we want to I mean this is a stylized version of thinking about optimal policy but one important ingredient that we need to add if we want to think in terms of optimal policy is some else dimension because of course I mean if we could just not shut down the sectors that would be optimal there would be no problem at all so but the reason why we are shutting down some sectors or there may be some some some problem but not as big as if you shut down but why you are you want to shut down sector because you think that that is going to help the health of the economy to improve so we need that dimension to think more broadly about so the way we are going to do it is in a very stylized way we are going to have an additive health component that depends on four things one on side that is the shock shock or no shock is binary and then the health of an individual depends on his own consumption in sector one and his own work in sector one because sector one is the contact intensive sector so if you if you are a both a customer in that sector or a worker in that sector you're exposed to contact and so you're exposed to the virus and so that's why the more you consume there the more you work there the higher the lower is your health the higher is your negative health shock and then so also the total level of production in sector one is going to affect your health and this is a standard delta externality so the more active the more people the more active the more production there is in sector one there are more contacts that are and the more people are there the highest is the chance that you get the virus if you think about restaurants if lots of people at the restaurants that you have a higher chance to get the virus okay and so there are three sources of inefficiency in this economy again we are in the complete market multiple sector version of the economy we have the health externality here as in a paper by Martin Sergio and Trombard we have a lot of issuance that comes from the complete components and then we have maybe we have also involuntary unemployment okay because because working in sector one is is a problematic what you have so what is that let's think about a shock an economy where there is a shock but the there are no lockdown policies so you could potentially work and consume in sector one still and and we're writing about optimal policy in terms of pretty much forced still in this economy where you do not have the lockdown policy maybe the production and output in sector one goes down simply because people in the personal decision optimal personal decision to work or consume there is to reduce their consumption their work there and their labor supply in that sector because of the higher virus free and so you can have involuntary unemployment but now the first remark is well can this is this unemployment inefficient well no this unemployment may actually not be socially inefficient in fact there are two forces that may make that unemployment efficient or inefficient on the one hand you have the Keynesian wedge so that of course tells us that the unemployment is bad because amplify the demand effect for the rest of the economy and so this is of course creates some inefficiency but on the other hand you have the PQP given externality that actually it's good that there is an employment because means that there is less contact in the economy and so less spreading now let's think about remark two okay once we think the shutdown for sector one is optimal and because the PQP an external effect is stronger than the Keynesian wedge effect well then these of course is going to generate the demand effect of the Keynesian supply shocks that we've talked about okay so in that sense there is some desire for stimulus in that dimension but as remark three is well within complete markets the way in which the stimulus that the the best way to provide that stimulus through targeted transfers and targeted transfers can help obtain the first best because they keep the three birds at the same time with one stone which is one they provide insurance solving the incomplete market problem two they help raising the natural rate helping if we are at the zero lower bound and three they may make the health policy more desirable because you can directly address the issue generated the the issue generated by the policy and that made it desirable to reduce the spreading of the do more more lockdown policies and reduce the spreading of the now um so socially is necessary and targeted transfer social insurance is key to obtain first best so how now of course this opens a big debate about how to implement the social insurance and how can we think about who who for example should we reinforce policies like unemployment insurance or should we reinforce incentive for firms to keep their work and so in the in the paper we also have an extension where we consider the importance of job matters so if we are in a business that has some workers well at that and your lockdown at that point you can keep your worker attached to your firm by paying the wage this is of course a cost but on the other hand once the the sector reopens your your workers that is valuable for you because maybe you invested in human capital because these are skills that have been developed in your firm and so forth then you have this valuable worker tomorrow on so v1 is the value of the worker that is the wages you have to pay today on the other hand of course you may just you may decide that the cost of paying the wage is too high and shut down and destroy the match so what is your choice what is what is that effect of your choice while the interest rate is important in affecting your choice because it determines how much more you you put value you put on the future rather than paying the cost today and and the interesting thing is that labor hardening provide perfect insurance at the same time is a way of providing insurance because it exactly gives the keep alive the wage for those workers who are in the lockdown sector so it kind of reduce the demand and adverse effect of the lockdown policy so we we make the claim that labor hardening is positive also if we think a longer around the consideration because in the longer run when the economy reopens at that point you are going to have your valuable matches back into the picture and so both because it provides perfect insurance today and before long run considerations of efficiency and avoiding job destruction labor hardening seems optimal so seems could incentivize labor hardening and monetary policy in general of course expansionary monetary policy is going to help in in that direction because by lowering the interest rate that help firms optimally choose to retain their workers now of course there are liquidity issues of firms and so on so when we think about expansionary monetary policy in that dimension for labor hardening consideration we want to think more broadly about the monetary stand and so think about providing liquidity and the all these type of measures to banks and so forth it's all these type of measures are going to help in terms of give the incentive to the firm and then give the liquidity to the firm so to keep the liquidity in the system alive so that the firms may have access to liquidity to keep their workers now i'm going on time i'm doing bad on timing right we have 10 minutes it's it's okay thanks okay so you tell me when i have to stop i'm gonna go for so one more one more issue is that i want to touch upon briefly in terms of policies well when we want to think about inflation of course you have ECB so i don't have a i need a slide on inflation and i want to say our model is very stylized in terms of inflation because we have like stark downward wage of nominal rigidity and of course so we have actually no deflation for example but the main message of the model if you think about generalization of the model we can easily generalize the model in a more general standard setting and the main message of the model in terms of inflation is that demand efficiency means no inflationary pressure so when the supply shock is a shocker than i have no inflationary infinity deflationary pressure pressure and to think about the one way of thinking about the the mechanism i just came before is to think through prices and i think he's useful to take his time so when you think about sector one what about prices in sector one we said sector one is price is essentially the integrity so if you could price the value of going to the restaurant with no risk of getting infected so it's like there is a spike in price in the price level in the shadow price level for sector one and so this means that actually if anything there is going to be a deflationary pressure because tomorrow these prices are going to go back to China so a different way to and this is you prefer to wait and consume more tomorrow okay so this is a different way of thinking about that but the main message is that when Keynesian supply shock arises deflationary pressure may be in the in the picture so effect on the real rate is also deflationary this complementarity among sectors that this seems important for the for our key results and as i briefly mentioned before one thing that i want to highlight is that input output linkages are an important way to think about this complementarity stronger so think about again about our rest accounting services and these account and then suddenly they stop demanding these accounting services the demand for account immediate inputs producing non-affected sectors actually there there is going to be even stronger demand the thing is that the supply chains in our that the the supply make demand the personal demand are supply chains okay not to so typically people think a supply chain in the other way from upstream to downstream so thinking about affected sectors that produce intermediate good for unaffected sectors that are not important but we want to emphasize this this dimension so how do we introduce that into the picture i don't have much time but the idea is that sector one who uses some intermediate goods that are producing sectors when now clearly the upward pressure the natural interest rate is going to be even higher because once you don't have any more the demand in sector two from coming from the demand for the intermediate inputs from sector one you're going to need even more to generate even also to kind of compensate for that so the effect of the demand effect is going to be so the last thing that i want to mention is that and i mentioned that before briefly so first of all these two sectors can be generalized to a continuum of sector you can potentially generalize it introducing an intensive measure if somebody has mentioned before the the one dimension that i want to highlight is that if you have a continuum of sectors okay and you have some monopolistic competition that generates profits and you assume that each variety is produced by a separate worker from so extending our model well then basically the results are exactly the same as in our model within complete markets and two sectors but then if you add an indigenous exit decision things becomes the demand effect becomes even stronger so if you now add a fixed cost a random fixed cost of businesses have to have to cover a fixed cost to stay active now it can be that when some fraction of sector shuts down there is going to be a fixed effect on another sector because this is going to reduce demand in other sectors but in these other sectors businesses suddenly may exit because the demand is not going to be enough to cover the fixed cost and so then this exit is going to fit back in demand from other sectors and so on generating a cascade of exit that amplifying the shutdown the endogenous shutdown in the economy i don't have much time and let me show you this figure this shows that the on the x-axis is the active firms and businesses and on the y-axis there is employment at time zero and you can see that the the business exit lock of the intersection between these firms is the effective active businesses so you can see that in the first here is before the shot the intersection is at one minus five at equal to one means all businesses are active in steady state suddenly if you have a you take out one minus five sector from the economy because there is shutdown then the firm exit lock shifts to the left but then you can see that the intersection between the red and the blue line is going to be at a point that is one minus five meaning that the active firms are going to be much lower than the one minus five in the economy so there is more on the top of the of the five firms that are shut down there is going to be an extra number of businesses that are going to exit endogenously and this is going to okay so let me conclude the conclude with the same slide table that I showed you at the beginning to re-emphasize our main results first a single sector we cannot the supply shock a pandemic is a simple supply shock so stimulus is not but if we introduce multiple sector and in particular if you introduce multiple sector we'd incomplete market and we'd enough complementarities among sectors which again can be interpreted in a loose way especially considering input output linkages well then it's a dependent so having demand side demand the effect that are stronger than the original and let me mention again the point the first question somebody asked again now that you have the model in mind I think it's natural to think about how about international like context like Europe and think about different sectors at different countries and this is a natural is an interesting avenue for thought and I think that this of course could call for a coordinated policy across countries thank you many thanks Veronica um one question that came in from Sebastian Schmidt is whether in the context of your model the type of policies we've seen in Germany it's called the Kurzarbeit where the government pays part of a worker's wage without them going to work from the kind of label hoarding perspective is that a good policy in the model yeah I would say so and this I think it's something similar to Casa Integrazione in Italy where the government kind of pay for the workers to stay attached to the firm I think that exactly this type of policy in this context seems positive I don't know if my quarters have anything to add on that dimension and the second question I guess that's sort of the last question I mean there's a lot I think your papers prompting a lot of thinking here in Europe in terms of more targeted policies targeting certain sectors of course primarily in terms of fiscal policy so as we interact a lot with these policy makers what's sort of the the main message coming from the model I mean you can also kind of step outside of the model like what what sectors to target is it sectors with strong complementarities so so first of all I want to say that this is a question that is particularly important so when we are at the top of the health emergency like in some countries like in Italy for example where all the sectors are shut down of course then the mechanism that we highlight and then the help of the sector is much more difficult to implement at that point the policy that you want to think are relief type of policy and so forth but as probably we are going to get out from the extreme lockdown measure then you're right that then our about different sectors is going to become important and how to target that insurance is going to become important and the the I mean one way of thinking about it is well of course you want to target you want to help the businesses in those and the and the workers who lose their their job so that you are attached to sectors that are affected from the mostly affected from the lockdown because these are the workers that then are going to spend in other sectors rather than help directly the other sectors that are the ones that have a complementary but of course if anything if you want to add some some help to other sectors that are active then you want to I mean the model would would would call for more action in those sectors that are somehow more linked to the sectors that are shut down in particular sectors that produce intermediate input for for sorry sectors that are producing yes intermediate inputs for those sectors that are shut down so that temporarily have a particular shortage in demand beyond what what would arise if the other sectors were not shut down thank you my quarters want to add something on that all right I don't hear anything so uh oh who's that yeah I yeah I just this is Guido yes Guido go ahead all right yeah no on the on the asymmetry between sectors I think that is I mean I totally agree with everything that Veronica said and yeah I think the idea is that when all the sectors are shut down like where we are in a really total lockdown then we're not saying that uh social insurance is not important but I guess at that point policy is essentially doing like disaster relief you're just helping people that get hurt but you don't worry at all about restarting activity because it just cannot be done and so in a sense like all our messages that the moment some sectors can be active then the question as are these sectors at the right level of activity or are they going to be too depressed and so we want to help them out and so I think is is when you start having some asymmetry when you have to start to have some sectors that could be operating like some manufacturing sector where you can do things without compromising the health of the workers then the question is like is activity in those sectors too too low that's what where we would want it to be and what kind of interventions can help so the asymmetry is very important thank you Guido so I'll surely be thinking about these issues and working with the input output tables and etc um if unfortunately time is up so I wanted to just end by thanking Veronica in a big way for preparing this very interesting presentation for us so on behalf of the whole ECB there were several our board members on the call as well thank you very much for doing this Veronica thank you please stay healthy all the best and I'm gonna close here for now thank you very much everybody bye bye now