 So my presentation is going to be based on my work with Guida Lorenzoni, Ludwig Straub and Ivan Werning. When we start working on this project back in March, when the pandemic starts spreading in the US and all over the world, and the original issue, of course, was how do we think about the pandemic in a macro model in order to think about possible policy that can help dampening the adverse effect of the pandemic? And at the time, governments and central bankers all around the world were thinking about different types of policy, and there still are, to try to help the economy. But the first fundamental question behind any monetary policy or any fiscal policy is, should we do any demand stimulus given the type of shock that hit the economy? And then if so, if this is the case, then we can start thinking about what is the best and most effective way of helping the economy. So the debate started to be framed in the context of the standard macro model as whether the pandemic was primarily a shock affecting the aggregate demand side of the economy or the aggregate supply side of the economy. However, we know that in a more rich macro model, demand and supply are interlinked. And what I'm going to emphasize in my presentation is that when the pandemic is the shock that hit the economy, this becomes particularly important. And so it's too simplistic to frame the question in terms of demand versus supply shock. In particular, what is why I'm saying the pandemic is a particular shock? So the defining feature of the pandemic, according to our view, is that the pandemic is an asymmetric shock. So it's a shock that needs to be thought of in a multi-sector model because it's a shock that hits some sectors in a direct way, and some other sector not at all in a direct way. I mean, it can still affect others. So when I thinking what are the sectors that are directly affected, those are the sectors that require some personal interactions and contact. And so where there is fear of spreading the virus if there is trading. And there are other sectors that can well operate without any personal interactions and then are not hit directly by the shock, but they still could be affected by the shock but only directly. And so our question then is, I mean, there are these sectors where there is a high contact, that are high contact intensive where actually it may be good, it is probably good that the activity goes down. So there is a reduction in the output and activity that is efficient. But the question is, when the shock propagate to the rest of the economy to the other sectors, does this generate actually inefficient shortage in demand? And we are going to label Keynesian supply shock a type of shock that is on the supply side directly on some sector but propagates to the others generating demand shortages that may dominate the strength of the original supply effect. And when I talk about supply shock, I mean supply shock in a macro sense. So I mean a shock that reduces the efficient level of activity in that sense. So when is that Keynesian a pandemic? Why could be that the pandemic is a Keynesian supply shock? And in our paper we emphasize different ingredients in a macro model that can amplify these demand shortages effect and so make Keynesian supply shock even more plausible. Two main components that we emphasize are complementarities across sectors and incomplete markets. But there are other things that we talk about in the paper and I'll touch on them a little at the end. Some of them the presence of input output linkages across sectors can actually amplify the complementarities across sectors. The possibility of business exit cascades or job match distractions may all be ingredients that can generate the demand shortages amplification. So now let me start giving you the mechanism of our model. Let's think about two sector. A which is the high content intensive sector and sector B which is we doesn't doesn't require any personal interaction. So of course it's an extreme world in reality we have multiple sector but it's going to be enough for me to give you the name. So there are going to be workers in sector A worker in sector B. This is like before the shock everybody's working gets income from its sector and spend in both sector. Now the question is how do we think about the pandemic? We are going to think about the pandemic as a complete shutdown of sector A. And so what's going to happen is that basically we don't have any more income going to the workers and nobody can spend any more in sector A. How do we think about that? Well you can think in a very simple way about that as at the beginning a full lockdown so sector A just is shut down by the government. But you can think about this also as in the absence of lockdown policies just as a decrease in the level of activity in the sector either because there is a reduction in the supply of labor. Or because maybe people are worried of working in person or because there is a reduction in demand for those goods because of a preference shock because people are just worried to go to the restaurant because they're afraid to get infected. So this is an extreme way now activity completely shut down. The question that we asked is what happened to sector B? And in particular let's assume first that there are complete markets so sector A workers are insured against this shock so they can actually spend in sector B. In sector B workers still get income and can spend still in sector B. The question is now the demand in sector B is going to be lower the same or more than before the shock. And now this depends in a complete market version of the model depends purely on the strength of the complementarity of the good produced in sector B and sector A. Relatively to the inter-temporal elasticity of something. So for example think about sector A being restaurants, activity in restaurants goes down what happens to take out food or grocery stores. And maybe there there is actually an increase in demand because a grocery store or take out food is a good substitute for restaurants. But on the other hand there may be other sector that are more complementary to the going to the restaurants. For example think about fancy clothing online shopping. Maybe you are going to reduce by less clothes if you go out less. And so the question become overall are these complementarities stronger if the complementarity are stronger it may be that overall you will see a drop in demand for sector B. If it is stronger you may see an increase in sector B. And this graph here represent the parameters where actually Keynesian supply shock happens which means that actually epsilon that in we will see is the elasticity of inter-temporal substitution is small enough. So let me introduce on top of that incomplete markets. What's going to happen then? Well if we introduce incomplete markets what happens? We lose the transfers between workers in sectors B and workers in sector A. So now workers in sector A don't get any income because sector A shut down. And let's assume an extreme version of incomplete market they cannot borrow at all so they cannot spend anything. Now we have only workers in sector B spending in sector B goods. So even in the case so you can see that there is going to be a drop in a further drop in demand beyond the fact of the substitutability versus cases where sector B goods substitute. So that incomplete market version of the model there would be no Keynesian supply shock actually may be that the lack of income for the worker the of insurance for workers in sector A may make demand shortages dominate. And you can see here that we have a larger parameter space where Keynesian supply shock. Now how so so far I was talking about complementarity and substitutability in the goods in terms of preferences. But it may be that the complementarity is across sectors as I was mentioning before is stronger if we consider input output linkages across sectors. In particular about restaurants now where they used to demand dishwasher machine or repair service accounting services. And now these demand is going to be not it's not going to be there anymore because the restaurants are closed. So then the sector B is going to see a drop a further drop in demand coming from the lack of demand for intermediate inputs that they produce. This makes the parameter space even large. So let me go in a little bit more details in the model. We have two sectors sector A and sector B. And they are the consumption is aggregated according to a CS where accident as I was mentioning before is the elasticity of intratemporal substitution across sectors. And the Sigma is the standard elasticity of inter temporal substitution. There is a linear technology in both sectors. So one to one from labor to output. Measure one of workers each one of them as a labor endowment given by and bar. Now there is among all the workers five of them a measure five of them are going to be attached to sector A. So they specialize in the production of sector A one minus five minus five of them is going to specialize in sector B. This means that in this baseline version of the mother labor is immobile. In the paper we also study the case where we relax this assumption and we consider also the other extreme where there is fully mobile. As long as there is some degree of immobility as long as all the messages are going to go through. Now what happened that what is the asset that the agents have access to only one asset one period bonds in zero net supply. So the budget constraints of these workers is simply the following. So what are they spending they spending good A good B and in assets and their resources are labor income and the returns on their initial asset position. Of the workers there is a fraction you this is the way in which we model incomplete market. There is a fraction of the workers that face an extreme boring constraint meaning they cannot borrow at all. Right. We are going to. So this is a very special way of representing incomplete market in the appendix of the paper. We show that all the results go through in the more general version of market incompleteness. But this is this is a way of formalizing incomplete markets is makes it very tractable and also allows us to do this. In limit case that like moving new from zero to into one to span all the possible degree of marketing. So there are two possible limit days that are interesting. One is when epsilon goes to infinity. Basically the two goods are perfect substitutes. It's like there is one sector model. The other thing is when new goes to zero, which is the limit where we are going to converge to a situation where nobody is constrained, which is equivalent in the aggregates to a complete market version of representative agent. Now let's think about the shock. How do we think about the pandemic. The pandemic is an MIT shock. So an unexpected shock that is going to hit the economy only for one period. So at time zero, we start in steady state and everybody has zero assets. So we're going to start from a situation where everybody has zero assets. And then at time zero, we're going to have this temporary shutdown of sector a in terms of our model. This means that all the time workers who are specialized in producing in sector a going to get an end up labor and diamond goes to zero. And then from time one on, we are going to go back to normal to a flexible price allocation. Now the question of whether this pandemic is a case on supply shock or not is going to be framed in the following way. They're going to say they're going to be downward rigid nominal wages. And let's assume that the central bank keep the interest rate unchanged. In that case at time zero is there is their assessment or insufficient demand. Okay, if there isn't sufficient demand, then this means that the pandemic is a case of supply shock. Otherwise is a standard supply shock. And the question can be also framed in a different way instead of saying central bank keep the interest rate and change. You can think how about if the central bank try to achieve full full employment is the natural rate going upward. Okay, so let me start from the one sector version of the model. So epsilon going to infinity. In that case, a negative supply shock always generate excess demand. So it's a standard supply shock. There is no possibility for a change of supply shock with or without incomplete markets. Why is that? Because if the shock is temporary. So if we have one sector, what's happening is simply that agents are going to see that shock as a good news about the future. They just have a negative income today and they expect a higher future income. So what the agents want to do is to borrow. And in the incomplete market version of the model, what happens is that agents, some of the agents may not be able to borrow. So, well, that's too bad for them. But this never means that they are going to be saving going overshooting the other direction. So at the limit, what's going to happen is when new is going to one that there is going to be no excess demand, but we are not never going to go the other way around where there is a shortage of them. How about when you introduce multiple sectors? Now, let's start with introducing multiple sectors with the complete market. Okay, in that case, we are going to have that is a negative supply shot can generate demand shortage if and only if sigma is greater than epsilon. This is the graph that I showed you at the beginning. We have the elasticity of intertemporal substitution on the y-axis, the elasticity of intratemporal substitution on the x-axis. You can see that the purple area is where the engine supply shock arise. What is that if you have a shortage in demand when epsilon is small and sigma is high? Well, there are two forces in this model that are opposite in opposite direction. And then it depends on which force dominate is going to give us if there is shortage in demand or not. On the one hand, when sector A shut down, what's going to happen? Well, there are some goods are not going to be available. So the shadow price of these goods is like spacking to infinity. So what does it mean that people know that now they can only consume a good being? And so they may decide that if they like intertemporal substituting now, they may decide to save today to consume more tomorrow when good A is going to be available again. People may prefer not to spend too much in takeout now to go to more to restaurants when they're going to be available. And you can see that from the standard earlier question. What's going to happen is that the price today are going to go up because of the shadow price of the goods A. And so this means that you want to save and buy how much is going to depend on this sigma here. On the other hand, we have the other force which is you want to spend more in good B because it's true today there are no good A anymore. This means that good B becomes relatively cheap. And so people want to, if they're relatively good substitute, they want to spend a little bit more in B. But of course, how much? Well, this depends on the degree of elasticity of intertemporal substituting. And depending on which force dominates, it may be that people decide that they overall prefer to spend less today and wait for good A being available tomorrow, which is going to be the case of our shortage in demand. Or it may be that instead good business, good substitute, good enough that they prefer to spend more. Now, how about when we're introducing complete markets? As I mentioned before, when we're introducing complete markets in our model, the workers who work in sector A will lose their income. Of those, there may be some of them that do not have access to credit markets so they cannot borrow. So in that case, you're going to be this part of the amplification in the drop in demand. So this increases the space of the k-jump supply shop. If mu goes to one, and as mu increases, this area becomes bigger. In fact, if mu goes to one, we're going to have this horizontal. Okay, so these are the main baseline results. Now the question is, well, let's first have a look at some data and then let's think about policy. In terms of data, I just want to give you, this is going to be just a graph for the US, but there are data available for other countries. I want to just make the point that if you look at real-time data that have been used in a bunch of new paper, for example, for spending or also for employment, this is a graph from a paper by Natalie Cox and co-author that used credit card data. You can see that there is a drop for different sectors. On the left-hand side, you have the essential and on the right-hand side, you have the non-essential sector. You see that overall, you see a drop in the ending in all the sectors, with a lot of heterogeneity. So there are going to be some sectors that are hit more directly, that are going to suffer a bigger, larger drop in demand, as expected. So for example, think about accommodation here, the blue line. Clear is one of the most affected sectors, but even sectors that are not directly affected like, for example, clothing, look at the green line here, has been dropping a lot. So there has been a pretty spread out drop in spending in all sectors, directly or not directly hit that kind of speaking favor of the fact that this shock is transmitting to the whole economy. The question is what can we do in terms of policy? What is best to do? Now, of course, if we believe that we are in a situation where actually the pandemic is a potential supply shock, so there is some shortage in demand, expansionary policy in general is something beneficial for the economy. But the question is which type of policy is more affected. So let's start thinking about the government spending. So this is the budget constraint for the government. G is under government spending, and then TA and TB are targeted transfer to workers in sector A and workers in sector B. And D is debt. Now, let's consider a small increase in government purchases financed by debt and future taxes on the workers. We can see that in the paper we show that in this case, the fiscal multiplier on government spending is going to be equal to one, which is smaller than the standard multiplier that typically is bigger than one is in new Keynesian models, where shocks are symmetric. So why is that? The fiscal multiplier is not as large as in other standard models. The reason is that the typical mechanism that creates a second round Keynesian cross that amplify the effect of government spending is not as effective when the shock is asymmetric, in particular because the workers in sector A now, who are the ones that are hit by the shock, are not going to be reached by this type of policy. Because if sector A is shut down, sector, the government spending is going to happen only in terms of goods of sector B. This means that this is going to increase the income for workers working in sector B, who are actually the low marginal propensity to consume guys. And then this is going to increase their demand and their income. But it's not going to achieve the income of the many unemployed workers in sector A, who are the ones that have the high marginal propensity to consume. In a sense, this is just similar in the spirit to the argument in the job market paper of Christina Patterson, who emphasized that the fiscal multiplier is large in recession because there is a correlation between the sensitivity to the shock and the marginal propensity to consume. In our model is the reverse is the similar argument in reverse. Here we have a negative relation between the sensitivity of the of being a beneficiary of beneficiary of the government spending and the marginal propensity to consume. So the ones who get the benefit of the spending are the ones with the lowest marginal propensity to consume. So this dumpens. Now going back to our two sector figure. So what about other types of fiscal policy? Well, then, of course, I mean, of course, I mean, it is in this framework, we can see that the problem of fiscal spending is that actually those workers in sector A who don't have income and may not be able to borrow are not reached by this standard fiscal multiplier, so it's going to be spending in sector B and so it's going to result only in increasing income for sector B workers. So they instead what can be beneficial in this framework is really transfer so direct targeted to the workers in sector A who are the ones that are mostly affected in this recession and are the ones that should insure us. So insurance here first are the are beneficial and is because of the asymmetry of this. And so let me write down targeted transfer to workers in sector A as the product of a replace the rate of law and the level of employment and bar before the shock. So what is interesting is that if we have the replacement rate on the x-axis here and we look at output, we can see two cutoffs, raw prime and rodable prime. There is a first raw prime cutoff where such that if raw increases below raw prime, output is going to increase because we know that there is shortage in demand and the lack of full employment in sector B. So increasing transfers is going to help that. The same time is going to help in welfare as well because it's going to also on top of increasing economic activity, increasing insurance. But if we keep increasing raw after raw prime after rodable prime output now achieve the full employment level so it doesn't increase. But yet welfare keep increasing because even if we are at the full employment, there is an improvement in welfare because workers in sector A can increase their insurance. And the rodable prime is the other cutoff where we achieve full insurance and then welfare is going to decrease after that. So raw prime to get full employment, we don't need 100% replacement rate because actually sector A is closed. So people don't need the same income as before to spend a cheap full employment. You only need to spend in sector B. But to achieve social insurance, full insurance, you actually need depending on the design of the taxes in the way we did it, you actually need 100%. You definitely need a replacement rate that is higher than the one that gives you. So to sum up, targeted transfers here are good not only because they increase economic output, but also because improved social welfare. That is one of the problems in when the shock is asymmetric and they are in complete mud. So overall, there have been fortunately, there have been many fiscal packages all over the world in the US, in Europe, in China. And these fiscal packages have included a different way of providing targeted transfers. And I believe they've been proven to be successful to some extent. So this is just for the US, the CARES Act. So this is a figure from a paper by Raj Chetty and co-authors who use microdata administrative data for on spending. And I think this graph is a good way to see that the CARES Act that included both extension in unemployment insurance and targeted business support was effective in providing more insurance. How do we see that? The green line here is spending for the top quartile of the distribution, the reach. And the blue line here is spending for the bottom quartile of the distribution, the poor. You can see that the spending of the rich has been actually hit by the recession. What is interesting, although the income of the poor has been hit harder, as you can see from employment data. But what is interesting here is that when the CARES Act was implemented, there has been a recovery in spending of the poor who are the most recipient of the CARES Act, much more than the rich. And so this is kind of a sign that this insurance mechanism is at work. Now, before moving to monetary policy, I want to make a last point about fiscal policy. So far, I have not mentioned health at all in a direct way, right? So in my model, the pandemic is a pure asymmetric shock. Behind the scenes, of course, there is an health dimension that matters. And for example, I've talked about possible lockdown. So to make sense of this world and of this policy, you need a health dimension in the model. And I think it's interesting to think about the complementarity between fiscal policy and public health policy. So if we introduce a health dimension in a simple way in the modeling preferences, where people care about consuming sector A, which is the eye contact intensive, because if they go more to restaurant, they get sick more with higher probability. They care about working in sector A, because if they work as waiters, they may have a get sick with higher probability. I think about the total level of activity in sector A, because if more people are in the restaurants and there are more tables around, there is a higher chance to get infected. This is a standard health extremity. And then this is just a dummy, a pandemic or no pandemic. So if we introduce this out, they mentioned now there are three issues that policy can address. One is a possible demand shortage in sector B that I emphasize in our main results. I think there is lack of insurance, and there is now also the health externality. So is it efficient to shut down sector A or not? There are now, of course, there is a tension between two forces. There is a Pyguvian externality. I would tell us that probably reducing activity in sector A is good to reduce the spreading of the virus. But there is also our occasion wage operating that if you reduce activity in sector A, you may have an efficient reduction in activity in other sectors, where the spreading of the virus is not affected. So it depends on the relative size of these two forces. But what is more interesting is that in this world, targeted transfer achieve first best. Why? Not only because they're going to help with economic activity, with stimulating demand in sector B, not only because they're going to help in creating more insurance for the targeted workers, also because there is a complementarity, as I once mentioned, between the fiscal policy and the public policy, in the sense that providing targeted transfers to the sector that are more responsible in spreading the virus reduces the cost of containment policies like lockdown or partial lockdown. And so this is going to possibly make this policy more desirable, which may be overall beneficial for the economy to reduce the spreading. How about monetary policy? Well, first of all, let's think about the standard stimulus effect of conventional monetary policy. If we think about conventional monetary policy, the main argument is similar to government spending, meaning that monetary policy is beneficial in the sense that we live in a world where the pandemic is occasionally supply-sharp, there is shortage in demand that is an efficient in sector B, and so monetary policy is going to help stimulating demand, and so achieving an achievable employment. However, it's not going to be able to help with insurance directly, because monetary policy is going to be not targeted, it's not going to reach in our model sector A workers, given that sector A is shut down. So there are more broadly two challenges that monetary policy face that are how do we think about inflation? So in our model, there are potentially different signals from inflation that comes from different sectors, because in sector B, you're going to see a drop in prices, because if there is demand shortages, but in sector A, you may see an increase in prices. So there are going to be some sectors think about the toilet paper where there is going to be shortage of supply, where you're going to have a supply and increase in prices, but there are going to be other sectors like retail, online shopping and so on, where you can see a drop in prices because of the shortage in demand. Even to complicate more things, there may be some goods that are missing. For example, the price of a football match that may be the same is just that the goods is not available. So how do we think about that? So it's going to depend then if you are going to think about that, if you think about inflation in terms of a measure of slack, then maybe it's good not to look at the goods that are missing and think about the measure CPI overall, that is what you see that are the price of goods that are traded and in that sense, this is going to tell you if there is need to stimulus. On the other hand, you may care about inflation also as a measure of cost of living. In that case, instead you want to introduce the goods that are missing, because that matters for the value of the cost of living or for welfare measure of inflation. So more broadly, what happens here is that some inflation may actually be just the symptom of the fact that relative prices need to change because terms of trading are changing. So it's a signal of a change in terms of trade. So how about, I mean, there is a other model. So there is an interesting new paper by Woodford that actually has a model in a similar same spirit of ours but with multiple sectors where agents may have different spending composition and where actually monetary policy can backfire. You can see that if you think about our model and think about that you have multiple sectors, some that are more complementary, the produce goods that are more complementary and some that shut down and some that are more substitute and spending are different and because monetary policy is just homogeneously affecting all these sectors, you can see that things can go either way. So I think that's an interesting paper. In our paper, monetary policy is always beneficial. Now, if there is beyond the standard stimulus effect of monetary policy, I think that there are other dimensions in which monetary policy actually may have positive effect on the economy. And in particular, more thinking about the medium run. Why? Well, because one run for the recovery are potential losses that comes from job destruction or from business exit and we have these two extensions in the model. For example, think about labor earning like when if you have some value in job matches, so if it is costly to find the right job match when you have some shutdown in some sector, you may want to keep alive the job relationship and try to provide stimulating firms to keep their workers on the books so that when activity can resume, they don't have to pay the extra cost of finding a right match. And so these may be good because it's not only providing the same social insurance can be provided with targeted transfers because now the workers are not losing their income, but it's also good because in the medium run when you have a recovery, this may speed up. Now, monetary policy here helps because it's kind of, in a sense, making the horizon for business longer by reducing interest rate and even more if we think about these effects of monetary policy, even more targeted monetary, non-conventional monetary policy like the liquidity provision or credit provision may actually even be more effective. Maybe other policy that can help in terms of incentivizing labor ordering, for example, a more fiscal policy like Pulsar Bay in Germany or Casa Integrazione in Italy and so forth, like one drug of this policy is that if we think that there is some structural transformation that is happening in response to this type of shock, actually this policy are going to kind of sclerotize the economy a bit. Now, you may make them more temporary and that can partially solve the problem, but monetary policy or credit easing or liquidity provision may be a more effective way of helping the jobs that are going to self-select themselves by their future prospects. So, let me conclude. What we want to emphasize is that the pandemic is a shock that is asymmetric to the economy. This is our main point. We think it's important to think about the pandemic in a multi-sector model and why it is important. Well, because economic activity is going to reduce some sector directly hit by the shock and it may be efficient that that happens, but you still may have feedback effect on the other sector in transmission of the shock that may generate demand shortages and so this may make policy important. Which type of policy targeted transfers are going to be more beneficial than government spending or conventional monetary policy in terms of stimulating demand in sector B because not only are going to achieve higher level of economic activity but also providing social insurance and in a situation where the shock is asymmetric, social insurance and there is lack of incomplete market, social insurance is particularly important. But monetary policy may actually be important if we think about more in terms of the fact that it's lengthening like making the horizon of firms longer because it may help in preventing job and business destruction. So it may be more important if we think about the medium run effect. Many thanks Veronica. So please send me your questions via the chat function and while they're coming in let me just start with one question Veronica. So you highlighted the importance of these exposed targeted transfers. So how should we think of the design of unemployment insurance through the lens of your model? Because typically that is a more blanket form of excellent arrangements and you seem to suggest it should be more targeted. So I think unemployment insurance is not the perfect way of achieving targeted, but it's a good way of achieving targeted insurance here because there is going to be the most unemployed people are going to be in the targeted sector. So I see the extension of unemployment insurance as a policy in line with the objective of achieving targeted accounts that is going to be more employed people in the target. Another question is when you speak about monetary policy of course you have a more narrow construct in mind than what central banks, many central banks are doing in the form of interest rate policy asset purchases and the like. So asset purchases also affect fiscal policy which you are also thinking about let's say an EMU context which your policy take away slightly change in terms of effectiveness. We have a very simple way of thinking about monetary policy so of course in the dimension it is very important that there is fiscal space for governments to do these fiscal policies that help achieving insurance and stimulate the economic activity and so of course any policy anything that the central bank can do to help giving more room in this fiscal phase is quite important. On top of also as I mentioned before I think helping I mean we don't have any explicit liquidity problem of businesses in the model but for example in the model we have an extension where we think about possible like endogenous exit and possible cascade effect on businesses and so there are clearly here problems of surviving in a moment in which you have a lack of demand and so providing as I was mentioning like in general more liquidity and reduce the cost of credit in the economy is going to help in this dimension while reducing this effect of demand chains that can disrupt economy especially even more in the medium run. Yeah so also so let's say the ECB type of actions in credit markets would also help to complete markets right? Yes. Okay so another question is coming in from Klaus Masuch he says if targeted fiscal transfers are more efficient than expansionary monetary policy would it be a problem if monetary policy is providing subsidies to banks with overtime maybe reduce fiscal space So basically asking about the scope for targeted fiscal transfers to work as hit by the shock more directly Sorry policy on banks or workers? No so if you have if you make monetary policy so if fiscal transfers are more efficient than expansionary monetary policy could it be a problem that's if monetary policy is actually indirectly providing subsidies to banks Well I guess yeah I mean I guess over time you reduce the scope for these targeted fiscal transfers that in your model are more efficient Yes I mean I guess as you were mentioning also I mean we take marketing completeness as given so of course anything that can help also improving in that dimension so maybe making bank more willing to land workers and so on can relax that incomplete market from a different angle instead of directly introducing targeted transfers assuming that that incomplete market are needed So I'm sympathetic to that possibility as well so targeted transfers are one way so the objective is to reach those workers who lose their income and try to make sure that they can spend and so you can do that either by the targeted transfers you can do that by helping them being able to borrow more and so that in that sense improving the financial setting can help as well Okay so thanks a lot Veronica