 Thank you very much for a very interesting presentation. I have one thing that I kind of, I mean keeps bothering me in the back of my mind regarding, I can certainly understand that poverty numbers go up because very small changes for vulnerable groups, you can shoot below or you can fall below the poverty line. That story I can kind of follow. But I have a little bit of a hard time, I mean getting it intuitively right in the back of my head about inequality. Because, I mean, I mean when you're sort of looking at it and then you sort of say, oh the rural sector was not affected so much, I mean, assumedly the poor are in the rural sector and the ones who are losing are the ones in the urban sector. I know that there are poor people in India in the urban sector but there's something about that combination of the major numbers of poor in the rural areas versus the urban centers where you have the relatively richer and then we are saying inequality is shooting up. There's something about that which I mean, and I was wondering whether somebody can kind of help me crack that knot so that I can instinctively understand it better. And I should say that the reason I'm asking the question is because we have struggled with exactly that issue in a paper with Peter Lamview and others. So, that's why I'm asking the question also. Okay, and one more question by Nassmo, yes. Thank you. So, my question actually relates to the context of your study. So, India didn't experience a V-shaped recovery and yet you have striking patterns of the poor, you are faster, you recover, so very sharp. So, I was expecting L-shaped pattern for the poor. So, and I was reflecting on profiles of households or groups. So, for example, female-headed households. So, my struggle is really to kind of connect with these patterns that for certain households, I mean, is there much heterogeneity in the way you have profiled or whether your household data does allow you to kind of look into these demographic specific patterns of recovery, just to sort of understand because Indian economy is not experiencing a V-shaped recovery pattern, but here you have striking contrasting patterns. Thank you. Okay, very quick answer and then we will do the second one. So, I'm not sure, so I, what your main point was that the poor are really concentrated in the rural areas and rural areas didn't suffer as much. Urban are richer people and they suffered more, so how is inequality widening? I'm not sure. So, there is a lot of heterogeneity in incomes in both rural and urban sectors and there's a large section of the urban sector, people who classify, who come under the urban sector, which are engaged in informal work and they are, if you haven't done the comparison, they can be much poorer than the rural counterparts. That's one. So, that distinction really will not bite into the inequality numbers. And second, the rural sector suffered less, so especially in the event study framework when we were seeing the regression coefficients, we find that the drop for the urban sector was around 45%. Rural sector was 35%. So, in itself, it's not a small number, that's one, and the rural has, I'll be wary before I say that the rural are poorer than urban. So, I'm not, I don't think there's a mismatch really between these, but if you have more to say, I'm happy to talk about that too. Hi, so the V-shaped recovery, so the decile graph that we were talking about, we are talking about dynamic deciles here. So, what I'm doing is I'm tracking deciles. I'm not tracking individuals, and that is primarily because of problems with the dataset. We were, we have our concerns with using this dataset in the panel format. So, the story that you're talking about is really tracking individuals and seeing how the poor were doing overtime, how the rich were doing overtime. Here, we are talking about deciles really, so how the poor as deciles was doing overtime and so forth. There's a lot of heterogeneity study analysis done in the paper. This is, by the way, a wider working paper, so I'll be happy to have your comments, and we can talk about the heterogeneity bit in person too, if you want. Thank you. Yeah, very, very quick too, okay. So, one is about, you know, why are Paraguay and Mexico behaving so differently from the rest of the countries, right? Because that can really help us understand what can we do better in terms of reducing these inequalities. The second one was about the employment rates plot that you showed, right? When you're looking at by gender, on the Y-axis, you had it at one, right? So, is, yeah, so maybe, you know, like I was not able to follow that part. One more question, thank you. You, okay. So, my question is about the sectorial decomposition of the people who lost their job. Because during COVID in the U.S., we realized that people who had jobs like taxi drivers, restaurants, attendants, didn't have to go to work. So, but people who were high-enders were able to work from home. So, how does that influence the disparity in income, given the job composition? Thank you. I was wondering if you can exploit the panel structure that sometimes some of these data sets for some countries have. I'm curious about the transitions, because some of the levels right of employment and all those are going back to normal. But I wonder if there are new household arrangements, for instance, where people who were employed before the pandemic are no longer employed, but someone is in the house who's not working. And that means that perhaps labor for us looks quite different now. Thank you. Thanks. I mean, very briefly, and it's basically the same question. I mean, the income share of the poor went up. The income share of the rich went down. Implication, inequality, went up. I mean, there's something with that which I just don't get. And then urban, more affected. So again, it bothers me a bit. And I know that there may be something that's very clear for somebody else. I just don't, I don't have it yet completely. Thank you. Thank you. Very good question. Okay, very quick answer. I'll be super fast away. So, one. One, we all basically, we treat it like the base. So it's all related to what the initial year that I think it was 2019 January or the first quarter. And so with respect to that, how it changes. So it's all normalized to one. That's why you see that. Why those two cases? Mexico and Paraguay, super interesting. I don't have the answer. If anybody has more ideas and more than happy to talk about that, could be related to how the government implemented lockdowns, for example, or other things, but brief answer, we don't know because it's not the government transfers and it's not the remittances, but it's super interesting. So definitely something to look, I agree with you, 100%. And the sectoral, the sectoral, we observed also in Latin America, a decrease in employment in the tertiary sector, but also in the secondary. And it's like increasing the primary sector. So this is how it changes. Now, for the income, for the inequality, I showed you up to 2020, but those sectoral changes are both in 2020 and 2021. The implications for that, we have some data in which we look at the income. I don't remember it now like this super fast, but we can look at that. Thank you for that. And panel would love to, absolutely. This was because we were going fast and we wanted to have a general idea of how much COVID changed. Is it something new that we have to are there new things that we have to keep in mind or is it just the previous issues getting deeper? But definitely something that we would love to explore further and on the income. So the initial one is something in general where we were looking at what happened between 2000 and 2019. So you see those increases or decreases are perfectly aligned with the decrease in inequality. Then what happened in 2020, I'm not showing there the poorest and the richest. And I understand the concern. My guess is that because the urban were more affected, it depends on the shares. It depends is like what we were saying before with you. My guess is that there is a large share of people in urban areas that were affected and became poorer and that overall impacts in inequality. But for the sake of it, I'm more than happy to keep on brainstorming on this. Thank you. Thank you. Thank you very much. Very interesting presentation, regional analysis and for all the insightful questions. So for the next presentation, we have another Rostrom from Japan, right? And so you can present the impact of the pandemic for MENA, because if you work on MENA, then you know that data set for MENA is even more difficult than other reason. Not because MENA is poorer, not because MENA is poorer, but because data access for MENA is more difficult. But this issue and from the Indian context, right? Because if schools are closed and women majorly are responsible for childcare responsibilities, then the opposite would have happened, right? We would have seen more women withdrawing, but I totally understand it's the income channel kicking in here because they need to support the husbands, right? But given that we are talking about a COVID period where jobs also were fewer, right? There was a crunch in terms of the available jobs. So how are these women able to find these jobs given that even men are losing the jobs that they had in the pre-COVID era? So that's what I was not able to reconcile both of these two findings together. And the last thing was about the losses, right? You find that the loss in terms of 25 hours less than that or even above than that were going down, right? That means they were supplying more intensively. But that could also be resulting from the fact that there are women who are moving out of the labor force, right? So this data is coming only from the set of women who are currently engaged in the labor market and it's probably the ones who are already engaged intensively that are capturing this effect, so. I think my problems are basically statistical questions because there's a lot of dummy variables in your model. So there's a tendency that you have a dummy variable trap. How did you actively solve, get out of that problem? Because if there's a dummy variable trap, there's going to be multi-coloniality between your variables which can actually affect the science you observe on some of the coefficients. And the next is the fact that you had a negative relationship on people who have more, six or some number of children, they are less likely to be unemployed. And then on the other equation, they are more likely to be out of the labor force. I think those results are sort of counter-intuitive, right? If you are less likely to be unemployed, then it means you are more likely to be in the labor force. I mean, what was the dependent variable? Because that result feel a little bit counter-intuitive for me. And also the labor force, what was your definition for the labor force? Very interesting work, thank you, Nada. So I think my first question is related to the first question that was asked. And I just wanted to know more about the sample of women that you're including in your regression, because whether those are women who were all working at some point before the pandemic, and now they can be unemployed or out of the labor force, I'm just a bit confused about this finding related to the less likely of being out of the labor force with more children, just a bit surprised. The other thing, I mean, I know you're controlling for countries fixed effects, but you have four countries, and for some countries you have two rounds, for some others you have four rounds. I mean, I was wondering would that be possible to see the findings by country? I know the sample sizes are very small, but I mean, I was wondering if at least for those countries for which you have four rounds, if you can show these, because the closure, the type of closure and the timing of school closures were different from one country to another. I mean, I'm not sure about the information, but I'm just wondering how this is affecting your findings. Last, related to this finding about the loss in income, I was wondering if that is related to a decline of the working hours of those women, and whether you're controlling for that. Thank you. Thank you. Okay, so we have a very useful question. Yeah. Okay, so starting with the question on unemployment, so the point is that we're not saying that women found jobs, they're just more interested in going back to the labor force. So that's why you can see that, like it's not that men did not like lost jobs and the women were able to find jobs, it's just that the women were more interested in just getting back. So that's the point. I'm not sure if it's answering questions or not. Okay, okay. Okay, I have a, like I think you had another question regarding to women. The loss in terms of the hours that was given, the density was going up since the density is conditioned on being in the labor force, so probably it's not even after the difference. Yes. I don't know, so the definition we had for this is that we were looking at just on the, in the particular wave, whether women were unemployed or not. So we didn't control for their pre-COVID situation basically. Okay, yeah, and we didn't do that because there was this like recall period problem and the data was very noisy with that. So we eventually dropped it, we didn't control that for that. So for the trap, for the dummy variable and the colonnality, yes, that could be a problem, but I'm honestly not pretty sure how we can solve that. They have an idea. We have explored how many dummy variables are in the model, so that's something. Yes, we will definitely look at. And also for the, I think you didn't, if I understand your question correctly, you were convinced by the idea that the unemployment was decreasing and out of the labor force was increasing like the opposite. So I think that's an interpretation that we saw in the literature for several papers. So usually when the out of the labor force coefficient increases, but the unemployment decreases, this does not mean that really unemployment decreased, but that more people were discouraged and went out of the labor force. So that's something that usually is interpreted this way. I'm not sure if someone has other comments. Question? Yes, and for the definition of women sample, we were looking specifically at women with, so we had, we were conditioning our sample for women, women that didn't have children and women with children. We also had the coefficient for being married, but usually they're very correlated like being married and having children or yeah, that's our sample. We have explored finding by countries and they didn't reveal different results like we have looked at that before. And for the school closure, yes, it varies by country, but then we have for each, it's pooled, it's a pool data set. So for each period or for each row, you have like this intensity of school closure. So it goes from one to three. And so the highest is three and so on. So I think it's still, it can be captured in the data well, so I don't think it's a big problem. And for the loss of income and the decline in the working hours, I think, yes, I think it's, I'm not sure if we showed anything related to the loss of income being significant. I can't really remember, but I think it's. So I'm surprised that you did not control for the length of the lockdown because I think that is a more useful variable when it comes to predicting the effect on unemployment or unemployment. I mean, the length of the lockdown, how many days did the country pursue the lockdown? I think that's a more important variable, the issue, including your model. Yeah, please, I'll be very brief, but first of all, I mean, it's very clear that it's always interesting to study the impact of COVID, but I mean, I think it is important to stress here that we are really only looking at it in 2020. So I'm kind of pondering. I mean, the COVID, the serious COVID really comes after. So I'm just wondering, do you have any hypothesis as to what happened in 21 and 22? And then there's a footnote. This is more for entertaining. I don't know how many people are aware of it, but the Vietnamese had actually hacked the Wuhan computers. So the Vietnamese government knew very early on that something was going on. So this is maybe an informal explanation of why Vietnam was able to act quite quickly. Any other questions? No, it's very interesting that, you know, the lockdown was for only for two months, but still you see impact, two weeks, I'm so sorry, two weeks, but still you are seeing impacts across all the four, you know, because it's either like it's the persistence that we are seeing here, I don't know because I would have expected that the minute the lockdown was removed, you should have moved back to the normal stages. And the second observation was regarding the unemployment rate, right? It went even below, like, but then after a couple of months, it's again increasing the unemployment rate, right? So probably the income effect or loss of income during the two weeks, but that's too short a time period to lead to changes in unemployment rates, right? So maybe you might want to shed light. Okay, thank you. Yes, so very interesting questions, yes. So maybe I just respond to, yeah, you first. Okay, so again, the lockdowns are only for two weeks in 2020, and the only province in Vietnam which is Da Nang City, right? They implemented the second way of lockdown with several more weeks, but other than that, you know, it's uniform, I mean, for the whole country. And as Sofie mentioned, you know, the Vietnamese government, as you know, they are socialist government, so they are in a way very similar to the Chinese government. And when they implemented the lockdown, you know, it's very, very rigorous because, you know, they have a mercenary, government mercenary going down to the community level, you know, weak field should know well, right? So basically, that means that, you know, they implemented it very strongly, right? And in lockdown, everything stopped, you know? You go to the street, nobody's walking on the street, you know, if you look at some recent picture, you know, photo, you know, from China, you know? Then you have some similar ideas. So perhaps, you know, because of that, you know, rigorous, you know, like the implementation, you know, of the lockdown, then we do see, you know, like all the negative impacts. And on top of that, we also analyzing the data, right? For, you know, more recent data. And we also see a lot of, you know, negative impacts. But of course, you know, Fin make a very good point, you know, we need to compare, you know, the impacts of more recent year in 2021 and 2022 with 2020, yes. So for your question, right? You said, did I respond to your question? I'm wondering why you did not incorporate the length of the lockdown? Oh, yes, yes, because it's the response, right? Because it's the uniform, you know, to which for all the lockdowns. So for the length, because it's the same for every province in Vietnam, you know? So same, you know, it's a constant, right? Yes, but indeed for follow-up paper, you know, in 2021, the government implemented, you know, lockdown length of varying length and then we do exploit that, you know, in the follow-up paper, yes. Only in 2021, we have that, yes. Okay, so, yes, so I would like to stop here and if you have any questions,