 Okay, thank you for the presentations. They are very interesting jobs. And my question is about how relevant are income taxes to these countries, for example, in context of high informality, I don't know how much the countries, they can, the share of the income taxes to the total revenues to these countries to do the relevance of these kinds of taxation, this is just my question. Thank you. Thank you very much for the presentations. My question for Javier, I'm wondering who are these people in the other group in your analysis? For example, can you observe if those people are, they were in different industries before and you track if they move, for example, from agriculture or from wholesale to the other? Or for example, those can be also people that were in the informality because we know that, for example, in developing countries, some people choose to be in the informality sector for some many reasons, one of the reasons is to pay taxes. And what I mean thinking is a lot of people saw that who were in the labor market and lost their jobs, they had some kind of social protection support. And I'm just wondering if it can be some movement from the informality to the formality in low-paid jobs in the case of Ecuador. And for Yucca, just can you observe some kind of descriptive evidence of bunching before this tax rate? Because what I mean thinking is what is the role of composition group in your control group before and after the reform if they changed behavior because of the change of the policy, basically. Thank you. Thank you. You might regret that because I have one for each. But let me try really quickly. So thank you for the great presentation. So for Xavier, the question on my question is on, you had changes over time of these shares and my question is whether your analysis is anonymous or non-anonymous in the sense you said at the beginning you were interested in these top share, these top people before the pandemic. So is it those people as they move or is it, you know, the top 1% at each particular point? And this is related to Amina's matrix, right? So I just wasn't sure. And then just to clarify, just to make sure I understood, there's no capital incomes, right? Xavier in your analysis. Okay, for Amina, two quick things. The capital gains that you have, are they just the realized ones or is this an estimate of the full capital gains including when there weren't sales? And then you mentioned that there wasn't work on merging the household surveys and tax data in South Africa and I was wondering, I thought I knew the Ingrid Woolard had done something on that. So asking about that whether I had dreamt that or not. And then Marcus, so you brought in the national accounts data to sort of have an estimate of income in the informal sector. And then you have your income data for the formal sector and there's a big gap between that and its representation in the national accounts. So again, just a clarification question. So you did scale that up as well by the national accounts, the formal part. So the informal part, I understood at least comes from the HFC from the national accounts and in the top part was there a national accounts adjustment. And for you two very quick questions. One is nerdy identification police question would worry about your parallel trends assumption in the diff in diff as they diverged the year before the tax reform. So... What was the word that was drawn in the year after? Ah, it was drawn in the year. Okay, well, maybe I want to redraw that line, but that's fine. No, that's fine, that's fine. Then if there isn't, I was just wondering whether the tax reform was anticipated or not. But if the line was not drawn in the tax reform, yeah, that's fine. And the second question to you was, so the interpretation of these elasticity, right? So you mentioned income shifting. So there's very different policy interpretations, I guess, or responses to the calculation of the elasticity. If the behavior responses primarily on, I don't know, labor supply or investment response, or whether it is hiding your income or shifting it away or moving it outside the country or not reporting it. And I guess it's impossible for you to tell, but I wonder if you had investigated that in any way. Thanks. Okay, thank you for the good question. So we'll go, I guess, in the same order. Would that be okay? So thanks a lot for the questions. On the first question, I don't have the exact number for Ecuador for, but it's low. The tax to GDP ratio is very small compared to developed countries. However, something interesting is that it is not only because of informal employment. So something that we have done with Ecuador is using micro simulation models is we have moved artificially people who are in informal employment in survey data to formal employment and recalculated for them social insurance contributions and personal income tax to see how much they would, how much it would cost for them to move, assuming they don't change incomes. Okay, something very, something very easy. And what we see is that actually personal income tax increases by very little by 20% from the revenue. This means that it's not a problem of informality only, it's a problem of the design of personal income tax. As I was saying, the exempted tax threshold is extremely high. We are talking about people only in the 10th decide of the income distribution who practically pay personal income tax in Ecuador. And you have this extremely general deduction from personal expenditures, so that make you fall out of the tax brackets. So I think it is both. And I think that's an interesting finding because it's not only about the labor markets, it's about policy design, okay? Rodrigo, we can, the answer is we can, the follow up of the answer is we haven't done it yet, but that's what we want to do. So, and this links to Chico's question. So it's anonymous in the sense that we are, at the moment, we just, especially for the top, we define it in terms of 2019. And for them, we do follow them, what's going on, okay? So these are top incomes that change from year to year. So there are people that were in the top incomes in 2019 and then we see what happens to the earnings and employment. But at the bottom is everyone, okay? So we are not facing for those who were at the bottom in 2019. So there's lots of movements. And we haven't done yet, and this done this yet, but I think this is the next step because what we want to see is mobility. What's going on in terms of mobility, you know? Do people who are top earners in 2019 are they moving to the bottom or are they dropping out from the data? So that's what we want to get a better grasp of. However, the main problem is that we can know what happened to these people only as long as they come back to the data. What do I mean? You have it in the register. If they go out, we don't know whether they became unemployed, whether they moved to informal employment, whether they died during the pandemic. It's impossible for us to know as the only way we can know is if they come back, okay? So what, you know, they went out and they came back. So we cannot do this mobility between formal and informal, but we can know what happened then, you know, in terms of exiting and coming back. And that's the next step that we want to take. Thanks. Okay, so for the first question, and maybe one thing I didn't mention, and maybe it's kind of different from some of the other presentations, is that in the South African context, the information that we actually capture from the tax data, so you're looking at about almost 70% of your employed, so yeah, employees in the country, so from your workforce, those who are employed, 70% of them are actually in the tax data. Informality is not as big, or the informal sector is not as big in South Africa, comparison to a lot of the other African countries. So that's, you know, maybe gives you some sense of that. Capital gains realized, yes, that's what I understand that to be. So it's a Wulad and Basir paper that you're talking about. Yes, and so we follow closely the work that they do and sort of expands on that. They don't quite merge the data, I think. I think they use it as complimentary because you can't actually put those two datasets together. They're just identifying different ways, different government departments that don't talk to each other. But I mean, I think there would be a project to actually do something like that. I think it's entirely feasible, and would be phenomenal, yeah. Hi, thanks, I'll address the first question about how relevant taxes are. In some sense, I don't care all that much about taxes here. I'm using this as a source for income distribution analysis, but of course, the fact that it's covered such a tiny fraction of the population is to be mildly a problem, but it's increasing very rapidly. Well, if 5% more of the population in a decade is rapid or not is, of course, a matter of opinion, but all the same, there is somehow more to it. And in fact, somehow understanding the interaction, I think there's a political economy project here lurking around to kind of try to figure out how formal informal sector composition changes across time and whether or not things like tax policy actually have a role in that, but it should probably be done by somebody who knows how to do that rather than me. But it's increasing, that's the short answer there. And on Chico's specific question, I don't scale up the gross incomes with GDP estimates of formal sector incomes in part because what I didn't say, but of course, is reasonably important, I ignore tax evasion here. And I mean, there are other differences between what GDP captures as formal sector income, but tax evasion is one of them. There are other differences as well. So I only use the GDP estimates to get essentially the control totals to compare with. I focused on the average income here because that's kind of interesting substantively, but the control totals function here is to figure out what the together quantile groupings right. And I don't do the rescaling, but I think that's because there are kind of other differences between the two. So I don't think the right way to think of it is that the taxes, I mean, the tax evasion problem could be kind of taking into account somehow doing this comparison, but there are other differences as well. All right, so regarding them, first the importance of personal income tax as a revenue generator. So I checked numbers from another paper of mine and for Uganda in 2019, it hovered around less than 2% of the GDP. So definitely, I mean, they rely very less on the sub-Saharan African countries than developed countries do on generating revenue from the income tax. Having said that, I mean, it's more than the country gets from the corporate income tax. So I mean, and then there's quite a bit of discussion on the role of multinational income shifting eroding the tax capacity and then the personal income tax, I think is understudied in that respect. And luckily, I mean, the revenue yields are increasing also in terms of share of GDP, not only as a number of taxpayers. Rodrigo, you asked about punching. So yes, we did check for punching. There wasn't any clear punching at the threshold. So really, it appears that where the response comes from are from taxpayers who are far away from the threshold level, so that they are above the threshold level. And that's why punching wouldn't actually identify, I mean, the response on these people. I think you also asked about the potential changes in the composition of the sample or the people in the analysis that we cover. And that's a good question. That's a potential pitfall of the cross-sectional analysis and also something that Syes discusses. Unfortunately, we don't have many other covariates to check because we're using the admin tax data and there's nothing on education, et cetera. We could do a little bit on regions and industry of the employer. So that's we could check. So that's a useful comment. And then finally, Chico, you asked about the parallel trends. So the figure was drawn so that the vertical line was for the first year of the reform and then it was normalized for the year before. And then two years before the reform, there were no differences across these groups. So maybe the figure was misleading but there's no problem in parallel trends. Yes, that's an excellent point regarding then what are the tax policy implications of this? So unfortunately, I don't think we can match the income and the dividend recipients. So in a sense to, I mean, ideally, we would like to, in a sense have a response of the broad income of these top earners but we are not able to match the dividend receipts or any other incomes and then the labor earnings to see how much the broad income covering all sorts of different tax heads would react. But the tax policy implications clearly are very different because if it's only income shifting then the formula that I showed you for the revenue maximizing top tax rates is actually not valid. We do discuss this a little bit in the paper but it's a little bit of a hand waving because there are limits on how much we can do actual analysis on that. All right, are there any other questions, comments? We still have time, so yeah, please. So thank you everyone. Amina, right? Yeah, thank you. So as Amina showed, there is of course like differentiation of the sources of income like between like the different person tiles, wine tiles, like everything, right? So as you know from yesterday's talk, we are in Colombia facing a change in the fiscal policy and there is a lot of debate on how do you tax gains from capital, dividends and so on. And today we have a lot of heterogeneity in that like they are taxed differently. So I just want to hear from you that are experts like in this, like how is this in your countries? How different is taxation to the different parts of income? And I'll be related to the last presentation of elasticity, I was wondering if, yeah, like if there are, like in the baseline we have different taxation, there is a reform and then maybe these parts are taxed differently. And if eventually like, I don't know, you have some experience and you know what happens because I was wondering regarding the transition that Amina was showing, right? Maybe when you change this, there are people that are very managing firms, they are exposed to more taxes then they need to change the source of income maybe the transition just shift all the way around because yeah, you are putting some people that know better to do some things to be better like in the income and distribution. So yeah, I just have that in my mind and I just want to hear from you what are your thoughts about this? Okay, thank you. I have a question on the paper from Quindo. The drop in earnings and employment, do we know the portion of that drop that can be attributed to the pandemic itself like COVID infections and also attributable to government policies like close on businesses and also certain days support on farms. Thank you. Thank you. Good question. And something maybe we need to think about a little bit more. It might be nice to add a slide to say, this capital gains has this tax rate and the dividends has that tax rate and labor income has this tax rate. Of the top of my head, I can't think of the tax rate for the capital gains tax but it is all different. And then yeah, Yucca, did you have some? I mean, I think we've had some small discussion about this. Well, I don't remember the rates. So the rates are different. So there's a flat tax on dividends. 20, yeah. South Africa, there's a flat tax, I believe, on capital gains. Can't remember now on top of my head what the rates are. So this is a very useful practice that I mean, it's hard to leave a progressive tax on capital gains because then there will be timing response. So yeah. Yeah. Thank you. So yes, the big drop in earnings and employment during the second quarter of 2020 is due to the strict lockdown measures taken by the government. These affected industries differently. As you saw in the graphic, there was a drop in all industries but to different extents. In agriculture, something that we found out is that there was a drop in earnings, although this is considered one of the activities that did not suffer or should not suffer much. So contrary to some studies that have assumed, now we survey data doing some now casting that have assumed that agriculture did not suffer during the pandemic, we have evidence that it did. So both in terms of employment and in terms of earnings, as I said, we don't have information about informal employment, but again, from survey data, we also saw a huge drop in employment, in informal employment and earnings with the best we can do. This is only, so what I presented is only earnings from labor. There were some support policies for people during the pandemic. I think it's also related to the question Rodrigo asked before, but we do not capture this year. More, I mean, perhaps also because these were targeted more to people in the lower part of the distribution. We think that maybe some of these people might not have received those aids and it would be quite hard to simulate. So we just focus on personal income tax and social insurance contribution for that reason. Okay, so if I understood your question correctly, you asked, I mean, what are the lessons learned in terms of the responses by various groups of taxpayers? Was that something that you wanted to know? And that's an excellent question. And what tax economists typically think that there's a sort of hierarchy of responses and this sort of reporting behavior, if you wish, that's avoidance evasion is the one which is the, in a sense, the most responsive. And about that, of course, doesn't represent then the real response in terms of anything, in terms of, I mean, investment behavior, labor supply behavior and such and maybe the main lesson in a sense is that you should fix the tax base first so that it becomes harder to do evasion slash avoidance and then take it from there. But most of this evidence comes from developed countries and it's a very rich research agenda to do all sort of this sort of granular level analysis for developing countries. And I think in the Latin American case, you have rich data sets that you can utilize to do this sort of work. Okay, so we are, unless there's a burning question, so maybe we are then finishing a little bit early. Thank you so much for coming. Thank you for the excellent questions. Thank you, colleagues, for presentations.