 We have some time for questions. I also have questions for them. But first, if anyone in the audience have a question, if not, I can go ahead with my questions for them. Okay. I'll go first. And if anyone has questions, I'll pass the mic. You have a question. Okay, good. So let's do a round of questions And then we can all reply. Okay, so for rango, i had a couple of questions. First, a clarification question was how do you construct the index? The index of inequality. I was curious to know a bit more about that. And then I have one question on the relationship you're trying to get at. I would have thought that a predictor of more child mortality is lower access to the quality of, sorry, lack of access to high quality water, to clean water. But I wasn't sure why the inequality within a country, unless I misunderstood the index, why inequality within a country in the access to clean water has implications for the absolute rate of mortality. I'm thinking if what we should be predicting perhaps is inequality of water is predicting inequality of mortality within the country, or if the levels of access to clean water predicts the levels of access, sorry, the levels of mortality. Those are my questions for you. For the, I had a question which is can you look at health outcomes? You were really getting into it when you were looking at doctor appointments and so on. Can you have an outcome that is the health status basically of the person to know if the overall reform is having an impact on the health status of these people? My second question I had is why do we care so much about that gap, that difference of beta 4 and beta 5 that you had in your slide? My interpretation from what I saw is that there is catch up of the rural area with the urban employed. And that to me is really very salient. I don't know why would we care necessarily about the other gap, the gap of where they are catching up in the same speed, so to say. It seems like for inequality, this one catch up on beta 4 or beta 5 that you had was already quite relevant. And that's all on my side. I don't know if anyone else has questions. I'll go over here and then I'll come back over there. Thank you. Appreciate it. I have another question for you that's actually related to that last question. And it has to do with the gap that you construct. As far as I understand, the gap is urban and rural gap, right? But your control group is in itself like composed of urban people. So when you're comparing beta 4 and beta 5, you're having like an urban employed population. Is it employed or not employed? The beta 4. I don't recall the rural. Yeah, so I'm not entirely sure if, like consistent with the last question, if that is the gap that you might want to look at. But perhaps you could go ahead and take the mean of your control group, which is the urban population, the urban non-employed population, and just take the coefficients and look at how much the mean is changing with respect to the urban group. Because what you're having is two coefficients which are, like, both of them are giving you an interpretation with respect to the control group, which is indeed, okay, you have it there, which is the urban non-employed population. So I'm not entirely sure if that gap has the type of meaning that you intended or maybe I'm just not understanding it correctly. So this is the first DID. And the first DID would measure the change in the gap between the urban employed population and the urban resident population, the urban non-employed population. And the second DID is between the top green line and the bottom blue line. That measures the change of gap between urban employees and rural residents. So if you do a diff for these two different diffs, then they will show the differential of change in gap, which is also the change in gap between urban residents and rural residents. You can also look at the illustration here. So this is delta 1. This is the first different DID. This is the second different DID. So if you do a DID between delta 1 and delta 2, that will be the change of gap between these two blue lines. Yes. This is just a control group. This is a benchmark. Yes, yes, yes. When I say urban rural gap, I'm specifically referring to these two populations. I'm not referring to the gap between urban employee and rural residents. Because the consolidation only consolidates these two groups. So this urban employees, they're not impacted by the consolidation. So my evaluation is only using them as the control group as a benchmark. Yeah, the, how do you say, the illustration is a little bit complicated and this is also pointed out by the reviewer of the journal, yeah. As for questions from Javier, so I do have, the data set do have variables like self-rated health, which is very, how do you say, subjective. And we're pretty much worried that the urban people and urban employed people and rural people, they will have different perception of health. So maybe the same disease and well, rural people thinks no big deal, but urban people will be like, things very serious, so we probably don't think it's very comparable for the self-rated health. And for more objective things like, for example, the blood pressure or stuff, the survey did not cover that. So yeah, I think that could be a future, how do you say, future study, a future direction for studies if such data is, such variables are available, particularly for the very subjective medical measures like, for example, blood pressure and things like that. And the second question, what was the second question? There was more comment on the, Yes, yes, why is it important? Okay, I think it's more related to ideology, I guess. It's more of a political thought that, I mean, equality is just very important. It doesn't matter if you're in the urban area or rural area because, like many, how I would say, even today the urban and rural inequality is still very large and just catching up with the urban employees with things probably not enough, yeah. So thank you for the question. These two questions are really, really thought-provoking actually, I've been prepared a bit. So your first question was about the estimation of a CIWAT variable. So water access data was collected from JMP data set, but it's correct that there is different levels of services and why I concentrated on only one of them is that the idea behind it is that the households have water access dependent on differentiated water access, dependent on their wealth groups, but still the households or the population having basic access to water can be also grouped down into the different wealth groups as well. So when it comes to the law-middle-income-country setting, not every family or every household has one single type of access. So therefore, depending on the wealth groups, even these different five types of wealth groups have differentiated access to the basic water services. Yeah, that is the main reason. And the second point is that the service level is an indicator of the water quality itself because the better quality, better health outcomes, more infected, worse outcomes, right? So therefore, I try to concentrate on the highest possible quality level that has available data. So that happened to be the basic one. And then the third point is that the concentration index approach was estimated based on the wealth quantiles. So the JMP data grouped down all the population data into the wealth quantiles. Yeah. There was a second question, right? I almost forgot about it. So basically, did I understand correctly? It is that the levels of water quality, how it actually, how the levels of water quality actually predict the levels of, actually predict the entire child mortality, right? What I was thinking is if you're looking at a measure of inequality and then you're looking at a level of mortality in the country, right? And I wonder if one should look at a measure of inequality on access and a measure of inequality in mortality or a measure of access to quality of water and a level of mortality. I wonder if we should look at inequality and inequality or level and level, basically. Yeah, yeah. It's actually a very valid point. But the idea behind it is that still the differences between differences in access to quality water still exist between the different wealth groups. Yeah. That's the reason why I brought this inequality in access variable to check whether there is impact in child mortality, specifically caused by the Patagon Diarrhea. Diarrhea caused in Patagon. Yeah, that is the reason behind it. But this point is very valid. And since this paper is still under progress, I will consider it. Yeah. Thank you. Yeah. Okay. Okay. So I have a question for Rangu. So I saw you use both the fixed effect model and the random effects model. So did you do a housement test to see which one is a better fit for the situation here? Yeah. Okay. This actually must be somewhere here. So housement test was actually run to decide whether fixed effects are random effects model. So yeah, the result indicated that the fixed effect model was better fit. Yeah, yeah. But for all LS model, whether I wanted to choose, you know, okay, FE over random effects, then whether FE over LS, I needed to decide that, right? And that consideration was, I will show you my empirical model, that consideration was because of this alpha, which actually the country-specific constant term where un-observable effects are embodied. So this is not the case for all this model. That's why I consider FS is better fit. Okay. Well, good morning. Thank you so much for your presentations. My question is for Javier. Well, I was wondering how was your parallel trends shaking or test, your parallel trends, test, how was it? And I was thinking that what is so is that you have a drop in the price and you were exploiting that change in different cohorts. But after the drop, you also have a rise. So you can also, like, kind of do similar exercises in those cohorts. I don't know if you have the opportunity of doing that kind of exercises and what was the results if you had the opportunity. Thank you. I can pass it. Audience or anyone else? No questions? Okay. So first the, thank you. Do you have the, yeah, the pointer? Thank you. The parallel trends, I have them here. This is a parallel trend. So first, before doing the parallel trends, we looked into the prediction that males are more likely to die than females. And that's something we pick up in pretty much all of the predictions. Let's focus for the sake of time here. This is a log cohort size. And males, according to this literature, it should be dying. They're more likely to die both in neutral and in early life. And we see for the panel of males that they are more likely to die. And they are not quite more likely to die in statistical significance terms. But the signs are what we would expect. Now, we then, in the paper, looked at the parallel trend. So we tried to predict missing children by gender. So the cohort sizes, we count them by males and females only. And this is a drop in 1995. This is a controlled cohort in 1984. What you see on the red line is the price. And that's the point they estimate. And males are here and females are here. So it's quite flat here. And then it drops for males. It does not drop a whole lot for women. What we see later, answering your second question, if we look at the extended graph, they do follow the pattern of the price. Now, the problem with making causal estimations with the pattern of the price that is over here. This one here. So your second question was, why don't we take advantage of this, right? We were trying to make very careful identification of where children are born and that there shouldn't be any expected variation on where they're deciding to relocate and so on. In fact, one could show estimates from 1996, which is still a very low price. But then the problem here is that there are behavioral responses confounding the measure of mortality. There is Plan Colombia over here, which increases the price because there is a shift, basically, of the production again and also the cancellation of this policy. This we cannot really tell if they are as exogenous of these first drops. We don't use them as the main identification of the paper. But there have been people asking about this upper part. So we're considering perhaps we should include a section where we show the follow of the mortality rates across the prices as supportive evidence that this continues to hit, but not as much of the causal effect of that price drop. Thank you. Those are good questions. So if there are any remaining questions, thank you very much to everybody. It was a pleasure and have a good rest of your day. Thank you.