 Okay, so now I open the floor for questions. Thank you, thank you for a very interesting discussion. Okay, so my question is that, you know, I mean, I'm hanging down with the World Bank, right? And, okay, we work a lot with microdata and we understand that, you know, there can be like inconsistency, right, for the same country over time. For example, South Africa, for Vietnam, we also have a lot of inconsistencies over time, right? And of course, I mean, when we look at across-country analysis, then of course, we also see that as Carlo mentioned, right, some countries, we have a lot of consumption, maybe mostly for lower-income countries, but then we come to lack, then we have mostly income data, right? So, okay, so my question is that, you know, given the competing, you know, like databases out there, right? We have here, we have new database, and then we have some, even some companions, I mean, to it, as you said, and then the Paris School of Economics, right? I mean, top income inequality database, and then the revised database, the World Bank, the PIP database. So, my question is that, what do you think, say, for example, when we look at some countries and then we see some inconsistency, right? For example, if we look at Ghana or if we look at China, right, we see only different estimates from the different databases, which one should we show? Or do we have to make a decision that if we use one main database, then we have to stick with it and we just ignore, you know, the other databases? So, yeah, any advice, I mean, guidance on that would be very much appreciated. Thank you. Maybe we take a couple of questions and then you answer. Any other questions? So maybe you can start answering. That's above my pay grade. Any of you want to take the first break? I think his question was about the fact that if you look at all the different databases that we have, there are differences. Which one do we choose? Basically, he's asking for advice, right, on which one to use. And I honestly wouldn't be able to give that advice. My suggestion would be that, you know, as economists, when you're doing all these analysis, you would want to check the robustness of your results, right? So I, if I'm doing any inequality research and, you know, there are so many databases, I would definitely use all of them and then, you know, put in a caveat or maybe try to understand how, you know, the different databases actually measure different things. So if you look at the case of Ghana, for instance, and if we compare the income inequality that is reported by the companion and what we have with our national database, we can clearly understand from the methodology that they used, we can provide a reason, right, for why there's that difference. Because Ghana was never part of that database. So it's more like using a regression to predict what Ghana would be, given that, you know, it's a country from sub-Saharan Africa, right? But then we know that countries in the region are very different, right? So in that case, that would be the reason for this difference. So I think the key here would be to understand how the different databases measure these different, you know, the same income, you know, measures so that that would help you to explain the differences. Other than that, I wouldn't be able to, I would think, I cannot sit here and tell you, okay, use the UNU wider database and forget the, you know, the World Bank database. I wouldn't be able to do that. But this is my advice or my suggestion. Thanks. Yeah, so maybe to mention also, I think one thing is to look at the inconsistency. For example, when you talk of Kenya, when you look at the three, for example, the three measurements that we have, that is, for example, Asia and also the, the, the, the, our, maybe the, the, the, the, the, the original. You find that those two later, those two are more like trading together. But then you want to look at how it was with the companion actually constructed. And in here, you'll find that you want to go into your data and see whether do you have those kind of particulars that perhaps were considered in there. And for example, in our case, we found that most of the information that we missed could have been partly what could have caused that kind of a scaling effect that we saw. But in terms of the trading, you see it's like the, the, the, the same, the same measurements actually giving the same trading. So in a way then perhaps as mentioned by Monica here also, the robustness of the, of the data that you have does it really speak to the, to the kind of issues that, to the kind of differences that you are observing there. But again, I think I don't have that. I can't really say for sure that you need to choose one, but I think it's important that you look at those which are more, which are looking more consistent in terms of the measure in terms of the trading. Thanks. I think that one's being asked to make a judgment call as well, right? So it's like a wood from the trees issue. So in South Africa, for example, we've got no database says we, that we've, that we don't have a high inequality that's off the charts so high that, that everything that people worry about, about inequality will, will permeate that society. That's the level sort of issue, right? The inequality doesn't change that quickly. And so what's the big picture then? The big picture of the trends is that it stayed high. It stayed high over a very important part of our, of our history. An important, of our sort of democratic era. It's remained high. That doesn't mean it stays the same, but that's our job to, to make sure that the, you know, that the two points are in the mix here. One that, that inequality hasn't, hasn't declined. And for a while that was a, a site of struggle with our government. They weren't very comfortable at all with us measures saying that. But, but come 2011 or so they wrote it into the national development plan as a fact. And then said, okay, what are we going to do about that? Now that's our job, right? I think, but you've got to get that big picture level right. And then, then it turns to the texture. Because if you're going to then give further advice about, okay, what are we going to do about this? You, you've got to bring an understanding. So I guess that's, that's my answer. One of our roles perhaps is to prevent people, you know, so if there's a, if somebody's favorite database exhibits some quick, some quick change in inequality and either down or up, somebody's going to be happy with that. Now, is that really true? And should that be fed in as the main point of the narrative? That's a very important role, I think for, for all of us. And it squares with the whole texture of the inequality. Okay, thank you. Finnta from the University of Copenhagen. I cannot simply resist not, not to try to add in a little bit here because I have had some experiences with sort of these types of complications that arise when you're sitting there with different databases and so on. And the first point is that there simply is no substitute for getting to understand what are the underlying assumptions, what are the potential reasons that particular databases have been put together in particular ways. And then maybe you do want to, when you have defined what your specific purpose is, then do look around and see and for example discover, ah, Professor Steven Jenkins from London, who is one of the most informed, bright, absolutely on top of this, has actually written a paper that's published on comparing the different databases where he actually gives advice on what you should use and what you should not use. So, I mean, now, unfortunately, this is older versions and, but I mean, there's still some quite deep insights in that. And I mean, just to give you one provocative example, I mean, one of the things that we, and this is now a while back and things have evolved, but I mean, some of us could really worry it when researchers started using the so-called SALT version, the so-called SWID, of the WID database. Why? Because they had basically plugged in numbers, the missing values had just been plugged in from an algorithm. So that in built patterns in the data, which then actually affected the results. And this actually was the database that the IMF then used when they did a number of quite influential studies that have been published of our journals and have been very influential and have certainly been changing the policy debates, but the question is whether we as researchers are happy with using what is clearly, what would I say, a dataset where you have inbuilt through an algorithm connections in the data which may not be there in reality. So I mean, the trick here is to get to understand what has actually been done to understand what might be the particular institutional interest that might lie behind a particular dataset, and then really get to understand that. And Stephen Dengler's paper is still very worthwhile reading to come on top of these issues. Sorry, this got a bit long. Thank you. But there are other people. Thanks Carlos. Thanks. Yeah, what Finn was talking about these comparison of databases, the paper by Stephen that you're talking about was in this special issue that Nora and I edited, and I think it was an interesting attempt to compare these compilations. And I just wanted to take this opportunity to say that I think WID, particularly under Carlos' leadership here, has done an amazing job in reacting to some of the critiques that were made in Stephen's papers. And I won't even comment on SWIT. I completely agree with Finn. But my real question to you guys is just to tell you a little bit about the Latin American example and then ask you whether that's the plan for Acer. And maybe I should know that already. I apologize. But what happened in Latin America was that there were all these different national statistical institutes producing the surveys, much as you have everywhere else. And then this one university in La Plata, Argentina, right? Sadness. Well, the University of La Plata. A couple of people there created this little center with a small seed grant from the World Bank where what they did is they had a lot of PhD students and they worked on the data collected in Latin America doing a soft harmonization. Not quite at the least level where variables are harmonized consistently, but at least some kind of harmonization on the income side. And they put it out as a comparable thing. I think you referred to this earlier. And they had a deal with the World Bank, which means that Povkaunets, which is one of the series here and one of the series in WID, uses the Settlers data. This meant that what Settlers did effectively means that what the data they get and treat is the Latin American data that predominates over the international discussion. SEPAL still has its different ones and so on. So as I listen to you guys, the question is, and again, I'm sorry, maybe I should know this already, but is this exactly what you're doing? Is this what you're planning to do? Is getting these deals with all these national statistical institutions, bring them in and harmonize them in this way and create... That's the plan, right? Because if that is the plan and also I wouldn't minimize the importance of as a former staff member, we have Hayan here who's a current staff member, but as a former staff member of the bank, if you get the bank's experts, which are serious people in the poverty global practice looking at this stuff, if you get them to look at it with you and say, okay, we're going to take the Povkaun data for Africa from these guys, then you are going to be in the Settlers position and you're going to be basically in business. Sorry. Hi, thanks. I'm Marcus Yanti from Stockholm University and this is a question which is directed to any of the panelists. Do you have a question, then, kind of follow-up comment? Do you make use of national accounts data in kind of standardizing the series across time? It strikes me that, for instance, one of the uses for that after-suitable kind of weeding out of different pieces would be to, for instance, the growth incidence curves that you get for Africa that should essentially coincide with household sector income growth across these countries. So I'm not suggesting kind of trying to replicate any distributional national accounts and also not suggesting that you think of national accounts as being God's final truth on income growth, but there is, at least in some kind of... There are pieces of information that can be sifted out which can be helpful in figuring out if your standardized survey results are being consistent with the national accounts numbers. So the question here is, do you make any use of the national accounts in doing the standardizing? And if you don't, then I'm kind of suggesting you do. Well, I just have a one additional comment, maybe also another question. You know, it's great that on the discussion from Fin, right, and from Chico, and from Makazanti, very, very helpful. Also from the panel, right. But then my follow-up question is that in that case, so let us say, you know, coming from the ground, right, from the local viewpoint, well, I guess you will know the most about South Africa than on the bus, then of course you can benchmark the data for South Africa, you can put your finger on whether the data makes sense, the chance makes sense, right. But then, you know, but then from an international viewpoint, from an across-country viewpoint, maybe what makes a lot of sense, right, the theory makes a lot of sense for a particular country, but maybe in terms of cross-country comparability, you know, it may not be so, right. So I wonder, you know, maybe can you comment a bit on the chat, you know, between like nationality, I mean of the data in a way, and then, you know, the internationality of the data, right. So that we have, you know, we can bring it to you guys for more comparison. Yeah, thank you. Okay, thank you. A lot of questions. Very short time. Who wants to answer? Maybe. Yeah, so maybe to reflect a bit on the national accounts, perhaps, of course, we've not used the national accounts, but we have used something very close to that in the use of what we call CEQ, where you use, of course, some national account numbers and also survey data, and we are, in that case, it's like we come up with an income concept which is comparable to some extent with a companion. To this extent, maybe we are able to compare that with a companion, and to some extent, then you could say that the CEQ was, which is of course an income kind of measurement, was very close. So to say, with the original video which has been using a consumption per capita data. Okay. Yeah, so I'll start with Chico, and I mean that's the dream, right? And it was written into the original proposal when ASAP proposed itself to the African Research Universities Alliance as the centre of excellence on African inequality and poverty. And that, yeah, that is still the aspiration. And so under who's funded our sort of, AFD funded our first few years of work and with a strong commitment to exactly that, it's, yeah, maybe we need some lessons in political economy as well from you because it's, I don't know how that came together, but in the African context it sounds almost miraculous. Why we proposed it was because there is a data centre at the University of Cape Town called Data First. That's where Taquanisa worked for a long, long time. And that's been doing exactly that for South Africa and like getting harmonised weights so that you can do comparisons over time, et cetera, et cetera. And it has the aspiration of doing that into the continent. But, yeah, so we need to find, yeah, I think your point about needing to integrate in with the World Bank and things, maybe that's the way to push forward a bit. We have that aspiration and this particular project actually was extremely helpful to us as well because people don't naturally dive into these issues of data quality. They just don't. And so I think Acer did quite well in that regard, but you can always do better. And I think this project has pushed us to do better. It's in our genes at the University of Cape Town. But we do, we run surveys and stuff. That's not uniformly true. And so you need some vehicle to set up the institutional capacity. I guess what you're talking about is a much stronger central role and I think that is the way to go. Actually, we spent a long time building the capacity in each of the nodes and I think that's incredibly valuable. But somebody's got to run a project like this. So that's a quick, yeah, that's our dream. So yes, thank you for reminding us. There is some work. We have done some work in the distributed national accounts sort of framework, but not a lot actually. And some of it, obviously there was a bit of an allergic reaction in a sense because at some stage, there was all this using in the growth, cross-country growth equation literature. There was use of household surveys for the distribution, but then some sort of reweighting around some national accounts means that was doing a lot of violence to what was going on. So that wasn't a good start on the continent. But your point is extremely well taken that we need to use the data that's available in the sort of triangulation. I think it's an excellent point. And we'll have to answer your point later, I think. But yeah, thank you very, very much. Okay, thank you very much to all the presenters.