 First question. Andrea. And this is for Tony. So I made a presentation in Santiago, in Sao Paulo, saying that inequality in Brazil was falling. And the guy, Giuseppe Verdi from Brazil, not from Parma, from Busseto, basically he said, are you counting the asset held abroad by the Brazilian? He said, no, there is no way to do that. But to say how large they are. And this gentleman, Giuseppe Verdi, said a huge amount, which would affect the distribution of assets and income. So now the correction, which is being done by Alvaredo, Sais and so on and so forth, they're correct for non-compliance with tax return. So you combine the tax evasion, I mean the declaration of the rich, which are missing from the surveys, but not the asset held abroad. And so the question is, how large is it? And would it alter your own picture? Now the second one is a comment. Now you mentioned that this very interesting picture that inequality in income can go up, but inequality in asset can go down. And from what I remember, because Lis, I think when Lis was in Luxembourg, there was a big wave of wealth surveys, including Italy, Germany, France, the Netherlands and so on and so forth. And I expected what you said, that the main component of the assets is stocks and bonds and all that. And this was true in the US and a little bit in the UK. But in all the other countries, 60 to 70% of the assets are buildings. So now what happens during a recession is basically the value of these assets or real estate basically collapses. And in a way, if I were to create a new measure of welfare, which combines inequality in income and inequality in asset, and well, you may come to the conclusion that human well-being is improving because the asset inequality is falling, which is not true, because it's not a distribution of the real asset, but the mark-to-market approach which is used to price them. That is the... So I think that this is quite an important issue because in the ECB, the European Central Bank data, which are updated, they basically show that when there is growth, the value of the real estate goes up and when there is a recession, the value of the real estate goes down. So I wonder how all this would affect your own conclusions, which are very interesting. Thank you. Another question. I have a comment for Martin. I am a great fan of relative deprivation and those type of measures. And if you use another definition, the Yitzaki way of measuring relative deprivation directly from the Ransiment's book, then you end up showing that at an aggregate level, relative deprivation is equal to the Gini coefficient. So you can say that what holds for the Gini holds for relative deprivation, and you don't need to say that you have to... So there can be a negative effect for comparisons with others. It doesn't have to be necessarily positive. And also with a negative, you can say that inequality went down, I think. So this is just a comment for all of you that so much of analysis of inequality happens at the country level, but a looming issue for us is about inequality at the regional or subnational level. And I'd love to hear about efforts to compile data to make that, you know, more possible. Yeah, hi. A question for Janet. I was just curious to see if you had emphasized everywhere that you were measuring inequality for non-elderly households and you separated them out. I mean, I find that a little bit unusual. I haven't done that myself, so I'd be interested to know just what does that imply in terms of all the things that you were saying about trends and levels and comparisons and so on. After all, in Europe, and I guess many of the rich countries, the elderly are a growing segment of the overall population, so it seems incomplete not to include them into the analysis. Thank you. Thank you. So would you please take them in that order? Tony and Martin, then Janet, and Carlos, perhaps you could take the third and Janet the fourth. Thank you. I'll just, perhaps responding to Andrea's comments. Yeah. Let me just say one thing is about what is the overall share of financial and non-financial wealth in the total wealth? Roughly speaking, it's about 50%, and has been. It trended towards non-financial assets in the first eight years of the century, and then since the crisis, it's gone back financial assets are growing faster, but they're roughly 50-50, but there's quite a lot of variation between countries. Usually in developing countries, or emerging markets, there's more real assets, as you can imagine, and then the richer countries tend to have more financial assets. And that does affect, the composition effects are quite important, and certainly when asset prices are changing. I think the other interesting point you make is, is all this just a change in the price of assets, or is it really something else? The trouble is, of course, you could say, alright, we've got our houses, we're not selling them, we're just, the price has gone up, we're lucky, but we're not really 10% wealthier, because we're still living in the house, and for us it's the same sort of asset that we had before. Of course, you can maintain that. I don't know, as a researcher, what one could do about that. You could just somehow, I suppose, in principle freeze these things, but of course we don't have, we don't have access to the data. We do what we can to sort of, with the data that we can access, but I don't even know really what sort of question you would be answering. In the short term, it is true, some of this wealth inequality may be, if there's an increase, simply because, Microsoft shares have gone up, and Bill Gates is richer, as if you look at the Forbes, of course, website, they actually have, they tell you each day whether someone's wealthier or not wealthier, and how much, you know, they've lost, you know, 10 million dollars today, because their wealth has gone down, but it's not a loss of wealth in the way that we normally think about it. So there is, if you like, temporal or some sort of temporary changes I don't quite know how much one could really, it's a huge amount of work to sort of identify that, and I'm not sure, frankly, whether it's going to be worth it, because if it is temporary, you'd expect it to adjust again sometime later. As regards income, I think there is a problem, and of course a lot of, with the wealth side, we know that there's a lot of missing wealth from the top of the wealth distribution. There's all issues about whether or not, what you do about all these people that keep their wealth in sure holdings in various offshore countries. Hopefully, we try and do this because we base our estimates on balance sheet data and we hope that the country agencies have some sort of way of guessing how much is held in these remote tax evasive places. For incomes, it is really very difficult to, we're in the process of trying to match the raw, weed survey data with the sort of top incomes database, but it's a non-trivial exercise for a number of reasons, and it happens actually to be easier to do that for wealth than it does for income. So all I can say is it's going to be a sort of tentative estimates, suggestive estimates, rather than I think very hard conclusive evidence that you can really be sure of. Can she just question about Shlomo Yitsaki's rising deprivation measure? It's really a bit obscure in the literature. I have a clever paper, but it doesn't really mesh with anything else. Sorry? Oh, sorry. The Shlomo Yitsaki approach to relative deprivation exists in the literature. It's sort of pretty much on its own. Everybody else has taken a different approach, including in sociology and as well as economics. So it really sort of sits there. I haven't taken that approach. I've taken an approach which it fits as much more mainstream within the literature as I see it, but it also allows me to encompass the relative deprivation approach within an approach which puts positive weight on national income. I'm trying to encompass that. I want relative deprivation as one extreme in a view that says that there's some generalized external effect of living in a richer country on your economic welfare. So I need an approach which allows that encompassing. I don't, as a recall from the Yitsaki paper, I don't see how you do that. But that's my only answer. It's really, it sits there in the literature, but it's sort of very different and doesn't work with what I want to do in my paper. Shareen's question about subnational. Yeah, there's a lot of work on subnational, not at the level of these kinds of studies in the literature, but I've done a lot of stuff on subnational in China, India, Indonesia, elsewhere, and others, many others too, people in this room. It's there. We're not yet at a point where we've got a comprehensive global picture of inequality and poverty built up from subnational data sets. It's not there. We will be there at some point. We're there for urban rural splits and we're there for subnational in a number of countries, including those I mentioned, but on a comprehensive basis it's not ready. It will happen, but not now. Well, I guess I fully agree that the geographical dimension of inequality is one of the most important, especially in developing countries in large countries like India or China, where they explain a lot, not only of the level of inequality in one given moment, but also somehow the trends over time. But unfortunately, I don't think there are databases that allow you to have a global perspective on that. You can have, I mean, when you have access to micro data, of course you can analyze the geographical dimension by regions, provinces, et cetera. In some cases, like China, it's difficult because sometimes the data you have will not have information for all provinces, for example, but for large geographical regions, at least you will have. For example, in the case of the WID, we have information for rural and urban, which is also a geographical dimension that is important in terms, again, of between group inequalities, but also on the evolution, the trends within each of these areas. But yeah, I fully agree that somehow, especially when it comes to large countries where a region can be larger than some of the most important countries that you find in Europe or in other areas, that should be taken into account as far as possible. Yes. Actually, I just want to say briefly, Andrea, your question just about the Luxembourg wealth, the data that we have at LIS, the wealth data. Many, many papers are disaggregating into the form of the wealth. I think for exactly the reason you said that housing prices are very complicated. So we have in almost every case, we can identify financial versus non-financial wealth, and the non-financial is even separated into housing investment versus residential housing. There's also the question of the debt on the housing because two people can have a house presumably of the same value, and one who owns 90% of it and the other less, and we have all these cases in the U.S. where the value of the mortgage is greater than the value of the house. So most people back out the housing to try to get a complete picture. The subnational, I think, Carlos said that already, there are some actually very nice list papers looking across the U.S. states, the German lender, the Canadian provinces, and what some people have done is also create some regions. So I remember a paper done a few years ago combining Luxembourg with the Belgian, French, and German neighboring regions, given that it's obviously an economic zone. But micro data, usually these papers are not at all theoretical, so they're somewhat lacking. Anyway, Pete, let me come to your question about the elderly, and you asked that, right? I'm trying to see over there. Actually, I did a research brief last year with a couple of years ago with Bronco Milanovic, called What Happens to the Story when you take out the elderly? I think that's literally the title. And actually, it's really important, so let me linger on it for just one second. I don't mean at all that we're not interested in elderly economic well-being, but this pre-versus post-story is completely different across countries if you disaggregate elderly and non-elderly. And the reason that we realize this, and I have to say perhaps immodestly, I think that was missed in the literature for a really long time, but we kept seeing the same thing, which is that market income inequality in the U.S. was not especially high, so it looked as if the disposable income inequality was high only because of meager redistribution, but everybody kept pointing out to us that seemed so odd, given that we have such huge wage dispersion in the U.S., the highest in the OECD of earnings dispersion, how come household market income inequality wasn't especially unequal? When you pull out the elderly, the story shifted exactly in that direction, and it's partly because the retirement ages are so different. So your 63-year-old American has positive earnings where your 63-year-old suite has zero earnings, say, and is a pensioner. So in a sense, it was distorting market income inequality in countries with earlier retirement ages and telling a story that made almost no sense. And so when we did it cutting at 65 and 60 and earlier, and when we sort of straightened out the variability in the retirement ages, and the fact, I mean, pension systems are, of course, by definition, are redistributional. U.S. market income inequality in terms of cross-national ranking soared, and then the story made much more sense and looked exactly like the OECD ranking of earnings inequality. So again, that's the story, is that if you've got the elderly in there, you're going to exaggerate the market income inequality in very pension-generous states with early retirement ages. So I think study the elderly, of course, but treat them, I would just make an age cut early and then put it back together. No further questions? Then I'll abuse the chair by making a comment. Karczak Barsou invited us this morning to think about history. So I'm reminded of Nick Kraft's work on the evolution of national income and life expectancy in northern Europe in the 19th century. And what happened at that time, of course, was that European meaning comes, rose began to rise steadily and they didn't rise in the south. And the other thing was that life expectation mortality fell and life expectation rose in Europe. So if you're of the opinion that if life is getting slowly better or moderately quickly better and you're living longer, then that's better still as it were. That's a reinforcement. And now if we go to 1950 look at the evolution of GNP ahead and life expectation in north and south, what do we see? Well, income ahead has grown more rapidly in the south but there's been a startling improvement in longevity in the south much more strongly than in Europe. And this suggests that global inequality if rising is mitigated to some extent by this effect. I don't know whether any member of the panel would like to respond to that. I guess I would just simply say I've heard Angus Deaton of course speak and I'm not sure he would put it the way you've said it that the changing longevity was mitigating the inequality but that both are happening at the same time of the floor under many morbidity and mortality statistics is rising while income inequality is also rising that they're both happening side by side. Isn't that the thesis of his book The Great Escape? Is that what it's called? I think so. There is a literature on this of course and it's all about how you aggregate income with life expectancy and a series of papers and look at where you have a generalized welfare metric incorporating the length of life as well as average income in an obvious way. I can bunch of papers but you can do that. You can make me a desert. Thank you. We are now indebted to our panel for splendid presentations. Let's give them a round of applause.